|
Author |
Thread Statistics | Show CCP posts - 15 post(s) |
Zagdul
Gallente Clan Shadow Wolf Fatal Ascension
|
Posted - 2011.07.28 06:15:00 -
[1]
Sweet tool and thanks for sharing.
I have to say that on reinforced nodes, fleet fights have actually become much more responsive.
I would like to ask if there's anything share concerning development on non-reinforced nodes or potential ideas you've had to assist combating lag when we have spur of the moment encounters. As you know, sometimes your enemy can use lag to gain an advantage and intentionally not requesting for reinforcement has been used as an offensive tactic a few times.
Again, thanks for all your hard work. For those of us who participate in large fleet fights, we've seen a difference.
|
Zagdul
Gallente Clan Shadow Wolf Fatal Ascension
|
Posted - 2011.07.28 15:45:00 -
[2]
Edited by: Zagdul on 28/07/2011 15:47:31 Edited by: Zagdul on 28/07/2011 15:45:19
Originally by: CCP Veritas
Originally by: Zagdul I would like to ask if there's anything share concerning development on non-reinforced nodes or potential ideas you've had to assist combating lag when we have spur of the moment encounters. As you know, sometimes your enemy can use lag to gain an advantage and intentionally not requesting for reinforcement has been used as an offensive tactic a few times.
A lot of the things we've been fixing help non-reinforced nodes the same, if not more, than dedicated nodes. I've noticed the performance between those two setups converge a bit, so things are getting better on that front as well.
Thanks for the answer.
One more if you get the time.
Does Tranq drop unused systems and resources from a node in "real-time" when load increases?
it seems to me that, possibly slamming on "Traffic control" for "common routes" to hot systems and booting a few systems off a node would help.
I believe that when spikes in a node happen, (VFK last week for example) that there must be something the server can do in order to assist the fight out.
|
Zagdul
Gallente Clan Shadow Wolf Fatal Ascension
|
Posted - 2011.07.31 16:02:00 -
[3]
Edited by: Zagdul on 31/07/2011 16:04:50 Edited by: Zagdul on 31/07/2011 16:04:27
Originally by: CCP Veritas
Originally by: Zagdul Does Tranq drop unused systems and resources from a node in "real-time" when load increases?
Not really, no. Unused systems are exceptionally low load though, being unused, so not much to be gained there. Moving systems with people in them causes those people to get disconnected, which is something we don't want an automatic system to do for us. We sometimes do those moves manually when we believe it will help more than it hurts.
Originally by: Zagdul I believe that when spikes in a node happen, (VFK last week for example) that there must be something the server can do in order to assist the fight out.
I agree, and that's why I'm working on Time Dilation
The thought I had was if someone decides to use an unused system as a mid point for supers, or one of the unused systems is a choke for getting people to the encounter.
By removing it from the node, if/when a fleet who is trying to catch up goes through, it would no longer effect the encounter happening.
Also, thanks for the work on time dilation.
|
Zagdul
Gallente Clan Shadow Wolf Fatal Ascension
|
Posted - 2011.07.31 16:08:00 -
[4]
Originally by: Vincent Jarjadian
The current CPUs are, according to the wiki, two dual core 3.3GHz per blade... If CCP used one or two newer X5690 Xeon CPUs which are 6 cores running at 3.46GHz... with an extra boost from turbo mode, these servers could handle the biggest battles, or even 12 of them simultaneously. Also... they can handle a lot of memory so that could be an advantage as well. You could even overclock them to or past 4GHz and get even more performance (I've seen some on searches running 4.5GHz)... It's not cheap but it could be beneficial...
Also... Some of the new sandy bridge CPUs are reaching 5GHz... are you aware of any plans to produce server grade versions of these?
The CPU's on tranq are more powerful PER CORE than the new Xeons.
|
Zagdul
Gallente Clan Shadow Wolf Fatal Ascension
|
Posted - 2011.08.02 07:29:00 -
[5]
Edited by: Zagdul on 02/08/2011 07:33:37
Originally by: DeadlyThreat
Originally by: Zagdul The CPU's on tranq are more powerful PER CORE than the new Xeons.
Are you saying that the Wolfdale architecture is faster than the Westmere-EP architecture?
No.
Faster != better per core performance.
Yes, the new Xeons 32nm 10 core tech is gorgeous with 30mb of cache for multi tasking.
however, for single threaded tech and tasking the way python handles it, right now the only thing that can match it is the cell tech or new GPU's nvidia are putting out on a per core speed and gflops. Both would require a full code rewrite of the server engine.
I'm very sure people will correct me for being a moron here if I'm wrong, but this has been stated in multiple dev blogs as to why sometimes the "fast" processor doesn't translate to "best" for tranquility as the servers require a specific type of processor, not always the one that benchmarks nicely.
|
|
|
|