| Pages: 1 [2] :: one page |
|
|
| Author |
Thread Statistics | Show CCP posts - 7 post(s) |
Darth Sith
Genbuku.
Nulli Secunda
7
   |
Posted - 2011.12.16 22:44:00 -
[31] - Quote
Well, at the time that CCP was discussing it at Fanfest last year, Intel had this system under really tight raps.
Recently however they have relaxed the NDA and we are now allowed to publicly disclose the details on the box in general terms.
Here is the box in question : Monster Toy
and it is making use of Intel's beast of a proc running at 4.4 Ghz (air cooled) aka : Intel Xeon X5698 4.4 GHz
Take a wild guess who I work for :)
Cheers. |
H3llHound
Nex Exercitus
Raiden.
3
|
Posted - 2011.12.17 00:23:00 -
[32] - Quote
Nice read but can you make(and show us :D) a server load graph of the fight in L-6BE1 on the 21.11.2011 from 22:00 to 24:00? Afaik that system wasnt reinforced and I would be interested how the servers kept up as it was lagging qiuet a bit for some time. pretty please :)
P.s. I will give you a 'like' if this happens ;) |
Mioelnir
Cataclysm Enterprises
Ev0ke
34
|
Posted - 2011.12.17 01:35:00 -
[33] - Quote
So, looking at the tama graph... its running on the supernode with +80% over regular dedicated nodes.
Which would put a regular node's 100% mark at around 55% of the supernode graph. So, had this not been on the node it was, there would have been constant 100% utilization for an hour between 19:30 and 20:30? |
Klandi
Consortium of stella Technologies
35
|
Posted - 2011.12.17 01:46:00 -
[34] - Quote
Great info - thanks for the update
Of interest to me - how much do you consider is the lag apportioned to the server hardware vs the client hardware? Ultimately if we had the best hardware on the market for the client with your current server configuration then what would be achievable in a fight. 1000 each side in system?
I won't hold you to any guesses :)
I am aware of my own ignorance and have checked my emotional quotient - thanks for asking |
Rognin
No.Mercy
Merciless.
0
|
Posted - 2011.12.17 02:06:00 -
[35] - Quote
LOL EKWB
Seriously, you guys went about that one wrong...
Edit: And Liquid was DI water + PT nuke, and if it weren't for EK, then a silver kill coil would have done the job.
At least put that old hardware to folding?
Phase change or a TEC WB would have done a better job at cooling than your air solution. |
ibnubis
Belligerent Enterprises
0
|
Posted - 2011.12.17 03:20:00 -
[36] - Quote
You guys are doing some great tuning. Very Impressive.
Have you guys also bumped the network along with the processor technology? If your servers get chatty to each other under load and you're using Ethernet, the TCP stack will load up your cores especially with lots of processes running. Or do you use InfiniBand? IB will scale linearly with core count and has no TCP stack. Very low latency, (700nS) and 56Gb/sec. IB can sometimes solve CPU headroom and scaling issues. As long as the application is dependent on server to server communication.
|
Sinq Arnolles
Center for Advanced Studies
Gallente Federation
2
|
Posted - 2011.12.17 04:57:00 -
[37] - Quote
2 fleets fights I have been in recently have lagged like crazy.
The worst one was http://eve-kill.net/?a=kill_related&kll_id=11742712
As a logistics pilot it took about a min to lock anything and a good 30 seconds to activate / deactivate a module and when the FC did a fleet warp it took 10 mins for it to actually start doing it. Sometimes your ship would break orbit and go in a random direction and it would just keep going for a few mins before it responded to your commands to go somewhere. Probing also took at-least 1 min to work.
http://eve-kill.net/?a=kill_related&kll_id=11762630 Similar situation for this one but it was slightly better.. still pretty impossible for logi to do anything though. |
Aineko Macx
Royal Amarr Institute
Amarr Empire
31
|
Posted - 2011.12.17 07:17:00 -
[38] - Quote
How is the fundamental research Jacky (CCP Warlock) is doing into Python multithreading, the GIL, etc. coming along?
Btw, I approve of client side performance improvements. Get them brackets performance fixed plz! |
Darth Vapour
School of Applied Knowledge
Caldari State
0
|
Posted - 2011.12.17 08:41:00 -
[39] - Quote
Aineko Macx wrote:How is the fundamental research Jacky (CCP Warlock) is doing into Python multithreading, the GIL, etc. coming along?
According to her LinkedIn profile CCP Warlock no longer works at CCP.
|
Camios
Tribal Liberation Force
Minmatar Republic
26
|
Posted - 2011.12.17 09:24:00 -
[40] - Quote
These are awesome news (this is the meaning of awesome).
I have some questions:
How much does the Frankenstein server cost, compared to a "normal" one? That is, how many such blades can we expect to have in place on tranquillity in the future, considering that TiDi will greatly reduce the impact of "lag" on game mechanics thus reducing the need of extreme performance hardware?
Slightly out of topic:
Is there any room for improvements in the jump and dock-undock procedure? In the event on dec the 8th TiDi reached the 10% limit (is this correct?) during a mass jump and a mass dock. And it seems that the load generated by these actions is far heavier than the one generated by usual fighting, even missile spam (right?). This could theoretically be exploited by a side in combat since a side with lots of ships could, by keeping jumping in and out of a system, resurrect the old lag monster.
In the worst case, how many jumps can a node handle per second without lagging behind without time dilating? And how many warpins? And is it correct that with TiDi at 10% that number will be multiplied by 10? With these numbers and considering the duration of the (dilated) session change timer one can estimate how many people are necessary to do the exploit I described above. |
|
|
Kerdrak
D00M.
Northern Coalition.
7
|
Posted - 2011.12.17 10:36:00 -
[41] - Quote
Change the EVE client to work as distributed computing so Jita traders and AFK people make the dirty job. |
FlinchingNinja Kishunuba
Perkone
Caldari State
12
|
Posted - 2011.12.17 11:57:00 -
[42] - Quote
Darth Sith wrote:Well, at the time that CCP was discussing it at Fanfest last year, Intel had this system under really tight raps. Recently however they have relaxed the NDA and we are now allowed to publicly disclose the details on the box in general terms. Here is the box in question : Monster Toyand it is making use of Intel's beast of a proc running at 4.4 Ghz (air cooled) aka : Intel Xeon X5698 4.4 GHz Take a wild guess who I work for :) Cheers.
Looks pretty damn good, unfortunately I work for a competing manufacturer so errrrr.... forget what I just said ;)
Thinking I need to find an excuse to get a crazy OC cpu for our R&D guys, 4.4GHz is pretty tame though understand that you would have trouble getting a reliable rig at a higher clock. Then again buy two and pop in to HA :-D
|
Tiger's Spirit
Troll Hunters INC.
27
|
Posted - 2011.12.17 12:10:00 -
[43] - Quote
Mioelnir wrote:So, looking at the tama graph... its running on the supernode with +80% over regular dedicated nodes.
Which would put a regular node's 100% mark at around 55% of the supernode graph. So, had this not been on the node it was, there would have been constant 100% utilization for an hour between 19:30 and 20:30?
So, looking the CCP GingerDude post.
https://forums.eveonline.com/default.aspx?g=posts&m=413112#post413112
"Well, the client lag was pretty bad in the blob..." |
Pinky Denmark
The Cursed Navy
Important Internet Spaceship League
48
|
Posted - 2011.12.17 13:02:00 -
[44] - Quote
I would like to know how that would go on a normalo not reinforced cluster... Especially with time Dillation in action |
Mortis vonShadow
Balanaz Mining and Development Inc.
387
|
Posted - 2011.12.17 22:12:00 -
[45] - Quote
If you're going to use liquid to cool the cpu and motherboard, use mineral spirits. Its doesn't conduct electricity and is an excellent heat transfering medium.
Some days you're the bug, and some days your the windscreen.
-á-á-á-á-á-á-á-á-á-á-á-á-á-á-á-á-á And some days, you're just a man with a gun. |
Inspiration
Focused Radical Energy Engineering
0
|
Posted - 2011.12.17 22:57:00 -
[46] - Quote
Mortis vonShadow wrote:If you're going to use liquid to cool the cpu and motherboard, use mineral spirits. Its doesn't conduct electricity and is an excellent heat transfering medium.
Not that i know the stuff you talking about, but cooling liquids need more properties besides thermal conductivity and electric insulation. Just to make an example, you don't want corrosion, you don't want heat to create boiling micro pockets. The later freeing gass from the liquid that would very locally kill thermal conductivity. Just saying things are more complex :)
|
|
CCP DeNormalized
C C P
C C P Alliance
0
|
Posted - 2011.12.18 00:02:00 -
[47] - Quote
we thought about trying to get some co-workers to drink it, saying it was Quafe Zero, but someone talked us out of it for some crazy reason (was likely for the best).
But the liquid is ThermoChill EC6 UV Blue and we had 2 independent loops. CPU was on it's own and the motherboard/ram were on the other loop.
Stats from the web:
GÇó Pre-mixed with no need for additional additives
GÇó Electrical Conductivity <2-¦S/cm at 24C *
GÇó Non-Toxic and RoHS CompliantGÇ¿
GÇó Freezing Point - 6C
GÇó 85% Bio-degradable Within 30 DaysGÇ¿
GÇó Shelf Stable UV Reactive Dyes of up to 2 yearsGÇ¿
GÇó EC6 is a true Eco Friendly Product
Also, while we had a ton of fun building this machine, others didn't quite enjoy it as much :) Our area backs onto the the maintenance guys room and they had a few spare tables where we could setup for the build. And even though they're no strangers to loud noise this thing was another level all together!
We had 6 insanely loud fans pulling air in/across the dual 320mm rads and had it running benchmarks for days. Before long we started to fear they would sabotage the project in order to protect their hearing/sanity :)
(Each of the fans put out over 200 CFM @ 6000 RPM and over 60 dB) |
|
Louis deGuerre
The Dark Tribe
Against ALL Authorities
100
|
Posted - 2011.12.19 11:01:00 -
[48] - Quote
Loving these tech blogs.
Keep up the good work
*cried at Old Yeller reference* Poor dog :_(
FIRE FRIENDSHIP TORPEDOES ! |
Sessym
Superstructure Exposure Service
19
|
Posted - 2011.12.19 11:27:00 -
[49] - Quote
CCP Red Button wrote:Doggy Dogwoofwoof wrote:FIRST So, just what IS the coolant for that liquid cooler?  Why, I would have thought that was obvious. Hamster juice ofc 
Juice made from the dead Hamsters? WTB Red Pill
Good to finally see how all those small improvements kick in in one place.
_________ ,,,,,,,,,,,,,;;;;;####;;-------======-]> --,,,,,,,,,....
//_###_________------;;;;;;;;;;;;'''----======-]> --'''''""""
//_____/ ------- |
Jim Orland
Furyan Federation
0
|
Posted - 2011.12.19 18:16:00 -
[50] - Quote
Why is the hardware such a secret? The processor of which you speak is public knowledge.
Edit: As are at least two of the servers that I can buy it in... |
|
|
DeBingJos
T.R.I.A.D
152
|
Posted - 2011.12.20 14:18:00 -
[51] - Quote
Technical blog = best blog
Fix FW ! |
Christian Schneider
3
|
Posted - 2011.12.20 17:15:00 -
[52] - Quote
In large fleet fights FPS generally drops below an acceptable amount if I zoom in while having graphic settings maxed. What hardware do I have to buy to keep FPS high in those situations? My Computer has 8GB RAM, a Quad Core CPU @2500MHz and an ATI Radeon 5780. I guess the graphic card is the problem? Or is it because EVE is limited to 2GB RAM? Is it possible at all to have high FPS in those situations even with the latest and best hardware? |
Rektide
School of Applied Knowledge
Caldari State
0
|
Posted - 2011.12.27 02:42:00 -
[53] - Quote
just because something has a name (X5698) doesn't mean it's publicly available or that a customer is allowed to repeat what it is they're using. it's very possible ccp is contractually obliged to be coy about a processor the world got wind of 9 months ago but that isn't in public channels and they have.
pretty painful how intel hands out cheap fast overclocks like crazy on air gear, then makes the server side people wait a year for parity. i'd always be tempted to buy & overclock fast failing servers. good luck adding HA or a quorum mode to the server. it'd be curious to note how determinstic stackless was, whether multiple instances of the server, handed the same inputs, would produce the same outputs, such that if one cheap super overclocked desktop went down another system in lockstep could take over, was already doing the same computation.
anyways, that's my boring theorycraft. night!
good luck seeing what your november 2010 fleet fights look like, equipped with westmere-ep cores and all your improvements. 800 person dev test is a good start, i'll keep my fingers crossed for your 1000+ person fight with caps & drones. who knows, get any more GHz out of those cores and maybe you wont need the time dialation feature, right? ;) |
Rektide
School of Applied Knowledge
Caldari State
0
|
Posted - 2011.12.27 02:59:00 -
[54] - Quote
Hey, you've had StacklessIO already for a while. That was offloading network IO to other cores right? Have these network cores any kind of utilization during Tama type testing or fleet fights?
And what about CarbonIO / BlueNet? Are there tasks other than network IO that you've offloaded out of the GIL lock that are happening in threads other than the main stackless event loop thread? What kind of CPU usages are you seeing out of this, what subsystems have given you the most payoff for offloading?
Gl ccp. Report back.
|
|
|
| |
|
| Pages: 1 [2] :: one page |
| First page | Previous page | Next page | Last page |