Pages: 1 2 3 4 5 6 7 8 9 :: [one page] |
|
Author |
Thread Statistics | Show CCP posts - 25 post(s) |
|
CCP Navigator
C C P C C P Alliance
|
Posted - 2010.08.25 15:42:00 -
[1]
During downtime on Wednesday, August 25, 2010, new hardware was deployed on a dedicated node for Jita and is being used to test the impact of a new processor specification and low latency memory. We encourage players to show up and give us feedback on performance.
Navigator Senior Community Representative CCP Hf, EVE Online
|
|
|
CCP Yokai
|
Posted - 2010.08.25 15:55:00 -
[2]
Will post tech specs here shortly.
|
|
Ix Forres
Caldari Righteous Chaps
|
Posted - 2010.08.25 15:56:00 -
[3]
Originally by: CCP Yokai Will post tech specs here shortly.
Don't skimp on the details! Exact model numbers and all the nitty-gritty details, plus comparisons to the rest of the cluster! :D
I'm not much of a Jita person; given this is for the location node, that means inventory movements (thus trading) should be faster/more prompt, and in-space activity likewise, yes? -- Ix Forres - 3rd Party Application Developer - EVE Metrics - accVIEW
|
Qoi
Exert Force
|
Posted - 2010.08.25 15:57:00 -
[4]
Originally by: CCP Yokai Will post tech specs here shortly.
We love you
|
RaTTuS
BIG Majesta Empire
|
Posted - 2010.08.25 16:02:00 -
[5]
1124 ATM and looking good --
|
Victor Valka
Caldari Endoxa Corporation
|
Posted - 2010.08.25 16:33:00 -
[6]
Originally by: CCP Yokai
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Rawr!
Originally by: Spaztick You are not outnumbered, you are in a target-rich environment.
|
Kazuo Ishiguro
House of Marbles
|
Posted - 2010.08.25 16:45:00 -
[7]
Originally by: CCP Yokai As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
I already have some quite nice cookies, thank you. They're chocolate, with white chocolate chips, topped with dark chocolate, adorned with chocolate-coated honeycomb and dried strawberries (a bit like raisins except using strawberries instead of grapes). --- 34.4:1 mineral compression |
|
CCP Yokai
|
Posted - 2010.08.25 16:48:00 -
[8]
Seriously... only 1200... thought you guys had alts or something.
|
|
Rowbin Hod
Cloak and Daggers Black Core Alliance
|
Posted - 2010.08.25 16:52:00 -
[9]
The the location node running on just one of those cores? what are the other cores doing?
|
Bagehi
Association of Commonwealth Enterprises R.A.G.E
|
Posted - 2010.08.25 16:56:00 -
[10]
Logging the Jita alt in now. Can we have pictures of this sexy new piece of hardware?
This signature is useless, but it is red.
|
|
Daniel Jackson
Caldari Planet Watchers
|
Posted - 2010.08.25 17:01:00 -
[11]
Originally by: Bagehi Logging the Jita alt in now. Can we have pictures of this sexy new piece of hardware?
yes please we need sexy **** hardware pictures :3 ___
|
QioX
Caldari State War Academy
|
Posted - 2010.08.25 17:21:00 -
[12]
Im wondering why is it only 1 quad core cpu and not 2 ? also why are you using a 10k rpm sas drive when these days most are using 15k rpm drives ? both of these would improve things alot more than just 1 cpu and an older slower drive
I used to be an engineer in a computer consulting firm in NYC and often I found that many companys understestimated what they needed for hardware and did not leave much for reserves and had issues with performance due to lack of hardware to handle larger than planed loads or future expansions ect. but 2cpus are better than 1 heck 4 for some apps that need this sorta of thing also the OS if its 32 bit ,64 bit makes a diffrence in performance VS 32
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Just my 2 cents
|
Kephael
Caldari GoonWaffe Goonswarm Federation
|
Posted - 2010.08.25 17:23:00 -
[13]
Originally by: CCP Yokai Edited by: CCP Yokai on 25/08/2010 16:20:27 This server is for benchmarking new chipset, Cache and RAM on Tranquility and is not the final or completed spec we have planned. Consider this the Control Group for our testing.
Also, while we are doing tests on Jita, we are also going to be profiling fleet fights and all kinds of other node/service types on the cluster.
Once we get a good load metric on none type we'll be moving the server to test other things.
Thanks in advance for helping get the load we need to see the performance impacts.
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
Why no 5677, you guys getting cheap?
|
Ix Forres
Caldari Righteous Chaps
|
Posted - 2010.08.25 17:25:00 -
[14]
Originally by: Kazuo Ishiguro
Originally by: CCP Yokai As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
I already have some quite nice cookies, thank you. They're chocolate, with white chocolate chips, topped with dark chocolate, adorned with chocolate-coated honeycomb and dried strawberries (a bit like raisins except using strawberries instead of grapes).
I am eating a very, very nice toasted sandwich with ham, cheese and brown sauce. And now I'm hungry again.
Originally by: QioX Im wondering why is it only 1 quad core cpu and not 2 ? also why are you using a 10k rpm sas drive when these days most are using 15k rpm drives ? both of these would improve things alot more than just 1 cpu and an older slower drive
I used to be an engineer in a computer consulting firm in NYC and often I found that many companys understestimated what they needed for hardware and did not leave much for reserves and had issues with performance due to lack of hardware to handle larger than planed loads or future expansions ect. but 2cpus are better than 1 heck 4 for some apps that need this sorta of thing also the OS if its 32 bit ,64 bit makes a diffrence in performance VS 32
The disk won't be doing anything, I imagine - only holding the base OS, maybe runtime executables. All DB work is on the central DB server. The servers are (all?) 64-bit OSes, multiple CPUs only makes sense if there's enough room on the box in terms of RAM throughput, latency, network capacity, etc for enough nodes to justify multiple CPUs. Rest of the time performance-wise you're better off with multiple machines. -- Ix Forres - 3rd Party Application Developer - EVE Metrics - accVIEW
|
Rowbin Hod
Cloak and Daggers Black Core Alliance
|
Posted - 2010.08.25 17:30:00 -
[15]
Originally by: Ix Forres
Originally by: QioX Im wondering why is it only 1 quad core cpu and not 2 ? also why are you using a 10k rpm sas drive when these days most are using 15k rpm drives ? both of these would improve things alot more than just 1 cpu and an older slower drive
...snip... but 2cpus are better than 1 heck 4 for some apps that need this sorta of thing also the OS if its 32 bit ,64 bit makes a diffrence in performance VS 32
The disk won't be doing anything, I imagine - only holding the base OS, maybe runtime executables. All DB work is on the central DB server. The servers are (all?) 64-bit OSes, multiple CPUs only makes sense if there's enough room on the box in terms of RAM throughput, latency, network capacity, etc for enough nodes to justify multiple CPUs. Rest of the time performance-wise you're better off with multiple machines.
AFAIK, they're only able to put a node on one core (multiple nodes fit on one core, but a single node won't fit on multiple cores) and as they're trying to test the Jita location node, they'll be looking at the performance of a single core. Of course, they may shunt other Jita nodes for other bits of the work onto the other cores, but this isn't a case where more cores = more performance.
|
Ix Forres
Caldari Righteous Chaps
|
Posted - 2010.08.25 17:35:00 -
[16]
Originally by: Rowbin Hod
AFAIK, they're only able to put a node on one core (multiple nodes fit on one core, but a single node won't fit on multiple cores) and as they're trying to test the Jita location node, they'll be looking at the performance of a single core. Of course, they may shunt other Jita nodes for other bits of the work onto the other cores, but this isn't a case where more cores = more performance.
Yup - that's what I was getting at. More cores may actually = less performance if you start bottlenecking RAM and network resources. -- Ix Forres - 3rd Party Application Developer - EVE Metrics - accVIEW
|
Arven d'Orrin
|
Posted - 2010.08.25 17:51:00 -
[17]
Edited by: Arven d''Orrin on 25/08/2010 17:51:52 1232 online atm. Feels good, man.
(as in: undocking doesn't take 15 seconds and buying/selling is VERY smooth for Jita)
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 17:54:00 -
[18]
Edited by: Zeba on 25/08/2010 17:54:31
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
You forgot to promise them free stuff and plentiful rewards for 'helping' you guys test stuff. Remember when you asked for help before they all said thats your job not theirs..
Eve online: Home of the Lazy Emo.
edit; logging on my trade alt brigade.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
Obsidian Hawk
RONA Legion
|
Posted - 2010.08.25 17:56:00 -
[19]
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
I do believe we just got trolled by a dev.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 18:02:00 -
[20]
Originally by: Obsidian Hawk
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
I do believe we just got trolled by a dev.
Heh, so lets see what the troll score is atm.
*glances at the front page*
Forums: over 9 billion Devs: 14
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
Monistat Seven
Caldari
|
Posted - 2010.08.25 18:13:00 -
[21]
1330! Still working well.
|
Darod Zyree
Gallente Zyree Holding
|
Posted - 2010.08.25 18:14:00 -
[22]
Originally by: CCP Yokai Edited by: CCP Yokai on 25/08/2010 16:20:27 This server is for benchmarking new chipset, Cache and RAM on Tranquility and is not the final or completed spec we have planned. Consider this the Control Group for our testing.
Also, while we are doing tests on Jita, we are also going to be profiling fleet fights and all kinds of other node/service types on the cluster.
Once we get a good load metric on none type we'll be moving the server to test other things.
Thanks in advance for helping get the load we need to see the performance impacts.
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
call that specs? I was expecting new hamsters, leaving thread disapointed! :p
-Darod- |
Vir Hellnamin
Gradient Electus Matari
|
Posted - 2010.08.25 18:23:00 -
[23]
Okay, no wonder that undock procedure felt bit funny.
-- "Entering MH means instant death. It's worse than 0.0. Even the asteroids shoot back." - Alex Harumichi [GRD]
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 18:28:00 -
[24]
Originally by: Darod Zyree call that specs? I was expecting new hamsters, leaving thread disapointed! :p
Here is the box art for the new server hardware.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
Atticus Fynch
Gallente The Scope
|
Posted - 2010.08.25 18:37:00 -
[25]
I have noticed a difference. Hard to explain since I'm a layman, but the "experience" at Jita is a lot smoother.
|
ghosttr
Amarr Muppet Factory QQ Connoisseur's
|
Posted - 2010.08.25 18:51:00 -
[26]
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
Sorry 18 17 months, you know Prospecting! |
Bagehi
Association of Commonwealth Enterprises R.A.G.E
|
Posted - 2010.08.25 18:52:00 -
[27]
Originally by: Zeba
Originally by: Darod Zyree call that specs? I was expecting new hamsters, leaving thread disapointed! :p
Here is the box art for the new server hardware.
Zeba wins this thread.
This signature is useless, but it is red.
|
|
CCP Navigator
C C P C C P Alliance
|
Posted - 2010.08.25 19:04:00 -
[28]
Thank you all for your feedback so far. Please keep it coming and let us know if your experience has improved, gotten worse or remained the same.
The war on lag is kicking up a gear and your reports are very important
Navigator Senior Community Representative CCP Hf, EVE Online
|
|
HeliosGal
Caldari
|
Posted - 2010.08.25 19:07:00 -
[29]
less lag shame about the rest of the game but good work i guess
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 19:10:00 -
[30]
Originally by: HeliosGal less lag shame about the rest of the game but good work i guess
Go back to the test server forums and troll there.
All this can be applied to nullsec and hopefuly will increase the amount of players that can be online and playing at once before the node croaks. Especially if ccp increases the total amount of servers that are seeded in nullsec.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
lythos miralbar
4 wing Dara Cothrom
|
Posted - 2010.08.25 19:43:00 -
[31]
Edited by: lythos miralbar on 25/08/2010 19:43:41
Quote: Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Just a random sugestion here, why not use SSD (solid state) drives? most have 1.5 million days MTBF, are MASSIVELY faster than old style HDD's and with garbage colection routines built in they dont degrade over time any more (even if the OS does not support TRIM).
Whats more, lifespans have been significantly extended with new memory chips with high numbers of write cycles supported.. on a 32GB drive (100gb + is more common), you can write 100GB of data every day for 8 years before the drive starts to wear out.
Just a random sugestion
I'm starting to roll them out throghout our datacenter, they are great IMVHO.
|
Erling Skakke
|
Posted - 2010.08.25 19:45:00 -
[32]
Better than ever. 1290 pilots in system. All command were immediately executed. I tried: dock, undock, warpto, browse some prices.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 19:49:00 -
[33]
Originally by: lythos miralbar Edited by: lythos miralbar on 25/08/2010 19:43:41
Quote: Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Just a random sugestion here, why not use SSD (solid state) drives? most have 1.5 million days MTBF, are MASSIVELY faster than old style HDD's and with garbage colection routines built in they dont degrade over time any more (even if the OS does not support TRIM).
Whats more, lifespans have been significantly extended with new memory chips with high numbers of write cycles supported.. on a 32GB drive (100gb + is more common), you can write 100GB of data every day for 8 years before the drive starts to wear out.
Just a random sugestion
I'm starting to roll them out throghout our datacenter, they are great IMVHO.
They already use ssd tech for the database. The drives on the actual nodes themselves are probably only for boot up and low level system stuff. Though who knows it might make some sort of diffence so couldn't hurt to buy one and stick it on the jita node since it seems to be the new 'live' lag test bed.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
CCP Yokai
|
Posted - 2010.08.25 19:57:00 -
[34]
I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
|
|
Textret
Gallente The Executioners Mongoose Legion
|
Posted - 2010.08.25 19:59:00 -
[35]
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
We will require BOTH pictures.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 20:00:00 -
[36]
Contact the head of Red vs Blue and ask them to stage a fleet fight in jita. That should give you plenty of additional data to chomp on.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
CCP Yokai
|
Posted - 2010.08.25 20:03:00 -
[37]
Originally by: QioX Im wondering why is it only 1 quad core cpu and not 2 ? also why are you using a 10k rpm sas drive when these days most are using 15k rpm drives ? both of these would improve things alot more than just 1 cpu and an older slower drive
SSD's would not get even one more person in a fleet fight or Jita. The OS is there alone... no real exe's to speak of neededing the I/O there... so SSD's on a server with no IOPS... Just a waste.
Actually running a single CPU is Faster than running two CPU's when you are only using one core on one CPU. The CPU's talk via the QPI buss and create overhead and latency. For this test we want a clean benchmark... no other services are running save for Jita to prevent any overlap in serverload.
Hope that helps explain things a bit more.
|
|
Teck Nine
|
Posted - 2010.08.25 20:04:00 -
[38]
Edited by: Teck Nine on 25/08/2010 20:05:44 I thought they would have gone with AMD 6 core or 12 core server CPU rigs. I think the 6 core 4 CPU ports can suport up to or around 165GB of ram or at least use a 6 core CPU in the server rig.
I'm kinda disappointed in the new server specs tbh.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 20:09:00 -
[39]
Something else to think about for getting higher numbers is the fact that not very many people who are ingame are reading the forums. Maybe put out an in ingame message like the traffic advisorys around the jita area asking for players to move into jita or log on an alt if they have the time to help?
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
CCP Yokai
|
Posted - 2010.08.25 20:10:00 -
[40]
Originally by: Teck Nine Edited by: Teck Nine on 25/08/2010 20:05:44 I thought they would have gone with AMD 6 core or 12 core server CPU rigs. I think the 6 core 4 CPU ports can suport up to or around 165GB of ram or at least use a 6 core CPU in the server rig.
I'm kinda disappointed in the new server specs tbh.
Again... this is the baseline server, not the production system. This is our benchmark... just need to see where she tops out to know how this would scale on more cores, RAM, and alternate clock rates.
Thanks!
|
|
|
|
CCP Yokai
|
Posted - 2010.08.25 20:16:00 -
[41]
Originally by: Textret
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
We will require BOTH pictures.
Make it 2000 and I'll post both!
|
|
alittlebirdy
|
Posted - 2010.08.25 20:17:00 -
[42]
Originally by: CCP Navigator Thank you all for your feedback so far. Please keep it coming and let us know if your experience has improved, gotten worse or remained the same.
The war on lag is kicking up a gear and your reports are very important
Only took what 9 months...
Slower for me, however don't matter seeing as have been running on onboard with no shader support, I am sure you guys do not care on that front.
|
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.08.25 20:19:00 -
[43]
Originally by: Kephael
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
Why no 5677, you guys getting cheap?
Because IBM HS22 and IBM BladeCenter H could not accomodate CPUs with TDP of 130W. Unforutantely Xeon 5667 3.06GHz is the fastest Quad Core with 95W TDP - the Xeon 5677 3.46GHz you've mentioned has a TDP of 130W...
Anyway, with Intel Turbo Boost feature (Xeon 55xx/56xx) the most loaded CPU core could go up to 3.73GHz (from 3.06), which is still more that the current 3.33GHz dual-core-4/5-generations-older Intel Xeons...Having better IPC (instructions-per-cycle), memory throughput and lower latancy, this node could bring 10-20% better performance (at least).
Let's see what the "Control Group" benchmarking will show ... ___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Hykke
Free Imperial Vikings
|
Posted - 2010.08.25 20:19:00 -
[44]
Originally by: CCP Yokai [if] we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
He heee, I vote for the picture of ... no wait On the other hand ... DAMN that's a hard question
|
Emberday
|
Posted - 2010.08.25 20:27:00 -
[45]
Originally by: Kephael
Originally by: CCP Yokai
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
Why no 5677, you guys getting cheap?
35W more power per cpu for 400 extra MHz ? In a blade farm, no thanks. You can see that the spec was equally performance and power conscious. VLP memory is "Very Low Power", very being relative, but say 15-20%.
Glad to see the HS22 setup is coming along. Given the recent cabinet moves, this may allow a systematic replacement of targeted blades for performance while not being worried about cooling.
It will be nice to see if the "Fleet Fight" blades are next.
I logged in 2 Alts, seemed ok, but can't say very much about it ... I'm camping in "Nordic Jungle", north of the 46th parallel on a 24Kbps link. Surprising how many "fleet lag" behaviors I can simulate by starving the modem (load a webpage in the background) and how well a single client can otherwise run/fight/etc...
|
|
CCP Yokai
|
Posted - 2010.08.25 20:30:00 -
[46]
Exactly what he said :)
Also, the 130W CPU's are possible to use in the H series... just need upgraded power supplies. For now we only need to know what makes the most impact for each part of the server and this test will help.
;)
Originally by: Hawk TT
Originally by: Kephael
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
Why no 5677, you guys getting cheap?
Because IBM HS22 and IBM BladeCenter H could not accomodate CPUs with TDP of 130W. Unforutantely Xeon 5667 3.06GHz is the fastest Quad Core with 95W TDP - the Xeon 5677 3.46GHz you've mentioned has a TDP of 130W...
Anyway, with Intel Turbo Boost feature (Xeon 55xx/56xx) the most loaded CPU core could go up to 3.73GHz (from 3.06), which is still more that the current 3.33GHz dual-core-4/5-generations-older Intel Xeons...Having better IPC (instructions-per-cycle), memory throughput and lower latancy, this node could bring 10-20% better performance (at least).
Let's see what the "Control Group" benchmarking will show ...
|
|
Jita Alt666
|
Posted - 2010.08.25 20:32:00 -
[47]
Edited by: Jita Alt666 on 25/08/2010 20:33:29
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
Evidence that your off track with your client base. The only reason anyone goes to Jita is to buy and sell. You seem to thnk that because you've made a new node for Jita more people will go there. Why? To compound the lack of insight you suggest that people in jita should log on alts... ...the characters you see in Jita are the alts. Why would a player have more than 2 fully trained trading alts unless they are permament scammers who need to look clean?
In fact the whole thread is a repetition of that same theme.
|
Wildcarrd
|
Posted - 2010.08.25 20:34:00 -
[48]
Originally by: CCP Yokai
Originally by: QioX Im wondering why is it only 1 quad core cpu and not 2 ? also why are you using a 10k rpm sas drive when these days most are using 15k rpm drives ? both of these would improve things alot more than just 1 cpu and an older slower drive
SSD's would not get even one more person in a fleet fight or Jita. The OS is there alone... no real exe's to speak of neededing the I/O there... so SSD's on a server with no IOPS... Just a waste.
Actually running a single CPU is Faster than running two CPU's when you are only using one core on one CPU. The CPU's talk via the QPI buss and create overhead and latency. For this test we want a clean benchmark... no other services are running save for Jita to prevent any overlap in serverload.
Hope that helps explain things a bit more.
reading this makes me feel: 1) proud that I play a game where the devs have enough respect for the intellect of the playerbase to have the conversation at all, and,
2) like my head's gonna explode from reading too much geek-speak
|
Kahar Dex
|
Posted - 2010.08.25 20:43:00 -
[49]
TY, CCP, huge improvement : )
|
Grozen
Caldari Titan Core
|
Posted - 2010.08.25 20:47:00 -
[50]
I had zero lag and performance unlike before was godly.Before it took bout 1sec to process a buy of item now its instant and smooth.Very good machine you got there.I can only imagine if all systems run something like this maybe fleet fights would also notice the difference. knowledge is power |
|
Louis deGuerre
Gallente Amicus Morte Shock an Awe
|
Posted - 2010.08.25 20:53:00 -
[51]
Originally by: Bagehi
Originally by: Zeba
Originally by: Darod Zyree call that specs? I was expecting new hamsters, leaving thread disapointed! :p
Here is the box art for the new server hardware.
Zeba wins this thread.
This. So much this lol Sol: A microwarp drive? In a battleship? Are you insane? They arenĘt built for this! Clear Skies - The Movie
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 20:56:00 -
[52]
Originally by: Jita Alt666 Edited by: Jita Alt666 on 25/08/2010 20:33:29
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
Evidence that your off track with your client base. The only reason anyone goes to Jita is to buy and sell. You seem to thnk that because you've made a new node for Jita more people will go there. Why? To compound the lack of insight you suggest that people in jita should log on alts... ...the characters you see in Jita are the alts. Why would a player have more than 2 fully trained trading alts unless they are permament scammers who need to look clean?
In fact the whole thread is a repetition of that same theme.
He is not asking the people in jita to log in alts. He is asking for anyone to log in some alts and go to jita to stress the new hardware. Hence the message asking for players to head over to jita on the news items when you get to the character screen.
In fact your whole poast is a repetition of the same troll you always troll.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
Joshua Cy
|
Posted - 2010.08.25 21:05:00 -
[53]
Just logged in my alt at Jita. Only 1300 there, but buying was smooth. Undocking was smooth. To get some more action I undocked in a Primus, but no one would gank it. So I undocked in a shuttle and randomly targeted and bumped large dangerous looking ships. Noticed no lag in targeting or movement.
I did see one bug: someone targeted me back and the yellow square about the ship is space was solid, not blinking. ( It blinked on the overview).
|
Jita Alt666
|
Posted - 2010.08.25 21:07:00 -
[54]
Originally by: Zeba
Originally by: Jita Alt666 Edited by: Jita Alt666 on 25/08/2010 20:33:29
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
Evidence that your off track with your client base. The only reason anyone goes to Jita is to buy and sell. You seem to thnk that because you've made a new node for Jita more people will go there. Why? To compound the lack of insight you suggest that people in jita should log on alts... ...the characters you see in Jita are the alts. Why would a player have more than 2 fully trained trading alts unless they are permament scammers who need to look clean?
In fact the whole thread is a repetition of that same theme.
He is not asking the people in jita to log in alts. He is asking for anyone to log in some alts and go to jita to stress the new hardware. Hence the message asking for players to head over to jita on the news items when you get to the character screen.
In fact your whole poast is a repetition of the same troll you always troll.
It was not a troll, it was direct feedback of the manner inwhich this testing is being conducted. Again CCP are asking for us to help them test their gear, and again ccp are expressing disatisfaction in the low number of paying subscribers who are willing to help them. I am simply saying that system wont get 1800-2000 players into Jita tonight. Jita is a system full of trading alts there to make ingame isk - why would I (an presumably many other subscribers) stop running my: r&d grind on an alt 30ish jumps out, pos erection on an alt 45ish jumps out, station reinforcing alt 35ish jumps out, to help test the load on Jita.
The best way to deal with trolls is to ignore them, but thankyou for pointing out that my post was unclear I hope I have clarified things in this post.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 21:17:00 -
[55]
Originally by: Jita Alt666 It was not a troll, it was direct feedback of the manner inwhich this testing is being conducted. Again CCP are asking for us to help them test their gear, and again ccp are expressing disatisfaction in the low number of paying subscribers who are willing to help them. I am simply saying that system wont get 1800-2000 players into Jita tonight. Jita is a system full of trading alts there to make ingame isk - why would I (an presumably many other subscribers) stop running my: r&d grind on an alt 30ish jumps out, pos erection on an alt 45ish jumps out, station reinforcing alt 35ish jumps out, to help test the load on Jita.
The best way to deal with trolls is to ignore them, but thankyou for pointing out that my post was unclear I hope I have clarified things in this post.
Feedback requires something helpful to be stated or a problem to be defined. You had nothing helpful to add and only pointed out the obvious which I promptly cleard up. Plus you were quite negative in your tone. Hence: troll. And no I will not ignore you.
As far as the last part of the poast well then don't arse yourself to help if its going to get in the way of your game. But do you speak for every single player with alts? Maybe there are a few with inactive alts they wouldn't mind to stick in a shuttle and shuffle off to jita on autopilot as they play the game with their main?
Always with the negative waves man. Have a little faith baby.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
Miss Wormwood
|
Posted - 2010.08.25 21:24:00 -
[56]
Popped to Jita to make up the numbers because I'm not doing anything else in particular.
Everything seems to be working smoothly, down to about 1170 bods now though.
|
Julie Vosgerichian
|
Posted - 2010.08.25 21:27:00 -
[57]
actually, what makes your post a troll is that you assume people do not have jita alts. pretty much everyone i know who lives in empire has a jita alt (alt living in jita) in order to check prices, and many of those people have 1 jita alt per account. GG
|
Jita Alt666
|
Posted - 2010.08.25 21:29:00 -
[58]
Originally by: Miss Wormwood Popped to Jita to make up the numbers because I'm not doing anything else in particular.
Everything seems to be working smoothly, down to about 1170 bods now though.
Expect Jita numbers to drop slowly over the next 13 hours. Currently still around 1140
|
Bagehi
Association of Commonwealth Enterprises R.A.G.E
|
Posted - 2010.08.25 21:36:00 -
[59]
Down to 1100 in local. It is running smooth as silk though.
This signature is useless, but it is red.
|
Jita Alt666
|
Posted - 2010.08.25 21:37:00 -
[60]
Originally by: Zeba
Originally by: Jita Alt666 It was not a troll, it was direct feedback of the manner inwhich this testing is being conducted. Again CCP are asking for us to help them test their gear, and again ccp are expressing disatisfaction in the low number of paying subscribers who are willing to help them. I am simply saying that system wont get 1800-2000 players into Jita tonight. Jita is a system full of trading alts there to make ingame isk - why would I (an presumably many other subscribers) stop running my: r&d grind on an alt 30ish jumps out, pos erection on an alt 45ish jumps out, station reinforcing alt 35ish jumps out, to help test the load on Jita.
The best way to deal with trolls is to ignore them, but thankyou for pointing out that my post was unclear I hope I have clarified things in this post.
Feedback requires something helpful to be stated or a problem to be defined. You had nothing helpful to add and only pointed out the obvious which I promptly cleard up. Plus you were quite negative in your tone. Hence: troll. And no I will not ignore you.
As far as the last part of the poast well then don't arse yourself to help if its going to get in the way of your game. But do you speak for every single player with alts? Maybe there are a few with inactive alts they wouldn't mind to stick in a shuttle and shuffle off to jita on autopilot as they play the game with their main?
Always with the negative waves man. Have a little faith baby.
Zeba, I feel we are getting off topic and derailing the thread. Which is not constructive feedback at all. However you raise a couple points that I feel need addressing.
1. If the problem I raised is obvious, why does CCP repeat it on a weekly basis. Do the same things and you get the same results. Indeed with our combined posting it is becoming more obvious but at the expense of the overall thread.
2. To test the node capacity you suggest "a few" players with in active alts could autopilot over. Jita has just dropped under 1100, by "a few" do you mean 1000 players auto-pilot over to Jita?
To test the node's capacity CCP need to engage with their client base, when your most vocal dissenters back you up or disappear, you know things are working well.
|
|
|
CCP Yokai
|
Posted - 2010.08.25 21:45:00 -
[61]
Just wanted to quickly say thanks!
We got over 1300 today... not bad for middle of the week. The hardware will stay there for now so there is plenty of time for testing.
Also, I wanted to point out that while I may have asked for more in people in Jita, it's all in good fun and in no way is anyone upset. I appreciate those who had the time and interest in logging on and respect those that couldn't or wouldn't. I hope that the next call for numbers will find you and yours available and interested.
We'll be doing some more testing shortly and will post in forums again when we are ready to try and break the record again.
CCP Yokai
|
|
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.08.25 21:48:00 -
[62]
Yep, but CCP Yokai was absolutely right and I was wrong for two things (a couple of weeks of holiday and IBM came up with new HW releases):
1. Since the begining of August 2010 IBM has released HS22 7870-F3x w/ Xeon 5677 3.46/3.73GHz 130W TDP. It is supported by the current BladeCenter H, but with the current power supplies you can't populate all blade slots. IBM has new 2980W PSUs with 95% (!!!) efficiency, which is much better than the current 2900W PSUs (82+% efficiency). New Fan trays would be needed as well.
2. The currently tested Xeon 5667 is 3.06/3.46GHz part (with Turbo Boost off/on) @ 95W TDP, which is ONLY slight improvement in terms of CPU frequency over the current 3.33GHz Wolfdale Xeon 5260...(3.46GHz vs. 3.33GHz). If the new Westmere CPU microarchitecture proves to be much better than the aging Wolfdale, CCP could justify upgrading the PSUs, FANS, Blades etc. to accomodate Xeon 5677...
One more thing - 48GB RAM with 1 CPU means you have used 8GB 1333MHz 1.5V VLP RDIMMs (49Y1431), right? Actually this is the only supported configuration that would allow memory running @ 1333MHz with 2 RDIMMs per channel (3x2x8=48), limited to only 1 CPU and that's for HS22 with 12 DIMM slots.
PLEASE, CCP Yokai, share some comparison graphs once you finish the test ;-). This would show the benefits of the microarchitecture & platform advances of Westmere over Wolfadlde for HPC applications like EVE :) That's how I found EVE - reading IBM whitepapers for HPC solutions for the gaming industry... :)
Originally by: CCP Yokai Exactly what he said :)
Also, the 130W CPU's are possible to use in the H series... just need upgraded power supplies. For now we only need to know what makes the most impact for each part of the server and this test will help.
;)
Originally by: Hawk TT
Because IBM HS22 and IBM BladeCenter H could not accomodate CPUs with TDP of 130W. Unforutantely Xeon 5667 3.06GHz is the fastest Quad Core with 95W TDP - the Xeon 5677 3.46GHz you've mentioned has a TDP of 130W...
Anyway, with Intel Turbo Boost feature (Xeon 55xx/56xx) the most loaded CPU core could go up to 3.73GHz (from 3.06), which is still more that the current 3.33GHz dual-core-4/5-generations-older Intel Xeons...Having better IPC (instructions-per-cycle), memory throughput and lower latancy, this node could bring 10-20% better performance (at least).
Let's see what the "Control Group" benchmarking will show ...
___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 21:51:00 -
[63]
Originally by: Jita Alt666 Zeba, I feel we are getting off topic and derailing the thread. Which is not constructive feedback at all. However you raise a couple points that I feel need addressing.
1. If the problem I raised is obvious, why does CCP repeat it on a weekly basis. Do the same things and you get the same results. Indeed with our combined posting it is becoming more obvious but at the expense of the overall thread.
2. To test the node capacity you suggest "a few" players with in active alts could autopilot over. Jita has just dropped under 1100, by "a few" do you mean 1000 players auto-pilot over to Jita?
To test the node's capacity CCP need to engage with their client base, when your most vocal dissenters back you up or disappear, you know things are working well.
Then make a suggestion as to how ccp can get more players into jita to test the new hardware. I have already made two suggestions and all you have done is snipe at ccp for 'failing' without any other content. I mean obviously ccp wants to get as many people into jita as possible in a short a time as possible so give it some time to perculate through the playerbase. Even an alliacne leader can't simply snap his fingers and get a 1000 ship blob formed on a whim. Besides if they don't get it today they will tommorrow or the next day as more people log in and read the forums or the news items. Maybe that offer for the dev dancing needs to be extended?
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
ShadowMaster
Gallente
|
Posted - 2010.08.25 22:00:00 -
[64]
Originally by: CCP Yokai Just wanted to quickly say thanks!
We got over 1300 today... not bad for middle of the week. The hardware will stay there for now so there is plenty of time for testing.
Also, I wanted to point out that while I may have asked for more in people in Jita, it's all in good fun and in no way is anyone upset. I appreciate those who had the time and interest in logging on and respect those that couldn't or wouldn't. I hope that the next call for numbers will find you and yours available and interested.
We'll be doing some more testing shortly and will post in forums again when we are ready to try and break the record again.
CCP Yokai
No, NOT OK! I want those pictures and videos! |
Scyyy
|
Posted - 2010.08.25 22:02:00 -
[65]
Originally by: ghosttr
Originally by: CCP Yokai Seriously... only 1200... thought you guys had alts or something.
Sorry 18 17 months, you know
"Better to remain silent and be thought a fool than to speak out and remove all doubt."
|
Jita Alt666
|
Posted - 2010.08.25 22:13:00 -
[66]
Originally by: Zeba
Originally by: Jita Alt666 Zeba, I feel we are getting off topic and derailing the thread. Which is not constructive feedback at all. However you raise a couple points that I feel need addressing.
1. If the problem I raised is obvious, why does CCP repeat it on a weekly basis. Do the same things and you get the same results. Indeed with our combined posting it is becoming more obvious but at the expense of the overall thread.
2. To test the node capacity you suggest "a few" players with in active alts could autopilot over. Jita has just dropped under 1100, by "a few" do you mean 1000 players auto-pilot over to Jita?
To test the node's capacity CCP need to engage with their client base, when your most vocal dissenters back you up or disappear, you know things are working well.
Then make a suggestion as to how ccp can get more players into jita to test the new hardware. I have already made two suggestions and all you have done is snipe at ccp for 'failing' without any other content. I mean obviously ccp wants to get as many people into jita as possible in a short a time as possible so give it some time to perculate through the playerbase. Even an alliacne leader can't simply snap his fingers and get a 1000 ship blob formed on a whim. Besides if they don't get it today they will tommorrow or the next day as more people log in and read the forums or the news items. Maybe that offer for the dev dancing needs to be extended?
I have made numerous suggestions regarding the long term development of eve online and how CCP could re-engage with its playerbase, in the appropriate forums/threads. Your correct that an alliance leader cant create a 1000man blob with a snap of the fingers. However when an alliance leader wishes to create a 1000man blob he/she informs the players well inadvance of the need to be present. Most of the alliances I have been involved with, have leaders who can - combined with other fleets - rustle up 1000 players in under 5 hours.
Using existing storylines and a bit of creativity it shouldnt be pretty difficult to rustle numbers: silly example: Utilise recent Sansha Incursion, "There are rumours of a sansha manipulated wormhole being targeted to develop in Jita. The last of these manipulated wormholes resulted in a sizable Sansha force emerging in high security space and momentarely overwhelming concord forces. Inside sources indicate the wormhole generation attempt will be between 1700 - 2300 on Wednesday 25 August 114. The same sources indicate that the Sansha Forces will be heavily dependent on their own faction ships and modules. Concord are appealing for capsuleer help to limit this attempted incusrion."
|
Ishina Fel
Caldari Terra Incognita Black Star Alliance
|
Posted - 2010.08.25 22:23:00 -
[67]
Holy cow what did you do to Jita
Admittedly, in the last year or so I have never had the huge lag that was there when I first joined in 2006 or so, but I have gotten pretty used to waiting 2-3 seconds for items to be moved from hangar to cargo (or to be fitted) whenever the system passes 1000 people. It doesn't really hurt or annoy if you only come there occasionally, but it's a notable difference from 'normal' systems.
Just now though? 1160 in local, items are moved as soon as I release the mouse button. Hot damn that new server you got is sexy alright
And yes, let's see some CPU load graphs for primetime hours. The devblogs delivered graphs for the old hardware, now we need some from the new rig to compare!
Signature? What signature? |
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 22:29:00 -
[68]
Originally by: Jita Alt666 I have made numerous suggestions regarding the long term development of eve online and how CCP could re-engage with its playerbase, in the appropriate forums/threads. Your correct that an alliance leader cant create a 1000man blob with a snap of the fingers. However when an alliance leader wishes to create a 1000man blob he/she informs the players well inadvance of the need to be present. Most of the alliances I have been involved with, have leaders who can - combined with other fleets - rustle up 1000 players in under 5 hours.
Using existing storylines and a bit of creativity it shouldnt be pretty difficult to rustle numbers: silly example: Utilise recent Sansha Incursion, "There are rumours of a sansha manipulated wormhole being targeted to develop in Jita. The last of these manipulated wormholes resulted in a sizable Sansha force emerging in high security space and momentarely overwhelming concord forces. Inside sources indicate the wormhole generation attempt will be between 1700 - 2300 on Wednesday 25 August 114. The same sources indicate that the Sansha Forces will be heavily dependent on their own faction ships and modules. Concord are appealing for capsuleer help to limit this attempted incusrion."
Other threads are not this thread. And thanks for the ontopic positive post.
anyhoo..
If this new hardware is any indication of the servers potential new performance levels especially if this is the 'low end' of the new stuff it looks like eve will be primed and ready for the influx of new players when incarna hits along with hopefully a real marketing campaign in all the major rags and other gamer sites. Who knows, maybe this is also part of the final pieces of the puzzle to fix the fleet lag issue. Then it would be a truely glorious expansion.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
Jita Alt666
|
Posted - 2010.08.25 22:40:00 -
[69]
Originally by: Zeba Other threads are not this thread. And thanks for the ontopic positive post.
anyhoo..
If this new hardware is any indication of the servers potential new performance levels especially if this is the 'low end' of the new stuff it looks like eve will be primed and ready for the influx of new players when incarna hits along with hopefully a real marketing campaign in all the major rags and other gamer sites. Who knows, maybe this is also part of the final pieces of the puzzle to fix the fleet lag issue. Then it would be a truely glorious expansion.
1. By responding to me you have effectively derailed this thread. Thats why suggestions go in suggestion threads and feedabck goes in feedback threads 2. You made no reference to my suggestion. This is also true of CCP and the myriad of suggestions people put forward. Thats why people fail to provide suggestions. 3. To get to the final pieces of the lag puzzle they need to find the first pieces. 4. If CCP dont fix the game as it is before the release Incarna, some one will figure out a way to make it rain *****es.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.25 22:50:00 -
[70]
Originally by: Jita Alt666 1. By responding to me you have effectively derailed this thread. Thats why suggestions go in suggestion threads and feedabck goes in feedback threads 2. You made no reference to my suggestion. This is also true of CCP and the myriad of suggestions people put forward. Thats why people fail to provide suggestions. 3. To get to the final pieces of the lag puzzle they need to find the first pieces. 4. If CCP dont fix the game as it is before the release Incarna, some one will figure out a way to make it rain *****es.
hoo boy..
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: CCP Adida Moved from missions and complexes. All other game discussions are only aloud in OOP.
|
|
Obsidian Hawk
RONA Legion
|
Posted - 2010.08.25 22:53:00 -
[71]
CCP Yokai -
Will we get a spiffy dev blog with graphs and pictures showing the test results.
And, could you explain the hardware more, im not a server person, im a geologist actually, but anyway, I would like a more indepth explanation of the new hardware.
|
Terrestes
Di-Tron Heavy Industries Atlas Alliance
|
Posted - 2010.08.25 23:19:00 -
[72]
Originally by: CCP Yokai Edited by: CCP Yokai on 25/08/2010 16:20:27 This server is for benchmarking new chipset, Cache and RAM on Tranquility and is not the final or completed spec we have planned. Consider this the Control Group for our testing.
Also, while we are doing tests on Jita, we are also going to be profiling fleet fights and all kinds of other node/service types on the cluster.
Once we get a good load metric on none type we'll be moving the server to test other things.
Thanks in advance for helping get the load we need to see the performance impacts.
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
Non-optimized and no affinity set. Intel tweaks will come after some counter collections get us stats.
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
Awesome, keep up the good work! Also hi
|
LyLak
|
Posted - 2010.08.26 00:38:00 -
[73]
im having issues i get to new caldari and everything freezes with autopilot heading to jita idk if its the new hardware or my account is stuck if u could please Send me an ingame message GMs id really appreciate it
|
RiskyFrisky
|
Posted - 2010.08.26 00:59:00 -
[74]
Honestly, I've noticed no difference.
|
Deadly Hobbitses
Caldari State War Academy
|
Posted - 2010.08.26 01:07:00 -
[75]
Originally by: CCP Yokai I hope that the next call for numbers will find you and yours available and interested.
Why not just Dev spawn a few 0.0 style Ice belts and 30 or 40 insta-repopulating Sanctums and let the isk farmers test the node for you?
|
Chris Camcer
|
Posted - 2010.08.26 01:10:00 -
[76]
Ships were stuttering back and forward a little today when i undocked. When flying with MWD, cap didn't get removed until half way through cycle which is abnormal.
Don't know if it affects market, but i could see some oddities there. Saw an increase in lag when purchasing items; items would appear about 20 seconds after i purchase it. Still have the incorrect price bug when selling many items from hanger on sell orders (market shows one item but prices for another).
|
Minchurra
Caldari Feudum Chalybis The Spire Collective
|
Posted - 2010.08.26 01:22:00 -
[77]
Originally by: Deadly Hobbitses
Originally by: CCP Yokai I hope that the next call for numbers will find you and yours available and interested.
Why not just Dev spawn a few 0.0 style Ice belts and 30 or 40 insta-repopulating Sanctums and let the isk farmers test the node for you?
Because if they ignore them they'll go away.
|
Lana Torrin
Minmatar Republic Military Skool HYDRA RELOADED
|
Posted - 2010.08.26 01:24:00 -
[78]
Want more people in jita? Put the belts back.. Ice miner bots alone would add another thousand clients..
Replying to threads without reading them since 2008 |
Tera Mcbitchypants
|
Posted - 2010.08.26 01:52:00 -
[79]
I would say this thing should be able to handle Jita with out too much issue. I figure its a quad core with hyper threading this allows a total of 8 threads to run at once and gobs of cache. So it should be able to handle all your little quires with out too much issue. Granted I dont know how the game works on the back end, but if this thing wont cut it might want to think about clustering some ATI cards to abuse there ******ed high number of Stream Processing Units. Simple operations but thats all this game really is. NOw as far as the harddrives 2 does kinda leave some questons are they in a RAID 1 set up or a RAID 0. If there in a RAID 0 what is gonna happen when one fails.
|
Drew121789
Caldari Royal Black Watch Highlanders
|
Posted - 2010.08.26 02:40:00 -
[80]
Originally by: CCP Yokai
As alwaysą tell us what you thinką if you felt itą if you want cookiesą etc.
I want cookies!!! Soft chololate chip please. And will they be delievered ingame to me or via real life mail?
|
|
lynch piin
|
Posted - 2010.08.26 02:55:00 -
[81]
The guest list loads alot faster than before and undocking has no lag problems. Using the market is better, but is still delayed be a few seconds. But all in all great job CCP, keep on fighting and we will keep bringing you logistics on the enemy.
|
Nemesis Factor
Caldari Telanus' Reach
|
Posted - 2010.08.26 02:59:00 -
[82]
Edited by: Nemesis Factor on 26/08/2010 03:01:34 You guys should buy a Sunfire for each system.
And I mean the big ones, not those little 1U/2U guys. ==================== ~/~ Sultan of Buruni |
Motsumoto
|
Posted - 2010.08.26 04:06:00 -
[83]
Worse. The delay in obtaining the list of market orders, considerable lag between purchase and the appearance of the item in a hangar.
|
Crusher Deac
Southern Cross Incorporated Flying Dangerous
|
Posted - 2010.08.26 04:31:00 -
[84]
I'll take a cookie!
What happens when we get over 9000? [FIGL] Flying Dangerous Alliance |
SFX Bladerunner
Minmatar Black Serpent Technologies R.A.G.E
|
Posted - 2010.08.26 04:40:00 -
[85]
Originally by: Ix Forres
Originally by: Rowbin Hod
AFAIK, they're only able to put a node on one core (multiple nodes fit on one core, but a single node won't fit on multiple cores) and as they're trying to test the Jita location node, they'll be looking at the performance of a single core. Of course, they may shunt other Jita nodes for other bits of the work onto the other cores, but this isn't a case where more cores = more performance.
Yup - that's what I was getting at. More cores may actually = less performance if you start bottlenecking RAM and network resources.
Aye, along with PSU, mobo throughput and heat generation that seems like more than enough reason for a LIVE GAMING SERVER to prefer the stability of a multi-system over a multi-cpu solution.. Also moar racks = moar e-peen.
Seriously, I have wet dreams about building a server park and adding racks on top of racks in a nice rackmount case over a period of x years and then one day finally having the entire rackmount filled up with servers I mean, just imagine all the networkcables... oh my I just wet myself __________________________________________________
History is much like an endless waltz, the three beats of war, peace and revolution continue on forever.. |
POS Trader
|
Posted - 2010.08.26 05:15:00 -
[86]
Originally by: Zeba Then make a suggestion as to how ccp can get more players into jita to test the new hardware.
UPDATE players SET location='Jita';
PS. /me already in Jita!
|
LyLak
|
Posted - 2010.08.26 05:15:00 -
[87]
I found that my ship got stuck coming into jita from new caldari had to relog and try again, 2nd time is always a charm.
but i noticed alot of chat lag. takes about 3 or 4 seconds for it to show up.
my FPS was higher then it normally is its usually around the 45 range in jita it was at 75 so yea. Mentioning this because i had some issues with jita in Dominion Exspansion while back when my computer wasnt too great. Id disconnect every like hour lol. my graphics card took a frying so i just bought a new computer running eve like nothing. Jita should have like 10 Ice Fields Amarr,Minmitar,Gallente and Caldari types
|
Teck Nine
|
Posted - 2010.08.26 05:45:00 -
[88]
Tried jita out for the 3rd time and once again still no lag even when you guys were using the old stuff. Either my internet connection is fast to help counter the lag problem or my comp's specs are beyond what the average eve online player has.
IMO I think its because the rig I built is a beast. I sunk over $400 just for the ram alone and using a 6 core black edition not o/c and running at 3.2GHz.
I have My gaming rig online, 1 comp downloading stuff most the time and playing my P3 online with no problems. Alot of people who play this game is using ADSL or Dile-up
|
Damon Eclipse
Absolutum Dominium de Stellae
|
Posted - 2010.08.26 06:02:00 -
[89]
Originally by: Teck Nine Tried jita out for the 3rd time and once again still no lag even when you guys were using the old stuff. Either my internet connection is fast to help counter the lag problem or my comp's specs are beyond what the average eve online player has.
IMO I think its because the rig I built is a beast. I sunk over $400 just for the ram alone and using a 6 core black edition not o/c and running at 3.2GHz.
I have My gaming rig online, 1 comp downloading stuff most the time and playing my P3 online with no problems. Alot of people who play this game is using ADSL or Dile-up
Report to CCP and become a <system> node.
|
Ishina Fel
Caldari Terra Incognita Black Star Alliance
|
Posted - 2010.08.26 06:09:00 -
[90]
Edited by: Ishina Fel on 26/08/2010 06:09:39
Originally by: Teck Nine IMO I think its because the rig I built is a beast. I sunk over $400 just for the ram alone and using a 6 core black edition not o/c and running at 3.2GHz.
Except that EVE barely even uses two cores (and nearly all of its load is on the first), and rarely goes over 50% load even on an age-old 2GHz CPU. Also, the performance increase in games from better RAM compared to the baseline modules is in the low single digit percentages - the same investment that'll get you 5% more gaming performance through RAM could give you 100% more if you invested it in your video card.
Of course, if you were to run half a dozen EVE clients in parallel, then yes, your system will be able to flex its muscles then. But for a single client? You won't see a difference to a 4 year old dualcore as long as the video card is identical.
Signature? What signature? |
|
Rakessh
Antares Shipyards Circle-Of-Two
|
Posted - 2010.08.26 06:24:00 -
[91]
Originally by: Damon Eclipse
Originally by: Teck Nine Tried jita out for the 3rd time and once again still no lag even when you guys were using the old stuff. Either my internet connection is fast to help counter the lag problem or my comp's specs are beyond what the average eve online player has.
IMO I think its because the rig I built is a beast. I sunk over $400 just for the ram alone and using a 6 core black edition not o/c and running at 3.2GHz.
I have My gaming rig online, 1 comp downloading stuff most the time and playing my P3 online with no problems. Alot of people who play this game is using ADSL or Dile-up
Report to CCP and become a <system> node.
Actually entertained this idea for about 5 minutes with an mmo tech I was creating. Not gonna work very well. In theory if there was no latency and no bandwidth limits across the Internet, having the client be a node could work great, but even then you have serious security issues with this setup.
Which of course is unfortunate or CCP could've tapped into a lot of nice customer hardware for running their game ;) Free cpu cycles 4tw
CEO Arachnea Phoenix Battalion |
Rasz Lin
Caldari Racketeers
|
Posted - 2010.08.26 06:32:00 -
[92]
I think DEV blog mentioned something about moving markettransactions to another separate node so thats why market feels snappier.
Checking prices is super fast, selling/buying also, but items appear with a noticable 1-2 second delay. Moving items = 1 second delay (was 3-5 seconds so it seems better).
flying around (700 players online in Jita) wasnt great tho, a lot of jerking back and forth, almost like you tried to optimize something and it backfired?
Undocking is super awesome, no more 5 seconds with black screen.
|
Abdiel Kavash
Caldari Paladin Order Fidelas Constans
|
Posted - 2010.08.26 06:52:00 -
[93]
One word: JITAFEST.
Jita has been working just fine for me before - admittedly I don't spend much time there, but even during the busier times (Sunday EVEnings) I had no problems doing shopping, undocking, moving around and that. I don't think I've ever been in a fight in Jita, but what's the point anyway.
How about keeping the server and using it for the nodes that are petitioned for reinforcement instead? Which would you rather have: "EVE online, with epic battles where YOU can fight hundreds of other people", or "EVE Online, where YOU can... buy and sell stuff"? ___________ EVE is dying! Now for real! |
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.08.26 07:02:00 -
[94]
Currently 735 in local in Jita: 0 Market lag (orders, price history stats for 1 year, bought items appear instantly) 0 EveMail lag 0 Chat lag 0 Undock / Doc lag
Let's see Jita on Sunday night
If my client specs matter: i7 Core 3.6GHz, 12GB 1866MHz DDR3, Radeon 5870 Toxic, Windows 7 x64, 100Mbps FO Internet connection, 8 hops to EVE cluster, 45-55ms network latency. ___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Spatula King
|
Posted - 2010.08.26 07:21:00 -
[95]
If they use an AMD, it would work better. The 1090Ts are way faster than any Intel chip, and then you have the Opterons which are even faster.
|
Katrina Bekers
Gallente Mia Corp
|
Posted - 2010.08.26 07:31:00 -
[96]
I'd love to compare the benchmarks now and when the affinity with a single core is set.
Single thread apps love to be bound to a single core. The context switch that may happen when hopping cores could really ruin your day. --- Kat |
Nanofoegan
|
Posted - 2010.08.26 07:43:00 -
[97]
can we get one of those servers for 0.0 as well :) we are kinda tired of lagging in fights with 50+ pilots. please do something do damn it!
|
Guilty Man
Minmatar Guilty People
|
Posted - 2010.08.26 08:02:00 -
[98]
Originally by: CCP Yokai Just wanted to quickly say thanks! Also, I wanted to point out that while I may have asked for more in people in Jita, it's all in good fun and in no way is anyone upset. I appreciate those who had the time and interest in logging on and respect those that couldn't or wouldn't. I hope that the next call for numbers will find you and yours available and interested.
You better make an announcement before the weekend, I will surely come and move all my alts to Jita, so many will do if they see the call.
|
Shingorash
Caldari Final Agony B A N E
|
Posted - 2010.08.26 08:45:00 -
[99]
Why use SATA drives anyway, you should be using RAID 5 SCSI drives. Bearing in mind the cluster with need fast response I/O on the drives SATA just wont cut it... --------------------------------------------------
I'm in your space, eating your ships! |
MrBreaker
Amarr Subatomic Research
|
Posted - 2010.08.26 08:46:00 -
[100]
I had no idea this testing was happening last night, or there had been an upgrade, but had to go to Jita to pick up some skills.
I did notice it was slightly faster.
|
|
Claire Voyant
|
Posted - 2010.08.26 08:48:00 -
[101]
Seems very simple to me. Leave the new hardware in through the weekend and don't put any limit on the number of players in Jita. You will soon get a "Crash Jita" movement going with thousands of alts orbiting the station. Why? Because it's Jita and because it is for a good cause. Once you have your data, reestablish the population limit (hopefully at a higher level) and tell everyone to go home. Just don't expect much to happen before the weekend.
|
ArchenTheGreat
Caldari Rapid Pod Transport
|
Posted - 2010.08.26 09:11:00 -
[102]
Originally by: CCP Yokai
We got over 1300 today... not bad for middle of the week. The hardware will stay there for now so there is plenty of time for testing.
Put a dev in Titan outside Jita4 and local will go up immediately. Just make sure it has lots of HP and drops good loot.
|
Mashie Saldana
BFG Tech
|
Posted - 2010.08.26 10:42:00 -
[103]
Originally by: Shingorash Why use SATA drives anyway, you should be using RAID 5 SCSI drives. Bearing in mind the cluster with need fast response I/O on the drives SATA just wont cut it...
That would be a waste as only the OS use the local HDD. All I/O is done to local RAM and remotely to the DB servers with the RAMSAN behind them.
|
Dan Wilkie
Gallente British Buccaneers The Worst Alliance in Eve
|
Posted - 2010.08.26 10:49:00 -
[104]
Originally by: Shingorash Why use SATA drives anyway, you should be using RAID 5 SCSI drives. Bearing in mind the cluster with need fast response I/O on the drives SATA just wont cut it...
Um, it's not SATA anyway? It's SAS (Serial Attached SCSI) which is pretty standard for direct attached drives these days...
As for the single drive only, as someobdoy commented further down it's purely for system files, DP ops are on the SAN iirc
|
Shingorash
Caldari Final Agony B A N E
|
Posted - 2010.08.26 10:52:00 -
[105]
My mistake didnt see that bit! --------------------------------------------------
I'm in your space, eating your ships! |
omgdutch2005
Gallente Advanced Planetary Exports Intergalactic Exports Group
|
Posted - 2010.08.26 11:12:00 -
[106]
ccp clearly said it is a serve r/w iops aka NOT much hdd activity, heck i think even a 40GB 2mb cache IDE disk would then suffice (might be over doing it but you get the point)
|
ceaon
|
Posted - 2010.08.26 11:21:00 -
[107]
CCP Yokai put a ramsan 20 and 2 gtx295 on sli so n00bs can be happy whit teh spec LOL
Originally by: CCP Adida The male thread was locked because the discussion turned into transsexuals and man boobs.
|
Darod Zyree
Gallente Zyree Holding
|
Posted - 2010.08.26 11:51:00 -
[108]
Post pic of eve server(s) -Darod- |
Koga Bisaya
|
Posted - 2010.08.26 12:04:00 -
[109]
I'm shocked and astonished but - yes, my corp had small gang battles with our wartargets in jita today that went alright! Now if the rest of the galaxy could become as lag free, I can go back to PVPing worrying only about losing my ship to other players, instead of losing my ship to lag
|
HellGate fr
Yarrbear Inc.
|
Posted - 2010.08.26 13:21:00 -
[110]
Edited by: HellGate fr on 26/08/2010 13:21:23
Originally by: ArchenTheGreat
Originally by: CCP Yokai
We got over 1300 today... not bad for middle of the week. The hardware will stay there for now so there is plenty of time for testing.
Put a dev in Titan outside Jita4 and local will go up immediately. Just make sure it has lots of HP and drops good loot.
That is a really good idea ! Put an Avatar like 200km from the station and people will come to be on the killmail!
|
|
Teck Nine
|
Posted - 2010.08.26 13:30:00 -
[111]
Edited by: Teck Nine on 26/08/2010 13:32:56 If they want to do a server load maybe they could do it during DT and do a test and could offer people ISK, items, ships, BPO, BOC and such that can grab the attention of alot of people or watch to see what time of day there's the most people online and do the testing then and offer people a reward(s) for helping out.
Sorry for posting 2 times but its easier to read this way because some people don't want to read a long thing.
|
Teck Nine
|
Posted - 2010.08.26 13:33:00 -
[112]
Originally by: Rakessh
Originally by: Damon Eclipse
Originally by: Teck Nine Tried jita out for the 3rd time and once again still no lag even when you guys were using the old stuff. Either my internet connection is fast to help counter the lag problem or my comp's specs are beyond what the average eve online player has.
IMO I think its because the rig I built is a beast. I sunk over $400 just for the ram alone and using a 6 core black edition not o/c and running at 3.2GHz.
I have My gaming rig online, 1 comp downloading stuff most the time and playing my P3 online with no problems. Alot of people who play this game is using ADSL or Dile-up
Report to CCP and become a <system> node.
Actually entertained this idea for about 5 minutes with an mmo tech I was creating. Not gonna work very well. In theory if there was no latency and no bandwidth limits across the Internet, having the client be a node could work great, but even then you have serious security issues with this setup.
Which of course is unfortunate or CCP could've tapped into a lot of nice customer hardware for running their game ;) Free cpu cycles 4tw
Haha actually there are games that do that. Doesn't the PS3 also use the same method for gaming is using People's PS3 as kinda a game server since in COD MW2 if a person's doesn't have a good enough connection they usaly get dropped and it looks for a new host.
I think once they get the game to a point where it would be costly to upgrade their server then they might start to make people's comp help take some of the load.
Originally by: Spatula King If they use an AMD, it would work better. The 1090Ts are way faster than any Intel chip, and then you have the Opterons which are even faster.
here's a rough list on the cost of AMD CPUs
1 - AMD 1090T6 Core 3.2GHz, 6MB Cache, 2000MHz (4000 MT/s ) FSB,Socket AM3, Processor would cost around $329.97
2 - AMD Opteron 2435 Six-Core Processor - 2.60GHz, Socket F, 6MB L3 Cache - $1300 3 - AMD Opteron 2431 Six-Core Processor - 2.40GHz, Socket F, 6MB L3 Cache - $900
4 - AMD Opteron 6172 Magny-Cours 2.1GHz 12 x 512KB L2 Cache 12MB L3 Cache Socket G34 115W 12-Core Server Processor - $1,070.00
With numbers 2 and 3 there's server boards that suport 2 to 4 CPUs and cau suport up to 165GB of ram and it also supports the new DDR3 type of ram.
For the 12 core there's boards that support 1 to 2 CPUs and can suport around 65GB of ram and it supports the new DDR3 ram.
So as you can see there are alot of options to make a powerful server rig that can deal with alot of load.
|
FireT
Gallente Royal Advanced Industries Imperial Crimson Legion
|
Posted - 2010.08.26 14:36:00 -
[113]
Thanbk you CCP and all your staff for your continued dedication to this great game.
|
Levitikon
Destructive Influence IT Alliance
|
Posted - 2010.08.26 14:41:00 -
[114]
Originally by: Jita Alt666
Utilise recent Sansha Incursion, "There are rumours of a sansha manipulated wormhole being targeted to develop in Jita. The last of these manipulated wormholes resulted in a sizable Sansha force emerging in high security space and momentarely overwhelming concord forces. Inside sources indicate the wormhole generation attempt will be between 1700 - 2300 on Wednesday 25 August 114. The same sources indicate that the Sansha Forces will be heavily dependent on their own faction ships and modules. Concord are appealing for capsuleer help to limit this attempted incusrion."
CCP Yokai, please, pleeeease, take a note. Call the event guys and let tell them to "make it so". You kill several birds with it - add some event to enrich EVE universe, you make RPers happy by providing them with high performance solar system and you get ****load of players stress testing new hardware.
I would also suggest that instead of Jita, you would designate two solarystems and have one run old node and other on new hardware. Without telling players which is which or even changing SOL affinity every day. This way you can get comparative data on usage models that model actual server use (instead of Jita'ish sitting in station) and reliable player feedback (it would be essentially quality single blinded scientific experiment, compared to basically gossip and self-suggestion you have now).
I know it's more work, but I do believe it's worth it (and it would be event team doing majority of heavy lifting).
|
something somethingdark
|
Posted - 2010.08.26 14:58:00 -
[115]
If only ecc reg ram had a faster timings ....
|
Purple Warlock
|
Posted - 2010.08.26 15:37:00 -
[116]
Originally by: Levitikon
Originally by: Jita Alt666
Utilise recent Sansha Incursion, "There are rumours of a sansha manipulated wormhole being targeted to develop in Jita. The last of these manipulated wormholes resulted in a sizable Sansha force emerging in high security space and momentarely overwhelming concord forces. Inside sources indicate the wormhole generation attempt will be between 1700 - 2300 on Wednesday 25 August 114. The same sources indicate that the Sansha Forces will be heavily dependent on their own faction ships and modules. Concord are appealing for capsuleer help to limit this attempted incusrion."
CCP Yokai, please, pleeeease, take a note. Call the event guys and let tell them to "make it so". You kill several birds with it - add some event to enrich EVE universe, you make RPers happy by providing them with high performance solar system and you get ****load of players stress testing new hardware.
I would also suggest that instead of Jita, you would designate two solarystems and have one run old node and other on new hardware. Without telling players which is which or even changing SOL affinity every day. This way you can get comparative data on usage models that model actual server use (instead of Jita'ish sitting in station) and reliable player feedback (it would be essentially quality single blinded scientific experiment, compared to basically gossip and self-suggestion you have now).
I know it's more work, but I do believe it's worth it (and it would be event team doing majority of heavy lifting).
Im down for that. CCP make it happen pliz. Ohh and Jita is AWESOME NOW. <3
|
Mashie Saldana
BFG Tech
|
Posted - 2010.08.26 16:11:00 -
[117]
Originally by: Teck Nine here's a rough list on the cost of AMD CPUs
1 - AMD 1090T6 Core 3.2GHz, 6MB Cache, 2000MHz (4000 MT/s ) FSB,Socket AM3, Processor would cost around $329.97
2 - AMD Opteron 2435 Six-Core Processor - 2.60GHz, Socket F, 6MB L3 Cache - $1300 3 - AMD Opteron 2431 Six-Core Processor - 2.40GHz, Socket F, 6MB L3 Cache - $900
4 - AMD Opteron 6172 Magny-Cours 2.1GHz 12 x 512KB L2 Cache 12MB L3 Cache Socket G34 115W 12-Core Server Processor - $1,070.00
With numbers 2 and 3 there's server boards that suport 2 to 4 CPUs and cau suport up to 165GB of ram and it also supports the new DDR3 type of ram.
For the 12 core there's boards that support 1 to 2 CPUs and can suport around 65GB of ram and it supports the new DDR3 ram.
So as you can see there are alot of options to make a powerful server rig that can deal with alot of load.
And you totally failed to realise that CCP is interested in max performace per core, not per CPU as EVE won't be multi-threaded anytime soon.
|
SFX Bladerunner
Minmatar Black Serpent Technologies R.A.G.E
|
Posted - 2010.08.26 17:47:00 -
[118]
Originally by: Mashie Saldana
Originally by: Teck Nine here's a rough list on the cost of AMD CPUs
1 - AMD 1090T6 Core 3.2GHz, 6MB Cache, 2000MHz (4000 MT/s ) FSB,Socket AM3, Processor would cost around $329.97
2 - AMD Opteron 2435 Six-Core Processor - 2.60GHz, Socket F, 6MB L3 Cache - $1300 3 - AMD Opteron 2431 Six-Core Processor - 2.40GHz, Socket F, 6MB L3 Cache - $900
4 - AMD Opteron 6172 Magny-Cours 2.1GHz 12 x 512KB L2 Cache 12MB L3 Cache Socket G34 115W 12-Core Server Processor - $1,070.00
With numbers 2 and 3 there's server boards that suport 2 to 4 CPUs and cau suport up to 165GB of ram and it also supports the new DDR3 type of ram.
For the 12 core there's boards that support 1 to 2 CPUs and can suport around 65GB of ram and it supports the new DDR3 ram.
So as you can see there are alot of options to make a powerful server rig that can deal with alot of load.
And you totally failed to realise that CCP is interested in max performace per core, not per CPU as EVE won't be multi-threaded anytime soon.
Ahum, that's a nicely inaccurate statement at the end there. Considering the location nodes alone provide a number of threads equal to the amount of solar systems in eve so saying 'eve' or the 'eve server' is not multi-threaded would be a grossly inaccurate statement as the 'eve server' is actually a cluster of threads divided by function, further subdivided if applicable (location nodes for instance), constantly communicating with eachother and the players and on top of that they are (going to be) dynamically spread around the available cores on the available servers.
Note that the term 'server' in the professional industry refers to a single xU mounting case containing a single motherboard and applicable hardware. These 'servers' in the form of database servers, storage servers, forum servers, service servers, load balancers/spreaders routers, firewalls and what not all fill up rackmount cases and the entirety of those interconnected servers filling up an x number of rackmount cases constitute the 'server park', which could be referred to as 'the eve online server' in a more traditional all-encompassing function-defined 'gamers' way. __________________________________________________
History is much like an endless waltz, the three beats of war, peace and revolution continue on forever.. |
Rakessh
Antares Shipyards Circle-Of-Two
|
Posted - 2010.08.26 17:48:00 -
[119]
Edited by: Rakessh on 26/08/2010 17:52:39 Actually, AMD Magnycours nodes might be better for holding all those 20 and below population systems, and the Intel one seem far far better at handling congested and contested systems and for node reinforcement use.
Edit: AMD 1090T6 is also out of the question as this is a desktop part, and will not find it's way into mainstream blade designs.
CEO Arachnea Phoenix Battalion |
Tinwan Sonomura
|
Posted - 2010.08.26 18:18:00 -
[120]
I just jumped over from Itamo and while there was a bit of lag, maybe 5 seconds jumping in from Maurasi, I noticed no other lag, docking, undocking, installing or deinstalling modules, or moving stuff to and from my cargo bay. Here here CCP, nice work on addressing the lag issues in Jita.
|
|
Teck Nine
|
Posted - 2010.08.26 18:29:00 -
[121]
Originally by: Rakessh AMD 1090T6 is also out of the question as this is a desktop part, and will not find it's way into mainstream blade designs.
I know AMD 1090T isn't a server CPU I put that there to give some people the idea of the price difference is between the Desktop CPU, the 6 core server CPU and the 12 core server CPU
|
Grez
Empire Assault Corp Dead Terrorists
|
Posted - 2010.08.26 20:13:00 -
[122]
They want clocks, not cores. In some cases, having more cores is absolutely the wrong answer. AMD's Bulldozer CPU is said to get up to 4.2Ghz, but we won't be seeing those til mid 2011, so that's out of the question. The new XEON's are absolutely a step in the right direction. ---
|
Mr Jaffar
|
Posted - 2010.08.26 21:35:00 -
[123]
My feedback is: 1215 pilots and as soon as I moved off the Sobaseki gate the client crashed... I tried 3 more times, client last about 5 seconds after the ship was drawn... So I'm now stuck in Jita. I hope its just teething problems.
|
EMO JAck
|
Posted - 2010.08.26 22:03:00 -
[124]
I would dearly love to know how the game functions on the back end.
Untel we know how the queries and activitys are handled we are just blowing smoke. That said me two sence. Now as for the memory abount on this system can the south bridge even handle that kinda load, and woudlent it make more sence to build racks or clusters. You could easly move the hardware around add and remove resorces at moments notice. Not to mention you could even set them up so that any lag would be shared across all of eve. Simply by having the servers share load. The software shuld be pretty compatable too.
|
|
CCP Explorer
|
Posted - 2010.08.26 22:22:00 -
[125]
Originally by: Rowbin Hod The the location node running on just one of those cores? what are the other cores doing?
... nothing. There is just one node running on this blade currently.
Erlendur S. Thorsteinsson Software Director EVE Online, CCP Games |
|
Victor Valka
Caldari Endoxa Corporation
|
Posted - 2010.08.26 22:24:00 -
[126]
I don't frequent Jita a lot, but when I've been there since the upgrade, with about 1000 to 1400 in local, it has seemed to perform better -- jumping in is faster, and so is docking and undocking.
(Undocking from Jita 4-4 without insta-undock bookmarks should be punished by CONCORD.)
Originally by: Spaztick You are not outnumbered, you are in a target-rich environment.
|
Johannus Shand
|
Posted - 2010.08.26 22:41:00 -
[127]
You could always try running jita on a cluster of recycled smart phones.
|
Stick Cult
|
Posted - 2010.08.26 22:44:00 -
[128]
Edited by: Stick Cult on 26/08/2010 22:44:00
Originally by: Johannus Shand You could always try running jita on a cluster of recycled smart phones.
PS3 cluster best cluster, obviously...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Berexia Cannin
|
Posted - 2010.08.27 08:53:00 -
[129]
guys i was in jita the today and i must say it was lagless the new stuff must be doing its job well.
|
Holly Hoover
|
Posted - 2010.08.27 10:33:00 -
[130]
Edited by: Holly Hoover on 27/08/2010 10:33:36 First the good, lag is mostly not noticeable in Jita. But there seems to be a bad side effect. During session changes in or around Jita, like warping up to a outside heavily populated station, if there is a lag, the client often gets dropped! If I jump thru a gate into Jita with lots of ships on the other side, I often get dropped! Is this how ther server solves the lag issue, by dropping clients to decrease the load??
"Yes I pvp fight with the market view open!" |
|
Corbeau Lenoir
ZER0. IT Alliance
|
Posted - 2010.08.27 10:57:00 -
[131]
Originally by: Berexia Cannin guys i was in jita the today and i must say it was lagless the new stuff must be doing its job well.
Same experience and opinion here.
|
Niccolado Starwalker
Gallente Shadow Templars
|
Posted - 2010.08.27 11:59:00 -
[132]
Edited by: Niccolado Starwalker on 27/08/2010 11:59:01
Originally by: Corbeau Lenoir
Originally by: Berexia Cannin guys i was in jita the today and i must say it was lagless the new stuff must be doing its job well.
Same experience and opinion here.
Now how could all this suddenly happen??? I know that the time flies by, but 18 months??????
Originally by: Dianabolic Your tears are absolutely divine, like a fine fine wine, rolling down your cheeks until they flow down the river of LOL. |
zz01shagsme
|
Posted - 2010.08.27 14:13:00 -
[133]
Would this have caused issues elsewhere, I was having real problems last night in Sinq Laison, Black screen, emergency warps, slow warping non-responsive...JITA itself seemed fine with my alt.
|
Super Whopper
I can Has Cheeseburger
|
Posted - 2010.08.27 16:46:00 -
[134]
While I'd rather see CCP use AMD CPU's it's inevitable they'd use Intel. This game purely runs on clock speed and Intel's CPU's are just faster than AMD's and I don't see this changing anytime soon. If you're building a super computer for purely number crunching then AMD is the best choice, as you get more CPU's, and cores, for your money but if the choice is pure power core speed then Intel are it.
Also they're not stressing the other cores of that CPU so the single core is overclocked automatically to max speed and max performance is squeezed out of the system. Also 1300 people in system checking markets doesn't immediately mean that 600 people shooting at eachother will also experience no lag, because they will. I am also confident CCP won't have a clue about this for another two years, until some CSM member points it out to them and shows them how awefully clueless they are.
Originally by: Niccolado Starwalker Edited by: Niccolado Starwalker on 27/08/2010 11:59:01
Originally by: Corbeau Lenoir
Originally by: Berexia Cannin guys i was in jita the today and i must say it was lagless the new stuff must be doing its job well.
Same experience and opinion here.
Now how could all this suddenly happen??? I know that the time flies by, but 18 months??????
You mean probably in 18 months? You should never forget the word probably because it simply means lolgo****yourselflolidiotcustomerslol.
|
Stick Cult
|
Posted - 2010.08.27 16:51:00 -
[135]
Originally by: Super Whopper Also 1300 people in system checking markets doesn't immediately mean that 600 people shooting at eachother will also experience no lag, because they will.
No, but the fact that the inventory system (the most apparent result of lag in Jita, other than actually being ~in space~) is more snappy means that things are being handled better by a faster server, which means a fleet fight will run smoother. Maybe only 5 or 10% smoother, but if you get a dozen changes each help by 5 or 10%, bam, things are running a lot better all the sudden.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Aldiree Antima
Caldari Science and Trade Institute
|
Posted - 2010.08.27 19:15:00 -
[136]
Edited by: Aldiree Antima on 27/08/2010 19:15:45
Originally by: CCP Explorer
Originally by: Rowbin Hod The the location node running on just one of those cores? what are the other cores doing?
... nothing. There is just one node running on this blade currently.
Is there a chance that we will see a multithread version of location node logik? (Maybe in the next three years or so?) Or is the logik itself not multithreading friendly?
|
Obsidian Hawk
RONA Legion
|
Posted - 2010.08.27 19:34:00 -
[137]
Went to jita yesterday to play around. Switched between brackets, no brackets, everything and nothing on overviews and several variations there of. no lag between them switches were extremely smooth.
now if i could find out why my client crashes randomly
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.27 20:40:00 -
[138]
Originally by: Obsidian Hawk Went to jita yesterday to play around. Switched between brackets, no brackets, everything and nothing on overviews and several variations there of. no lag between them switches were extremely smooth.
now if i could find out why my client crashes randomly
Probably because they are tweaking around the server settings. Had a couple myself after doing some session changes playing with the alts but its not very common at least for me.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Teck Nine
|
Posted - 2010.08.28 10:06:00 -
[139]
Ever since they've done stuff I've been getting more d/c now then before they did the upgrade
|
Sev Morde
|
Posted - 2010.08.28 16:54:00 -
[140]
I know this is an old thread.
But if you really want to do some stress testing, Involve the entrants to DreamHack, should get a Few thousand in the One system.
|
|
Professor Tarantula
Hedion University
|
Posted - 2010.08.28 18:43:00 -
[141]
Originally by: Obsidian Hawk
Just posting to say that's great.
My Warmest Regards. Prof. Tarantula, Esq. |
Jita Alt666
|
Posted - 2010.08.28 21:15:00 -
[142]
I had a small amount of market lag while looking at my orders then checking market details on each of them. 08.34eve time 1400ish in jita local
|
Infested Stukov
Infested U.E.D.
|
Posted - 2010.08.29 06:18:00 -
[143]
Can i still make lag with shuttles? HAO OOPSE arvld here, jst wnted to say im going to bes gonde for a whle, bnad for froms for almost a year, and csnt resub, i luve you oope, your all awesom, n ot as awesome as mer, but still aseome.
|
Jim Luc
Caldari Rule of Five Lucky Starbase Syndicate
|
Posted - 2010.08.29 16:52:00 -
[144]
I'm guessing there's no way to split a system up between nodes.
For instance, what people do in station does not effect anyone outside the station, right? Also, can a systems be split up into dynamic "grids", with each on a separate node if needed? This seems the best way to scale, as more and more people start playing Eve.
Also, on a side note, can you set up a station cam, showing what's going on outside the station? Especially when Incarna hits - the ability to look outside would be amazingly cool.
|
carebear one
|
Posted - 2010.08.29 22:45:00 -
[145]
Hi,
well I watch the performance in Jita some times with this new hardware. There is abolutly no change. Jita is a slow and laggy as always. Business transactions take up to 5 min til they are processed by the server. Backwards-Warping as allways and so on.
I think its time that CCP accepts the fact that Jita is THE trading hub in EVE and dont try to change that.
I think the way to get Jita going is just simple. Make the system like a POS-Forcefield. No target-locking possible in Jita. All the ship just hanging around locking and scanning every ship will have to move to the systems around jita. EVE has thousands of systems where a ganker or a griefer can enjoy his fetish. Just throw them out of Jita so the Capacities on the Cluster goes to the players who do what Jita is intended for. Trading. Also the CPU-time needed for the CONCORD-Scripts could be safed.
I think this simple and easy to implement suggestion would free up so many Capacities on the Jita-Cluster that no new hardware is needed.
|
Stick Cult
|
Posted - 2010.08.29 23:14:00 -
[146]
Originally by: carebear one well I watch the performance in Jita some times with this new hardware. There is abolutly no change. Jita is a slow and laggy as always. Business transactions take up to 5 min til they are processed by the server. Backwards-Warping as allways and so on.
I've never had issues with that.. market transactions have always taken max 20 seconds for me, same with fitting things to ships. Maybe it's your internet?
Quote: I think its time that CCP accepts the fact that Jita is THE trading hub in EVE and dont try to change that.
I'm pretty sure they have accepted that...
Quote: I think the way to get Jita going is just simple. Make the system like a POS-Forcefield. No target-locking possible in Jita. All the ship just hanging around locking and scanning every ship will have to move to the systems around jita. EVE has thousands of systems where a ganker or a griefer can enjoy his fetish. Just throw them out of Jita so the Capacities on the Cluster goes to the players who do what Jita is intended for. Trading. Also the CPU-time needed for the CONCORD-Scripts could be safed.
Certainly an interesting suggestion... I do like the suicide ganking in Jita, though...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.29 23:16:00 -
[147]
stick cult: trolled and rolled.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Stick Cult
|
Posted - 2010.08.29 23:22:00 -
[148]
Originally by: Zeba stick cult: trolled and rolled.
Naa, just ****ing bored. Oh god am I bored...
/me goes off to get moar forum warnings...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.08.29 23:24:00 -
[149]
Originally by: Stick Cult
Originally by: Zeba stick cult: trolled and rolled.
Naa, just ****ing bored. Oh god am I bored...
/me goes off to get moar forum warnings...
Go to S&M and bait Intigo. He is always good to slap around for some jollies. Willing too.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Stick Cult
|
Posted - 2010.08.29 23:34:00 -
[150]
Originally by: Zeba
Originally by: Stick Cult
Originally by: Zeba stick cult: trolled and rolled.
Naa, just ****ing bored. Oh god am I bored...
/me goes off to get moar forum warnings...
Go to S&M and bait Intigo. He is always good to slap around for some jollies. Willing too.
Meh I'm gonna go play a game or something... Not eve, of course! Maybe Supreme Commander 2 or Borderlands...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
|
Tres Farmer
Gallente Federation Intelligence Service
|
Posted - 2010.08.30 00:18:00 -
[151]
Originally by: Aldiree Antima Edited by: Aldiree Antima on 27/08/2010 19:15:45
Originally by: CCP Explorer
Originally by: Rowbin Hod The the location node running on just one of those cores? what are the other cores doing?
... nothing. There is just one node running on this blade currently.
Is there a chance that we will see a multithread version of location node logik? (Maybe in the next three years or so?) Or is the logik itself not multithreading friendly?
Read the Follow-Up-Thread on CCP Warlocks Dev Blog.. I think she entioned some stuff into this dierction, but said explicitely that she can't and won't give details/timeframes.. only that its on the table as far as I did understand it
|
|
CCP Yokai
|
Posted - 2010.08.30 13:02:00 -
[152]
Okay so here's the deal.
Testing is going REALLY well.
We tried the server on its default configuration to get a baseline. We have since disabled hyper threading... turned on turbo boost to max... and let the little hamsters take a few more pills.
Yesterday Jita peaked about 1550 players... with the server CPU hanging around at 80%.
Less than a month ago the CPU was at a near flat line 100% at just 1400 players. Part of this change is because of the super smart software guys making some changes (not going to steal their dev blog thunder) and then this new hardware.
Just has now been set to an 1800 limit from its previous 1400. I'd love to see that limit tested too. The goal is not to have a massive fleet fight in Jita. That's not a true indicator of the performance heuristics of Jita. But, if you can and if you are willing... please spend a bit of time in Jita today.
|
|
Jim Luc
Caldari Rule of Five Lucky Starbase Syndicate
|
Posted - 2010.08.30 15:43:00 -
[153]
I think CCP might get more results if they send a notification to everyone's EVEMail when they want us to participate in a test. I don't want to get a bunch of spam from CCP to buy things, but a simple notification and a personal email requesting aid would probably bring people by the thousands.
|
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.08.30 19:57:00 -
[154]
Yes, CCP, you have to post fresh announcement, because the "New Jita hardware" thread is not part of the blog-roll, it was posted on the home page and now it's burried below other topics.
Why don't you just ask people to log in to Jita prior to the mass testing on Thursday? Post a Blog, a New item + EVE Mail ;-) ___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Clee eelc
|
Posted - 2010.08.30 21:17:00 -
[155]
CCP could give 1 million to everyone on during the test. it might attract new players.
|
Tessty
|
Posted - 2010.08.30 21:55:00 -
[156]
@Clee - no that would be, give CCP 1 million isk and they will double it for testing the new hardware.
|
mechtech
SRS Industries SRS.
|
Posted - 2010.08.30 22:01:00 -
[157]
Edited by: mechtech on 30/08/2010 22:05:03
Originally by: Jim Luc I think CCP might get more results if they send a notification to everyone's EVEMail when they want us to participate in a test. I don't want to get a bunch of spam from CCP to buy things, but a simple notification and a personal email requesting aid would probably bring people by the thousands.
That's what the calender is for.
edit: Assuming changing sockets isn't too big a deal, you might as well wait for 32nm Sandy Bridge Intel processors coming out very soon. They're an additional 15% faster clock for clock over Nehalem (and run higher clocks with a more aggressive Turbo Boost). That means that any advantage seen from moving from Wolfdale to Nehalem (what's currently being tested) will be matched again from moving to Sandy Bridge.
|
HeIIfire11
|
Posted - 2010.08.30 22:43:00 -
[158]
I remember the times when jita died as there were 600 or so players..no it runs good with over 1000..good job on that
|
Fam Trinly
|
Posted - 2010.08.31 02:58:00 -
[159]
I was in Jita yesterday (7 times from 16 to 18 GMT), have not activate my modules, only jumping/docking.
My PC config is: Pentium D945 3,4Ghz (2 cores), 2+GB DDR2 RAM, Eve client on sata SSD (SLC), ATI Radeon HD 5870, WinXP SP3, fresh DirectX 9c. 1920*1200 all shaders at high quality, brackets in overview.
I watched with in-game fps-counter and results was: 1) 1k local in Jita - lower fps on undocking 4-4 and jumping to/from perimeter - near 15 2) 1,3k local in Jita - lower fps on undocking 4-4 and jumping to/from perimeter - near 7
on undocking spot wich is 3k km far from 4-4 station - no lags, average fps was around 55 (like in other solar systems).
trading and contracting with 1k and 1,3k local was normal ... sometimes 3-5 seconds of waiting contract data to refresh after accepting.
P.S. If you want more IO perfomance (for sql) on Jita node hardware - I'm advising you to think about SSD storage solutions (which should be faster than 10k rmp sas hhd, and more asynchronosly with no sector-seek delays, no need to defragment file-system), or combination of little volume on SSD (temp, write log and may be indexes) and large on 15k rpm hhd for tables.
|
Jita Alt666
|
Posted - 2010.08.31 23:52:00 -
[160]
I heard that a ccp employee told a csm that we were allowed to attempt to break jita. That they simply wanted as many people in local as possible.
Is this true? Will I get into trouble for logging all my accounts on and sitting on all gates/undocks and insynch launching smart bombs?
|
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.01 02:53:00 -
[161]
Originally by: Jita Alt666 I heard that a ccp employee told a csm that we were allowed to attempt to break jita. That they simply wanted as many people in local as possible.
Is this true? Will I get into trouble for logging all my accounts on and sitting on all gates/undocks and insynch launching smart bombs?
For once Jitafest is an 'approved' activity.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
NekoKitten
|
Posted - 2010.09.01 10:52:00 -
[162]
CL9 memory isnt particulary low latency ;) There is faster memory available .. And I would've chosen Intel solid state disks, as those are lightning fast. (when you're spending money on 10k SAS disks, then you can spend the same money on SSD's as well) Just my idea ..
|
Ban Doga
|
Posted - 2010.09.01 13:02:00 -
[163]
Originally by: NekoKitten And I would've chosen Intel solid state disks, as those are lightning fast. (when you're spending money on 10k SAS disks, then you can spend the same money on SSD's as well) Just my idea ..
Then you would have spent money for useless hardware, since the node does not store game data on local disks.
|
Crieton
|
Posted - 2010.09.01 14:50:00 -
[164]
What chassis do you guys use? BladeCenter H? Also, I know that blade supports 2 addins, have you used any of slots? I assume the FC backhaul card is installed but what about extra 1gbit ports?
|
|
CCP Yokai
|
Posted - 2010.09.01 19:45:00 -
[165]
Originally by: Crieton What chassis do you guys use? BladeCenter H? Also, I know that blade supports 2 addins, have you used any of slots? I assume the FC backhaul card is installed but what about extra 1gbit ports?
Blade Center H
No need for fiber channel or extra gig-E cards on the server... none of that really helps. No disk I/O local or remote attached. And, the network throughput of EVE is very minimal.
|
|
|
CCP Yokai
|
Posted - 2010.09.01 19:47:00 -
[166]
Originally by: Zeba
Originally by: Jita Alt666 I heard that a ccp employee told a csm that we were allowed to attempt to break jita. That they simply wanted as many people in local as possible.
Is this true? Will I get into trouble for logging all my accounts on and sitting on all gates/undocks and insynch launching smart bombs?
For once Jitafest is an 'approved' activity.
We are asking playes to do what they normally do in Jita as much as possible not to skew the test results anymore than we need to. but some people do that stuff in Jita too :)
Jita is protected by a population cap (currently set to 1800), in the evet the server is over capacity we can simply reduce the cap to protect system stability. Breaking Jita completely is not... dare I say it... possible.
|
|
Jita Alt666
|
Posted - 2010.09.01 20:18:00 -
[167]
You havent answered my question: Is it true a ccp employee has given the playerbase permission to attempt to break the Jita node? In other words before I contemplate this notion I want to have it in writing that I and others are not going to face bans for causing issues.
|
|
CCP Yokai
|
Posted - 2010.09.01 20:27:00 -
[168]
Originally by: Jita Alt666 You havent answered my question: Is it true a ccp employee has given the playerbase permission to attempt to break the Jita node? In other words before I contemplate this notion I want to have it in writing that I and others are not going to face bans for causing issues.
We've asked players to log on in Jita and play EVE.
If you do that and just that... there will be no issue. If you decide the best way to help us TEST or BREAK Jita is by means outside the EULA... then you do so at your own risk.
Does that answer your question?
|
|
Jita Alt666
|
Posted - 2010.09.01 20:38:00 -
[169]
So the ccp employee who told a csm to: "Log in as many as you can in jita we challenge you to break it" needs to be censured by his employer.
|
Stick Cult
|
Posted - 2010.09.01 20:39:00 -
[170]
Edited by: Stick Cult on 01/09/2010 20:39:49
Originally by: Jita Alt666 So the ccp employee who told a csm to: "Log in as many as you can in jita we challenge you to break it" needs to be censured by his employer.
No you're just twisting his words...
They want to test Jita by getting people in it. If it breaks, meh. If you're spamming lag inducing **** like there's no tomorrow, I bet they don't want that, as it wouldn't be an accurate test.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
|
Jita Alt666
|
Posted - 2010.09.01 20:50:00 -
[171]
Originally by: Stick Cult Edited by: Stick Cult on 01/09/2010 20:39:49
Originally by: Jita Alt666 So the ccp employee who told a csm to: "Log in as many as you can in jita we challenge you to break it" needs to be censured by his employer.
No you're just twisting his words...
They want to test Jita by getting people in it. If it breaks, meh. If you're spamming lag inducing **** like there's no tomorrow, I bet they don't want that, as it wouldn't be an accurate test.
Quote:
CCP Yokai contacted CSM today with a challenge: Let's Break Jita Tonight!
As most everyone knows, Jita consistently places a heavy load on servers. Because of that, CCP began testing recent lag-related improvements to hardware and software in Jita. Test results indicate that Jita capacity has increased at least 25% compared to a month ago! The big question is: "Where's the tipping point now?"
yes totally twisted
|
ScreammShadow
|
Posted - 2010.09.01 20:54:00 -
[172]
Originally by: Stick Cult
... If you're spamming lag inducing **** like there's no tomorrow, I bet they don't want that, as it wouldn't be an accurate test.
Guess u havnt been there in a while?, ALL hubs have spamming like nuts. So it would, in a way, be acurate.
|
|
CCP Yokai
|
Posted - 2010.09.01 20:55:00 -
[173]
Originally by: Jita Alt666
Originally by: Stick Cult Edited by: Stick Cult on 01/09/2010 20:39:49
Originally by: Jita Alt666 So the ccp employee who told a csm to: "Log in as many as you can in jita we challenge you to break it" needs to be censured by his employer.
No you're just twisting his words...
They want to test Jita by getting people in it. If it breaks, meh. If you're spamming lag inducing **** like there's no tomorrow, I bet they don't want that, as it wouldn't be an accurate test.
Quote:
CCP Yokai contacted CSM today with a challenge: Let's Break Jita Tonight!
As most everyone knows, Jita consistently places a heavy load on servers. Because of that, CCP began testing recent lag-related improvements to hardware and software in Jita. Test results indicate that Jita capacity has increased at least 25% compared to a month ago! The big question is: "Where's the tipping point now?"
yes totally twisted
There are no issues with the post. I asked the CCP employee to post that for me. It's quite clear we want as many players as possible in Jita. Do that :)
Thanks
|
|
Jita Alt666
|
Posted - 2010.09.01 21:01:00 -
[174]
Originally by: ScreammShadow
Originally by: Stick Cult
... If you're spamming lag inducing **** like there's no tomorrow, I bet they don't want that, as it wouldn't be an accurate test.
Guess u havnt been there in a while?, ALL hubs have spamming like nuts. So it would, in a way, be acurate.
Thats one of my issues - technically spamming local is against the eula. However CCP have increasingly allowed it to roll unihibited. But if we succeeded in breaking a node and one of the tools we used was extreme local spam CCP would pull out the eula and claim we were different than what already occurs.
Likewise with anchored cans, timed drone release, grid manipulation, concord activation.
|
mechtech
SRS Industries SRS.
|
Posted - 2010.09.01 21:14:00 -
[175]
Originally by: Jita Alt666
Thats one of my issues - technically spamming local is against the eula. However CCP have increasingly allowed it to roll unihibited. But if we succeeded in breaking a node and one of the tools we used was extreme local spam CCP would pull out the eula and claim we were different than what already occurs.
Likewise with anchored cans, timed drone release, grid manipulation, concord activation.
Oh come on, there's no need to be so technical. Cops can arrest you for J-walking, doesn't mean they have to.
Obviously CCP isn't going to gag/ban every noob who rolls into Jita and starts begging for isk. Likewise, CCP asking us to heavily load Jita is a special case for testing new server hardware.
On the other hand, people going into a system and deliberately trying to crash it by spamming cans and local is a different scenario, because the goal is not to test server hardware, but in most cases to annoy or kill another corp/player. If you're deliberately trying to use lag w/ extreme methods such as anchoring 1000 cans, timing drone releases, or activating 10k concord ships to gain an advantage, that's obviously overstepping the line.
|
Jita Alt666
|
Posted - 2010.09.01 21:19:00 -
[176]
Edited by: Jita Alt666 on 01/09/2010 21:19:49
Originally by: mechtech
Originally by: Jita Alt666
Thats one of my issues - technically spamming local is against the eula. However CCP have increasingly allowed it to roll unihibited. But if we succeeded in breaking a node and one of the tools we used was extreme local spam CCP would pull out the eula and claim we were different than what already occurs.
Likewise with anchored cans, timed drone release, grid manipulation, concord activation.
Oh come on, there's no need to be so technical. Cops can arrest you for J-walking, doesn't mean they have to.
Obviously CCP isn't going to gag/ban every noob who rolls into Jita and starts begging for isk. Likewise, CCP asking us to heavily load Jita is a special case for testing new server hardware.
On the other hand, people going into a system and deliberately trying to crash it by spamming cans and local is a different scenario, because the goal is not to test server hardware, but in most cases to annoy or kill another corp/player. If you're deliberately trying to use lag w/ extreme methods such as anchoring 1000 cans, timing drone releases, or activating 10k concord ships to gain an advantage, that's obviously overstepping the line.
It's not a line its a grey band that CCP get to choose where you have gone to far. That is why when I recieved a mail saying: "CCP challenge us to break Jita" I logged in here and have asked for clarification.
|
Abdiel Kavash
Caldari Paladin Order Fidelas Constans
|
Posted - 2010.09.01 21:28:00 -
[177]
CCP, put this new node or whatever into Pure Blind, we will gladly break it for you in zero time. ___________ EVE is dying! Now for real! |
Cergorach
Amarr The Helix Foundation
|
Posted - 2010.09.01 21:49:00 -
[178]
Scenario: 1800 EVE players sitting in station Jita 4-4, they all decide to hit the undock button at once. What happens on/to the server?
|
Stick Cult
|
Posted - 2010.09.01 21:54:00 -
[179]
Originally by: Cergorach Scenario: 1800 EVE players sitting in station Jita 4-4, they all decide to hit the undock button at once. What happens on/to the server?
Smoke.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Jita Alt666
|
Posted - 2010.09.01 22:19:00 -
[180]
Originally by: Stick Cult
Originally by: Cergorach Scenario: 1800 EVE players sitting in station Jita 4-4, they all decide to hit the undock button at once. What happens on/to the server?
Smoke.
Followed by CCP banning the accounts that are alts of non IT members, because again such an action is against the EULA
|
|
Mr Dogg
Amarr Double-Down Demolition Notice
|
Posted - 2010.09.01 23:00:00 -
[181]
Remembers Yulia, The original "market-hub-node-nightmare". All this time and I'll give it to CCP they have improved EVE EXPONENTIALLY since those days.
Smart-arsed comments GO!....
|
Jita Alt666
|
Posted - 2010.09.01 23:15:00 -
[182]
Originally by: Mr Dogg Remembers Yulia, The original "market-hub-node-nightmare". All this time and I'll give it to CCP they have improved EVE EXPONENTIALLY since those days.
Smart-arsed comments GO!....
was that when i was i running eve on my pIII 800 with a 25gig hd and 384mb ram? The world has moved faster than eve development.
|
Ager Agemo
|
Posted - 2010.09.01 23:34:00 -
[183]
excuse me if i sound ignorant, but if nodes can only run on a single core, wouldn¦t it make sense to have a machine apart that has a single/dual core super clocked to handle special cases ? ( big fleet battles and alike ) something like an AMD procesor at 5000mhz or so? i believe a less core new procesor can easily be overclocked much more than a multicore procesor. or i m wrong?
|
Stick Cult
|
Posted - 2010.09.01 23:46:00 -
[184]
Edited by: Stick Cult on 01/09/2010 23:47:48
Originally by: Ager Agemo excuse me if i sound ignorant, but if nodes can only run on a single core, wouldn¦t it make sense to have a machine apart that has a single/dual core super clocked to handle special cases ? ( big fleet battles and alike ) something like an AMD procesor at 5000mhz or so? i believe a less core new procesor can easily be overclocked much more than a multicore procesor. or i m wrong?
What chip can reach 5 ghz, even overclocked? And then, more importantly, what server chip can reach 5 ghz? (Not sure on the ease of overclocking thing single vs multiple core thing) Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM). The new Jita node has faster RAM, which is why it can handle more people.
edit: Part of CCP's plan with the servers was them thinking "whoa, chips are getting so much faster ever day! We can just upgrade our servers every few years to reduce lag..." back in 2000. Well, they kinda flopped. (lolololol)
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Professor Tarantula
Hedion University
|
Posted - 2010.09.02 00:53:00 -
[185]
Clearly, CCP must begin harvesting human brains to act as servers.
My Warmest Regards. Prof. Tarantula, Esq. |
Breaker77
Gallente Reclamation Industries
|
Posted - 2010.09.02 01:10:00 -
[186]
Originally by: Stick Cult Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM).
I thought that the reason they use 1 core is that EVE is not programmed to use multiple cores therefore they are a waste
|
Stick Cult
|
Posted - 2010.09.02 02:17:00 -
[187]
Originally by: Breaker77
Originally by: Stick Cult Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM).
I thought that the reason they use 1 core is that EVE is not programmed to use multiple cores therefore they are a waste
Right, but I thought I remember them saying they could pack basically 2 copies of the server code into 2 seperate cores on one blade. Which they don't, because of the RAM issue. And Eve is only running on one core.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
thegallenteamarrguy
|
Posted - 2010.09.02 02:20:00 -
[188]
Edited by: thegallenteamarrguy on 02/09/2010 02:22:34
Originally by: Jita Alt666
Originally by: Stick Cult
Originally by: Cergorach Scenario: 1800 EVE players sitting in station Jita 4-4, they all decide to hit the undock button at once. What happens on/to the server?
Smoke.
Followed by CCP banning the accounts that are alts of non IT members, because again such an action is against the EULA
Keep tryin' real hard to get a negative response from CCP. I think it's working.
For an actual question: Why not just get a couple hundred of the android bots that you talked about earlier to come into jita to see what it'd be like?
|
Zendoren
Aktaeon Industries
|
Posted - 2010.09.02 05:02:00 -
[189]
Enjoy
IBM HS22
Over 4,000!!!!!!!!!!!!!!!!!!
|
NekoKitten
Gallente Novus Culpa Astra The Jagged Alliance
|
Posted - 2010.09.02 05:19:00 -
[190]
I was in Dodixie last night and I think people got it mistaken with Jita, cos selling of stuff was beginning to crawl ;) |
|
Lieutenant Obvious
Lobster Sticks To Magnet
|
Posted - 2010.09.02 05:44:00 -
[191]
Originally by: Stick Cult
Originally by: Breaker77
Originally by: Stick Cult Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM).
I thought that the reason they use 1 core is that EVE is not programmed to use multiple cores therefore they are a waste
Right, but I thought I remember them saying they could pack basically 2 copies of the server code into 2 seperate cores on one blade. Which they don't, because of the RAM issue. And Eve is only running on one core.
Something to think about (assuming it hasn't been stated already, didnt read entire thread): New Nehalem/Westmere processors (5500 and 5600 series) have a feature called TurboBoost - if you only have 1 core doing work and all others are idle it'll overclock itself significantly. Don't quote me, but I remember numbers like 3.8GHz. For a single-threaded application that only cares about pure, unadulterated single-core grunt, that is a huge bone to throw it.
Which is why I can see this certainly being a very successful test against a dedicated, single-system node.
|
ma perke
|
Posted - 2010.09.02 06:13:00 -
[192]
Originally by: CCP Yokai Edited by: CCP Yokai on 25/08/2010 16:20:27
Okay... the spec:
IBM HS22 1 x Xeon 4C X5667 95W 3.06GHz/1333MHz/12MB 6 x 8GB PC3-10600 CL9 ECC DDR3 1333MHz VLP RDIMM 146 GB 2.5in Slim-HS 10K 6GB SAS HDD
at 4000USD this server is overpriced nothing to brag about... get the new SSD instead of SAS, theu have much more IOPS which should be the key here
|
HeliosGal
Caldari
|
Posted - 2010.09.02 06:23:00 -
[193]
how about fixing lag in other areas
|
Bloodswain Fresch
|
Posted - 2010.09.02 06:29:00 -
[194]
Sorry, but it's worse than ever. Just before posting this, sitting in the station Jita IV - Moon 6 - Hyasyoda Corporation Refinery, I tried to get a 6 month price history table for Medium Anti-thermal Screen Reinforcer I.
When I started, there were 693 local; and I was alone in this station. After 10 minutes of waiting while nothing happened, I gave up and tried again--same result.
Been playing on various accounts for three years; never seen it anywhere near this bad.
|
Ban Doga
|
Posted - 2010.09.02 06:55:00 -
[195]
Originally by: Stick Cult
What chip can reach 5 ghz, even overclocked? And then, more importantly, what server chip can reach 5 ghz? (Not sure on the ease of overclocking thing single vs multiple core thing) Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM). The new Jita node has faster RAM, which is why it can handle more people.
http://www.legitreviews.com/article/836/1/ That method might not be viable for day-to-day use but it shows that that CPU can run at over 5Ghz. Almost two years later and with a more conventional cooling method you might still get exceptional results.
|
Marc Mactravis
|
Posted - 2010.09.02 11:09:00 -
[196]
Originally by: NekoKitten I was in Dodixie last night and I think people got it mistaken with Jita, cos selling of stuff was beginning to crawl ;)
Yeah I was there too.
|
Ender Dax
|
Posted - 2010.09.02 17:17:00 -
[197]
Originally by: Ban Doga
Originally by: Stick Cult
What chip can reach 5 ghz, even overclocked? And then, more importantly, what server chip can reach 5 ghz? (Not sure on the ease of overclocking thing single vs multiple core thing) Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM). The new Jita node has faster RAM, which is why it can handle more people.
http://www.legitreviews.com/article/836/1/ That method might not be viable for day-to-day use but it shows that that CPU can run at over 5Ghz. Almost two years later and with a more conventional cooling method you might still get exceptional results.
Alright alright, pass around the hat for liquid nitrogen money and get this over with..
|
RiskyFrisky
Suddenly Ninjas
|
Posted - 2010.09.02 23:06:00 -
[198]
Edited by: RiskyFrisky on 02/09/2010 23:08:37 Okay, so after a few days of doing nothing but sitting around in Jita.
I've noticed that the time for the 'Employment History' tab in character panels now loads significantly quicker along side undocks etc. Jumping into the system also seems to be a bit quicker, as before node I was getting a blank screen for a good 5-10 seconds, it is now as if I was jumping into any other system.
You did good CCP. Now do better. _
Ohh hai, I saw you like to kill people, so I brought starcakes. |
Jita Alt666
|
Posted - 2010.09.02 23:40:00 -
[199]
Originally by: Ender Dax
Originally by: Ban Doga
Originally by: Stick Cult
What chip can reach 5 ghz, even overclocked? And then, more importantly, what server chip can reach 5 ghz? (Not sure on the ease of overclocking thing single vs multiple core thing) Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM). The new Jita node has faster RAM, which is why it can handle more people.
http://www.legitreviews.com/article/836/1/ That method might not be viable for day-to-day use but it shows that that CPU can run at over 5Ghz. Almost two years later and with a more conventional cooling method you might still get exceptional results.
Alright alright, pass around the hat for liquid nitrogen money and get this over with..
That hat has come around to me every 3 months for last 4 years and everytime it comes around I put 2 x $38.85 in it and then another 12+ x plex.
How many times will the hat have to go around to enable this fix to go ahead?
Fake edit: Oh another 6 times you say?
|
Barakkus
Kaalakiota Innovations
|
Posted - 2010.09.02 23:40:00 -
[200]
Originally by: Ban Doga
Originally by: Stick Cult
What chip can reach 5 ghz, even overclocked? And then, more importantly, what server chip can reach 5 ghz? (Not sure on the ease of overclocking thing single vs multiple core thing) Another bottleneck in the servers is RAM (Another reason CCP only uses 1 core on multicore processors, it's just more CPU trying to get to the same RAM). The new Jita node has faster RAM, which is why it can handle more people.
http://www.legitreviews.com/article/836/1/ That method might not be viable for day-to-day use but it shows that that CPU can run at over 5Ghz. Almost two years later and with a more conventional cooling method you might still get exceptional results.
I saw some video of someone overclocking like a 386 or some old crap to like 4ghz with some liquid nitrogen, ran for like 5 minutes before practically exploding lol
Originally by: CCP Dropbear
rofl
edit: ah crap, dev account. Oh well, official rofl at you sir.
|
|
HeliosGal
Caldari
|
Posted - 2010.09.02 23:58:00 -
[201]
ok so you have fixed jita whats next
are we talking more mass testing, software upgrades, finishing 00 factional warfare missions or factional warfare 00
|
KUN TEWK
|
Posted - 2010.09.03 04:29:00 -
[202]
I've noticed some unwelcome changes after the upgrades to Jita.... Now, at times, I'll click on a ship to put in my overview and my ship will suddenly approach that target w/out a command. Also I'm always getting a "You cannot do that during warp." instruction at inappropiate times.
|
Jake Falcor
|
Posted - 2010.09.03 06:13:00 -
[203]
Originally by: KUN TEWK I've noticed some unwelcome changes after the upgrades to Jita.... Now, at times, I'll click on a ship to put in my overview and my ship will suddenly approach that target w/out a command. Also I'm always getting a "You cannot do that during warp." instruction at inappropiate times.
Fix Your Mouse.
End.
|
|
CCP Yokai
|
Posted - 2010.09.03 09:50:00 -
[204]
Edited by: CCP Yokai on 03/09/2010 09:50:53 In response to questions about when do we do this for 0.0:
After Sunday we should have all the data we need to prove and measure the impact on Eve from:
Changes in Memory Latency Changes in Chip Architecture Changes in Cache Changes in Clock Rate Addition of Turbo Boost
When we have this we can much more easily spec the optimal system for Jita and similar systems and move on. This same server will be then used to benchmark fleet fights in 0.0
With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0
I hope that helps better explain the intentions of this testing.
|
|
Alara IonStorm
Agent-Orange
|
Posted - 2010.09.03 10:09:00 -
[205]
Umm... Yay!
I is happy!
-- Tactical Responder who is Organized and a Leading-edge Linguist |
Guns nButter
Stormlord Battleforce Vanguard.
|
Posted - 2010.09.03 17:14:00 -
[206]
i be thinking that CCP done good with 1300 in local and zero lag.
|
TriadSte
Gallente
|
Posted - 2010.09.03 19:14:00 -
[207]
I dont know much about servers but I know alot about desktop PCs and I have this question.
Core I7 are right now the very best CPU known to mankind. Why not use one of these for Jita? How is a Xeon better? theres no faster CPU for number crunching than a I7.
|
Vomit Comet
|
Posted - 2010.09.03 21:34:00 -
[208]
Edited by: Vomit Comet on 03/09/2010 21:34:28
Originally by: TriadSte Core I7 are right now the very best CPU known to mankind. Why not use one of these for Jita? How is a Xeon better? theres no faster CPU for number crunching than a I7.
Use some google-fu: the server blade mentioned at the start of this thread uses the server version (Xeon Westmere) of the Core i7.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.04 04:45:00 -
[209]
Originally by: Vomit Comet Edited by: Vomit Comet on 03/09/2010 21:34:28
Originally by: TriadSte Core I7 are right now the very best CPU known to mankind. Why not use one of these for Jita? How is a Xeon better? theres no faster CPU for number crunching than a I7.
Use some google-fu: the server blade mentioned at the start of this thread uses the server version (Xeon Westmere) of the Core i7.
Server cpu >>>>>>>>>>> consumer cpu.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
mechtech
SRS Industries SRS.
|
Posted - 2010.09.04 16:40:00 -
[210]
Originally by: Zeba
Originally by: Vomit Comet Edited by: Vomit Comet on 03/09/2010 21:34:28
Originally by: TriadSte Core I7 are right now the very best CPU known to mankind. Why not use one of these for Jita? How is a Xeon better? theres no faster CPU for number crunching than a I7.
Use some google-fu: the server blade mentioned at the start of this thread uses the server version (Xeon Westmere) of the Core i7.
Server cpu >>>>>>>>>>> consumer cpu.
The Xeons are re-branded i7s with additional QPI lanes for multiprocessor setups. They are effectively exactly the same as an i7, no better or worse.
|
|
hydlided
|
Posted - 2010.09.04 16:57:00 -
[211]
1200 people in jita... I think we can do much better then this guys :)
|
Rootax
Gallente International Research Labs Ltd.
|
Posted - 2010.09.04 17:33:00 -
[212]
Originally by: CCP Yokai Edited by: CCP Yokai on 03/09/2010 09:50:53 In response to questions about when do we do this for 0.0:
After Sunday we should have all the data we need to prove and measure the impact on Eve from:
Changes in Memory Latency Changes in Chip Architecture Changes in Cache Changes in Clock Rate Addition of Turbo Boost
When we have this we can much more easily spec the optimal system for Jita and similar systems and move on. This same server will be then used to benchmark fleet fights in 0.0
With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0
I hope that helps better explain the intentions of this testing.
Can't you wait for the Sandy Bridge architecture ? Single thread performances seems great on it. ---------------------------------- - I'm french, sorry if you don't understand me - |
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.04 19:03:00 -
[213]
Originally by: hydlided 1200 people in jita... I think we can do much better then this guys :)
Well this ongoing lack of topping out jitas population has certainly killed the theory that if you raise the pop cap in jita they will only fill it to capacity and then demand MOAR!
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Lanu
0utbreak
|
Posted - 2010.09.04 21:23:00 -
[214]
Really smooth experience so far.
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.05 04:30:00 -
[215]
Originally by: Lanu Really smooth experience so far.
Yeah. I'm also curious about how this new hardware will help out with general nullsec lag. Also now that we know that 1200ish is about the limit for a normal prime time jita population I wonder what fleet fights will eventually top out at?
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Hun Jakuza
Roving Guns Inc.
|
Posted - 2010.09.05 16:08:00 -
[216]
Edited by: Hun Jakuza on 05/09/2010 16:08:04
Originally by: CCP Yokai Edited by: CCP Yokai on 03/09/2010 09:50:53 In response to questions about when do we do this for 0.0:
After Sunday we should have all the data we need to prove and measure the impact on Eve from:
Changes in Memory Latency Changes in Chip Architecture Changes in Cache Changes in Clock Rate Addition of Turbo Boost
When we have this we can much more easily spec the optimal system for Jita and similar systems and move on. This same server will be then used to benchmark fleet fights in 0.0
With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0
I hope that helps better explain the intentions of this testing.
Jita and fleet fights not equal, so i think this it's an useless test for fleet fights. You need testing fleet fights. No hundreds drones in jita, no hundreds activated gun cycles etc. You know where needed reinforced nodes, when big fleet fights coming, just use there new hardwares there and check it , what happen there.
|
Stick Cult
|
Posted - 2010.09.05 16:10:00 -
[217]
Originally by: Hun Jakuza Edited by: Hun Jakuza on 05/09/2010 16:08:04
Originally by: CCP Yokai Edited by: CCP Yokai on 03/09/2010 09:50:53 In response to questions about when do we do this for 0.0:
After Sunday we should have all the data we need to prove and measure the impact on Eve from:
Changes in Memory Latency Changes in Chip Architecture Changes in Cache Changes in Clock Rate Addition of Turbo Boost
When we have this we can much more easily spec the optimal system for Jita and similar systems and move on. This same server will be then used to benchmark fleet fights in 0.0
With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0
I hope that helps better explain the intentions of this testing.
Jita and fleet fights not equal, so i think this it's an useless test for fleet fights. You need testing fleet fights.
It's not a test for fleet fights, it's a test for Jita. We know Jita's better now, but they want to know how much better, so they want to (almost) break it. Then he said they're going to have a node like they have now in Jita to do fleet fighting on, to test fleet fighting.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Hun Jakuza
Roving Guns Inc.
|
Posted - 2010.09.05 16:20:00 -
[218]
Originally by: Stick Cult
Originally by: Hun Jakuza Edited by: Hun Jakuza on 05/09/2010 16:08:04
Originally by: CCP Yokai Edited by: CCP Yokai on 03/09/2010 09:50:53 In response to questions about when do we do this for 0.0:
After Sunday we should have all the data we need to prove and measure the impact on Eve from:
Changes in Memory Latency Changes in Chip Architecture Changes in Cache Changes in Clock Rate Addition of Turbo Boost
When we have this we can much more easily spec the optimal system for Jita and similar systems and move on. This same server will be then used to benchmark fleet fights in 0.0
With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0
I hope that helps better explain the intentions of this testing.
Jita and fleet fights not equal, so i think this it's an useless test for fleet fights. You need testing fleet fights.
It's not a test for fleet fights, it's a test for Jita. We know Jita's better now, but they want to know how much better, so they want to (almost) break it. Then he said they're going to have a node like they have now in Jita to do fleet fighting on, to test fleet fighting.
Read it what they wrote, "This same server will be then used to benchmark fleet fights in 0.0 With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0" So Jita is not 0.0 fleetfights.
|
Ellissdee
|
Posted - 2010.09.05 17:27:00 -
[219]
Well ichiroya is now ****ed, I guess due to the "testing" at jita
|
THE SOP
|
Posted - 2010.09.05 17:55:00 -
[220]
Originally by: Ellissdee Well ichiroya is now ****ed, I guess due to the "testing" at jita
I cant even log into Ichy now, and there was only a 146 in local......
|
|
Valixithor
Caldari The Eidolon Synergy
|
Posted - 2010.09.05 18:52:00 -
[221]
Confirming Ichoriya is currently unplayable. About a 15 second latency.
Originally by: Yoravian
If by 'you guys are pretty nasty' you mean 'you guys are sure good at explaining game mechanics to guys like me', then yes. Yes we are.
|
Grez
Empire Assault Corp Dead Terrorists
|
Posted - 2010.09.05 19:01:00 -
[222]
Originally by: Ellissdee Well ichiroya is now ****ed, I guess due to the "testing" at jita
Jita runs on its own node. It can't affect other systems. ---
|
Kliump
|
Posted - 2010.09.05 19:02:00 -
[223]
Please remove the 1800 limit in Jita :)
|
BearUkraine
|
Posted - 2010.09.05 20:04:00 -
[224]
Remove limit 2k people pls!!!
|
Lanu
0utbreak
|
Posted - 2010.09.05 20:33:00 -
[225]
The glorious moment ;)
http://www.youtube.com/watch?v=4EZuKEuk9pk&fmt=22
|
Phantom Slave
Universal Pest Exterminators
|
Posted - 2010.09.05 20:47:00 -
[226]
Originally by: Hun Jakuza Read it what they wrote, "This same server will be then used to benchmark fleet fights in 0.0 With all of your help the test in Jita gets us the predictable load we need for baseline data collection and we move right into improving 0.0" So Jita is not 0.0 fleetfights.
I read that as, "We test jita with this hardware first, get a benchmark for how market hubs will perform, then we'll take the server over and reinforce a 0.0 node on Sisi or TQ and benchmark a 0.0 fight."
|
mrfg
Caldari Jnana Yoga Vera Cruz Alliance
|
Posted - 2010.09.05 21:06:00 -
[227]
My Drake was warping sideways in Jita... there were about 2000 people in there. lol
|
Cipher Jones
Minmatar
|
Posted - 2010.09.05 21:17:00 -
[228]
Heres my feedback on the new Jita hardware after coming to the event.
With 2000 people in local it was totally unplayable. less than 1 fps.
with 1600 people in local it was laggy but playable. Greatly reduced FPS but able to lock shoot and all of that.
I hope this helps. When you say "fanboi" try to picture a fat man doing burlesque with 2 big ass fans that say CCP on one and HTFU on the other. Because that dude is me. |
Mikk36
Caldari
|
Posted - 2010.09.05 21:23:00 -
[229]
Originally by: Cipher Jones Heres my feedback on the new Jita hardware after coming to the event.
With 2000 people in local it was totally unplayable. less than 1 fps.
with 1600 people in local it was laggy but playable. Greatly reduced FPS but able to lock shoot and all of that.
I hope this helps.
1950 people and I was still managing around 50-60 fps when orbiting the station at 30km. Specs: Q6600 @ 3.0Ghz, 8GB DDR2-6400 4-4-4-12 and HD4870 512MB unclocked, graphics set to highest in EVE, 1024x768 (alt's account). Locking time of other people varied between 1 to 4 seconds ( that is, the time between hitting 0 and getting a lock).
|
RentableMuffin
|
Posted - 2010.09.05 21:28:00 -
[230]
it was pretty decent even with 2000+ in system. granted I was just in a shuttle zoomed out all the way
should have started spawning npcs outside to get some added combat stress
|
|
mechtech
SRS Industries SRS.
|
Posted - 2010.09.05 21:40:00 -
[231]
The client has some serious problems with rendering high numbers of players. I have an i5 running @ 4.0ghz and a 5850 highly OCd as well with hardware monitors running. Looking in certain directions at Jita 4-4 caused me to drop to 5fps. Checking the hardware monitors my CPU no core was over 50% load and my GPU was at 60% use (had 4GB free memory as well).
In short, I had lots of power that the client wasn't using yet the game was running at 5fps. CCP needs to re-examine the client under high load situations to figure out what's going wrong with the hardware utilization.
As for basic module lag, pretty amazing for 2k people, but then again it ain't no fleet battle.
|
Mikk36
Caldari
|
Posted - 2010.09.05 21:53:00 -
[232]
Oh and, memory usage dropped by 1.2GB when I warped off the 4-4 station (64bit Win7)
|
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.09.05 21:53:00 -
[233]
2010-09-05: Jita Stress Test Gallery - 1979 to 2025 players!!! Jita Stress Test EVE-Files Gallery
If you prefer, open the direct links to EVE-FILES: SS01: 1997 Players @ 60FPS / Gate View SS02: 2003 Players @ 55FPS / 4-4 Station View SS03: 2025 Players @ 41FPS / 4-4 Station View SS04: 1979 Players @ 20FPS / 4-4 Station - Concord Fight SS05: 1992 Players @ 24FPS / 4-4 Station - Concord Overview SS06: Windows 7 x64 - Resources @ 5 EVE Clients
Feedback on lag will follow soon ;-)
___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
|
CCP Explorer
|
Posted - 2010.09.05 22:40:00 -
[234]
Originally by: Cipher Jones Heres my feedback on the new Jita hardware after coming to the event.
With 2000 people in local it was totally unplayable. less than 1 fps.
with 1600 people in local it was laggy but playable. Greatly reduced FPS but able to lock shoot and all of that.
I hope this helps.
What you are describing is client lag, in particular your video card and your video card driver were unable to keep up. Did you go to a planet or a moon or a less busy station than 4-4? How was the experience there?
Erlendur S. Thorsteinsson Software Director EVE Online, CCP Games |
|
|
CCP Explorer
|
Posted - 2010.09.05 22:42:00 -
[235]
Originally by: mechtech The client has some serious problems with rendering high numbers of players. I have an i5 running @ 4.0ghz and a 5850 highly OCd as well with hardware monitors running. Looking in certain directions at Jita 4-4 caused me to drop to 5fps. Checking the hardware monitors my CPU no core was over 50% load and my GPU was at 60% use (had 4GB free memory as well).
In short, I had lots of power that the client wasn't using yet the game was running at 5fps. CCP needs to re-examine the client under high load situations to figure out what's going wrong with the hardware utilization.
As for basic module lag, pretty amazing for 2k people, but then again it ain't no fleet battle.
Sounds like your video card wasn't able to keep up. How did you configure the overview, and more importantly, how did you configure the in-space brackets?
Erlendur S. Thorsteinsson Software Director EVE Online, CCP Games |
|
|
CCP Explorer
|
Posted - 2010.09.05 22:45:00 -
[236]
BTW, Jita peaked at 2023 pilots between 19:56 and 20:16 GMT.
Erlendur S. Thorsteinsson Software Director EVE Online, CCP Games |
|
mechtech
SRS Industries SRS.
|
Posted - 2010.09.05 22:58:00 -
[237]
Originally by: CCP Explorer
Originally by: mechtech The client has some serious problems with rendering high numbers of players. I have an i5 running @ 4.0ghz and a 5850 highly OCd as well with hardware monitors running. Looking in certain directions at Jita 4-4 caused me to drop to 5fps. Checking the hardware monitors my CPU no core was over 50% load and my GPU was at 60% use (had 4GB free memory as well).
In short, I had lots of power that the client wasn't using yet the game was running at 5fps. CCP needs to re-examine the client under high load situations to figure out what's going wrong with the hardware utilization.
As for basic module lag, pretty amazing for 2k people, but then again it ain't no fleet battle.
Sounds like your video card wasn't able to keep up. How did you configure the overview, and more importantly, how did you configure the in-space brackets?
I had my standard overview setup, so all brackets were on, and every ship was showing up in the overview. If I turned my view 30 degrees I would go 30-40fps, but when I centered the ships onto the screen (at the same zoom level) fps simply tanked.
The weird thing though was my CPU utilization was at 50% and my GPU was at 60% use. Unless the FPS is capped the engine should always be using either 100% of the CPU or 100% of the GPU, no? (in my experience Eve is very GPU bound under normal conditions). Actually, Mafia 2 with APEX enabled did the same thing on occasion, so maybe there are fringe cases where this can occur. Still, a well optimized client should always be using all of the resources available to it under ideal circumstances.
If you guys do any in-house testing under fleet fight conditions, take a looks at CPU/GPU utilization, because tbh the GPU should be pegged at 100% use during any situation where you're running at 5fps.
(system: 5850 w/ 10.7 drivers, i5 @ 4.0ghz, 8gb RAM, win7-64)
|
Cipher Jones
Minmatar
|
Posted - 2010.09.05 23:05:00 -
[238]
Originally by: CCP Explorer
Originally by: Cipher Jones Heres my feedback on the new Jita hardware after coming to the event.
With 2000 people in local it was totally unplayable. less than 1 fps.
with 1600 people in local it was laggy but playable. Greatly reduced FPS but able to lock shoot and all of that.
I hope this helps.
What you are describing is client lag, in particular your video card and your video card driver were unable to keep up. Did you go to a planet or a moon or a less busy station than 4-4? How was the experience there?
I specifically ran as little resources on my client as I could. I ran in safe mode 1024x768 with nothing else open. I tried to open 'ctrl shift alt m' and 'ctrl F' in several locations including the station but was unable to. The lag on the station was considerably less than in space however. I managed to dock and undock in under 45 seconds a couple of times. I did not notice less lag anywhere in the system itself. I managed to enter the system, fly to the station, fly away from the station, fly back, dock, undock. I only went back and forth between the perimeter gate, there were significantly less people there than 4-4.
So when I was at that gate with 200 people, it may have been less laggy than the 4-4 station. But it was so laggy I would not engage in a fleet battle by any means. With 1600 people in the system all balled up in front of the station I would have been capable of combat.
Also I was unable to turn brackets on/off when the system was overloaded, but it started working around 1800 people in the system.
I hope this is more concise and helps.
When you say "fanboi" try to picture a fat man doing burlesque with 2 big ass fans that say CCP on one and HTFU on the other. Because that dude is me. |
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.09.05 23:20:00 -
[239]
OK, my Jita Stress Test Feedback Post:
Hardware & OS Info Windows 7 x64 Intel i7 Core 920 @ 3.6GHz (O/C) ATI RADEON HD5870 Toxic 2GB GDDR5 (Catalyst 10.8) 12GB 1800MHz DDR3 (6 x 2GB @ 3 channels) System Drive: 4 x Intel X25-E 32GB SSD (Adaptec 5405 RAID) Internet connection: 100Mbps Upload/Download Fiber / 54ms ping to 87.237.38.200 / 14 hops (5 internal within my corporate LAN)
EVE Clients Info EVE Client 1: Peak Working Set (Memory) @ 1.54GB - Outside 4-4, Orbiting 15km EVE Client 2: Peak Working Set (Memory) @ 1.48GB - Outside 4-4, Orbiting 30km EVE Client 3: Peak Working Set (Memory) @ 1.36GB - Jita @ Perimeter Gate EVE Client 4: Peak Working Set (Memory) @ 0.54GB - Inside 4-4 EVE Client 5: Peak Working Set (Memory) @ 0.97GB - Inside 4-4
Total Resource Utilization: 9.76GB RAM, CPU @ 12-20% Average / 70-80% short peaks
EVE Clients Graphics Settings Resolution : Full Screen @ 1920 x 1200 Color Depth : 24 bit color with 8-bit alpha Depth/Stencil : 32 bit Z-buffer with 8-bit stencil Present Interval : Interval Defualt (Vsync = ON)
Shadows Enabled Resource Cache Enabled HDR Enabled
Bloom: High Shader Quality: High Texture Quality: High LOD Quality: High
All effects ON (except Drone Models, which were OFF)
Feedback @1600 Players Market Buy (various T1 items): NO LAG Market Sell (various T1 items): NO LAG Contracts Buy (BPO): NO LAG Market History Graphs (Fast Clicking in the Market Browser): NO LAG Undock / Dock: Normal Speed, NO LAG loading the grid Local Chat: NO LAG Corp Chat: NO LAG Frame Rate: 55-60FPS Inside / Outside Station
@1800 Players Market Buy (various T1 items): NO LAG Market Sell (various T1 items): NO LAG Contracts Buy (BPO): NO LAG Market History Graphs (Fast Clicking in the Market Browser): NO LAG Undock / Dock: Normal Speed, some LAG loading the grid (2-5 seconds black screen) Local Chat: NO LAG Corp Chat: NO LAG Frame Rate: 40-60FPS Inside / Outside Station
@2000 Players Market Buy (various T1 items) : 1-2 seconds LAG Market Sell (various T1 items): 5-10 seconds LAG Contracts Buy (BPO): N/A (not tested) Market History Graphs (Fast Clicking in the Market Browser): NO LAG Undock / Dock: Slow Speed, some LAG loading the grid (10-15 seconds black screen) Local Chat: 1-2 seconds lag (not every time) Corp Chat: 1-2 seconds lag (not every time) Frame Rate: 40-60FPS Inside / 15-45FPS Outside Station (10-15 FPS when CONCORD is in the overview)
Hope this helps! ___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Samulus
Caldari
|
Posted - 2010.09.05 23:47:00 -
[240]
My client crashed when Concord loaded grid in force at station 4-4. CPU usage was at 60% (dual core)and memory usage was at 97% (of 3 Gigs). NOTE: My graphic settings had not been optimized for Jita blob. Sam |
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.06 00:45:00 -
[241]
2023 in jita? Reverse psychology ftw!
Thank me in faction ships and isk plx.
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Jita Alt666
|
Posted - 2010.09.06 01:00:00 -
[242]
Originally by: Zeba 2023 in jita? Reverse psychology ftw!
Thank me in faction ships and isk plx.
Thank you?
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.06 01:08:00 -
[243]
Originally by: Jita Alt666
Originally by: Zeba 2023 in jita? Reverse psychology ftw!
Thank me in faction ships and isk plx.
Thank you?
Your welcome!
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Grideris
Gallente Fleet Coordination Commission
|
Posted - 2010.09.06 06:00:00 -
[244]
Originally by: CCP Yokai
Originally by: Textret
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
We will require BOTH pictures.
Make it 2000 and I'll post both!
Right. We got over 2000, now where are the videos?
|
BearUkraine
|
Posted - 2010.09.06 09:15:00 -
[245]
Originally by: Grideris
Originally by: CCP Yokai
Originally by: Textret
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
We will require BOTH pictures.
Make it 2000 and I'll post both!
Right. We got over 2000, now where are the videos?
Much interesting. Where is the report and what is the next CCPs step?)
|
Mashie Saldana
BFG Tech
|
Posted - 2010.09.06 10:15:00 -
[246]
So now when Jita can handle 2k+ people isn't it about time to move all assets there back to Yulai and reintroduce the old highways?
|
sir gankalot
|
Posted - 2010.09.06 12:22:00 -
[247]
Witnessed it go over 2k, sitting in station spinning my ship was still fine. Did not do anything else but was also able to caht in my usual channels without lag. God knows that has not always been the case
|
Zeba
Minmatar Honourable East India Trading Company
|
Posted - 2010.09.06 18:56:00 -
[248]
Originally by: sir gankalot Witnessed it go over 2k, sitting in station spinning my ship was still fine. Did not do anything else but was also able to caht in my usual channels without lag. God knows that has not always been the case
And the coolest thing is that apparently the hardware that is being used isn't the absolute best they could get as this is simply a test run to see if it would be worth it.
Welp, ccp get to ordering the really good stuff nao!!
Originally by: CCP Oveur My milkshake brings all the boys to the yard.
Originally by: Ryhss There is no paranoia in Eve, everyone is out to get you....
|
Hawk TT
Caldari Bulgarian Experienced Crackers Circle-Of-Two
|
Posted - 2010.09.07 15:34:00 -
[249]
Hello, CCP? Is anybody out there?
Could you share with us some results from the Jita Stress Test? As you already know, "silence kills trust" ___________________________________ Science & Diplomacy Manager @ BECKS Circle-of-Two |
Frederick Corvinus
|
Posted - 2010.09.07 18:04:00 -
[250]
Least give them a few days to work on the data... I'd rather they spend some time crunching through the results from the testing and then come back to use with accurate and meaningful data rather than rushing it out of the door to appease people.
|
|
Stick Cult
|
Posted - 2010.09.07 18:46:00 -
[251]
Originally by: Hawk TT Hello, CCP? Is anybody out there?
Could you share with us some results from the Jita Stress Test? As you already know, "silence kills trust"
Give them a few days to make some colorful charts.
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Jita Alt666
|
Posted - 2010.09.08 07:03:00 -
[252]
4 days and nothing. I can make pretty charts at the rate of 1 every 15 minutes (and I know I am slow) so thats 4 an hour during a working day thats 32 pretty charts and over 4 days thats 128 charts.
To be honest I would be happy with 4 charts (one a day) and suitable analysis. Although even a simple post a single developer would go a long way.
|
|
CCP Yokai
|
Posted - 2010.09.08 11:08:00 -
[253]
Still collecting data and writing up the details. We are on it and plan to get something out in a Dev Blog format A.S.A.P
Give us just a touch longer please.
Thanks!
|
|
Frederick Corvinus
|
Posted - 2010.09.08 11:09:00 -
[254]
Originally by: Jita Alt666 4 days and nothing.
Uhhh 4 days? It was Sunday night and it's now Wednesday morning. That's 2 1/2 working days or so at most.
|
Squid Prime
|
Posted - 2010.09.08 13:42:00 -
[255]
Originally by: CCP Yokai
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
Make it 2000 and I'll post both!
You'd better not forget those pics Yokai! Gief sexy pics with the beast that was able to do this!
...to prevent any confusion, I'm referring to the server
|
Stick Cult
|
Posted - 2010.09.08 19:33:00 -
[256]
Originally by: Squid Prime
Originally by: CCP Yokai
Originally by: CCP Yokai I thought if we posted about new hardware we'd get a few more people in local testing. Tell you what... we get this puppy above 1800 tonight I will post a vid of a dev dancing in a cabinet in the new London datacenter.
Or if you choose a server pic :)
Make it 2000 and I'll post both!
You'd better not forget those pics Yokai! Gief sexy pics with the beast that was able to do this!
...to prevent any confusion, I'm referring to the server
I'd accept sexy pics of Yokai...
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
|
CCP Yokai
|
Posted - 2010.09.09 11:58:00 -
[257]
I can't promise sexy... but there are pics of... Me, CCP Explorer, servers, graphs and some words... Stay tuned.
|
|
Stick Cult
|
Posted - 2010.09.09 19:29:00 -
[258]
Originally by: CCP Yokai I can't promise sexy... but there are pics of... Me, CCP Explorer, servers, graphs and some words... Stay tuned.
I. Must. See. This.
(I'm sure they are sexy... there's servers involved! )
Originally by: CCP Tuxford my bad. Rest assured I'm being ridiculed by my co-workers.
|
Alara IonStorm
Agent-Orange
|
Posted - 2010.09.10 07:31:00 -
[259]
Originally by: CCP Yokai I can't promise sexy... but there are pics of... Me, CCP Explorer, servers, graphs and some words... Stay tuned.
I was promise a half naked dancing Yokai, don't Dispoint me!
Can it be infront of the new server!
-- Tactical Responder who is Organized and a Leading-edge Linguist |
|
CCP Yokai
|
Posted - 2010.09.10 16:24:00 -
[260]
New Dev Blog located here: http://www.eveonline.com/devblog.asp?a=blog&bid=792
|
|
|
Jita Alt666
|
Posted - 2010.09.11 01:57:00 -
[261]
Dev Blog Summary:
Jita performance has reached a new benchmark level, testing of fleet fight performance will on same hardware will now commence.
I do like the sound of that.
|
|
CCP Yokai
|
Posted - 2010.09.11 11:49:00 -
[262]
Originally by: Jita Alt666 Dev Blog Summary:
Jita performance has reached a new benchmark level, testing of fleet fight performance will on same hardware will now commence.
I do like the sound of that.
Me too! I'm pretty excited to get a good fight on the new test hardware.
Thanks!
|
|
Squid Prime
|
Posted - 2010.09.13 08:19:00 -
[263]
Originally by: CCP Yokai New Dev Blog located here: http://www.eveonline.com/devblog.asp?a=blog&bid=792
Awesomesauce :D. Thank you for all the effort!
|
illford baker
Caldari Provisions
|
Posted - 2010.09.15 22:04:00 -
[264]
Originally by: CCP Yokai
Originally by: Jita Alt666 Dev Blog Summary:
Jita performance has reached a new benchmark level, testing of fleet fight performance will on same hardware will now commence.
I do like the sound of that.
Me too! I'm pretty excited to get a good fight on the new test hardware.
Thanks!
enterprise techs and pc hardware enthusiast are not so different, we both want to stress our hardware to its limits. i remember acting the same way when i got my new i7, or GTX 465.
|
HeliosGal
Caldari
|
Posted - 2010.09.21 11:42:00 -
[265]
should we suicide in jita more often to test the node and discourage ppl from trading there ?
|
|
CCP Navigator
C C P C C P Alliance
|
Posted - 2010.09.25 16:51:00 -
[266]
Thank you all so much for your feedback. This thread will be unstickied now but please feel free to keep your comments coming
Navigator Senior Community Representative CCP Hf, EVE Online
|
|
|
|
|
Pages: 1 2 3 4 5 6 7 8 9 :: [one page] |