Forged Alliance Forever Forged Alliance Forever Forums 2015-04-10T16:17:27+02:00 /feed.php?f=42&t=9751 2015-04-10T16:17:27+02:00 2015-04-10T16:17:27+02:00 /viewtopic.php?t=9751&p=98162#p98162 <![CDATA[Re: Show CPU model in lobby]]>
Aurion wrote:
E8400-CV wrote:Did you ever test it? Run Furmark @ Extreme Burn-in while at the same time also having some torture test load up the CPU. Then monitor CPU clock frequency and see if the same frequency persists for 20 minutes or if it drops. And if that frequency even is full turbo.


That load is way heavier than FA load, so really not representative. It would be nicer to have a better benchmark though, but I think to test the actual thermals you have to let it run longer (and it's too long to run when you open the lobby).


Also, it is. My previous laptop was the G74SX, that one's predecessor. That entire line of notebooks has an incredibly beefy cooling system perfectly capable of maintaining 100% load across CPU and GPU even when overclocked. The same goes for the larger Alienware machines, the larger Clevo chassis (Which are used by most custom laptop builders, are inexpensive, and VERY common among laptop gamers), and my own MSI GT72 and its brethren from that company.

The fact is that gaming laptops capable of maintaining full turbo, or even overclocks, without even going past 70 degrees, are fairly common, don't have to be stupidly expensive (Though some are), and in a community like this there will be a lot of them around. The CPU model is used in them, and in £400 ultrabooks, making it a useless metric because you cannot tell what the cooling is like.

The days when laptop cooling on proper dedicated gaming machines couldn't keep up with the hardware are long, long gone I'm afraid. Time to squash this myth.

Statistics: Posted by IceDreamer — 10 Apr 2015, 16:17


]]>
2015-04-10T13:26:09+02:00 2015-04-10T13:26:09+02:00 /viewtopic.php?t=9751&p=98153#p98153 <![CDATA[Re: Show CPU model in lobby]]>
E8400-CV wrote:
Did you ever test it? Run Furmark @ Extreme Burn-in while at the same time also having some torture test load up the CPU. Then monitor CPU clock frequency and see if the same frequency persists for 20 minutes or if it drops. And if that frequency even is full turbo.


That load is way heavier than FA load, so really not representative. It would be nicer to have a better benchmark though, but I think to test the actual thermals you have to let it run longer (and it's too long to run when you open the lobby).

Statistics: Posted by Aurion — 10 Apr 2015, 13:26


]]>
2015-04-10T11:41:56+02:00 2015-04-10T11:41:56+02:00 /viewtopic.php?t=9751&p=98142#p98142 <![CDATA[Re: Show CPU model in lobby]]>
Aurion wrote:
E8400-CV wrote:No, I would enable the option to kick if it slows down bad and welcome you in.

Haswell CPU's run a bit cooler than previous generations. If I run my i7-2860QM full-bore, it will run a lot hotter.


So what is this function for? You only need the one from your other topic.


To call people on their bullshit; the 99% that lies about high cpu score. Strictly; yes, the suggestion from the other topic is better. I guess it boils down to politics, where if that would not make it, this would do as a (lousy) partial solution instead.

Also; current cpu bench does not correlate that much with actual sim speed ability. My 2008 Core 2 Duo E8400 gets a CPU rating of 250, my 2011 Core i5 2500k gets 220. Late-setons it's more a comparison of -2 vs +1 though.

Zoram wrote:
E8400-CV wrote:Decent laptops for FA gaming are unicorns.


I must be playing with a unicorn.
(the unicorn being an ASUS ROG G75VW for over two years ago (i7 -3610QM @2.30Ghz - 8GB RAM)
Not to say it never lags, as I don't often end up in over 60 minutes seton games, but so far, so good.


Did you ever test it? Run Furmark @ Extreme Burn-in while at the same time also having some torture test load up the CPU. Then monitor CPU clock frequency and see if the same frequency persists for 20 minutes or if it drops. And if that frequency even is full turbo.

rootbeer23 wrote:
hotfog wrote:well I have a question.

I have often heard that FA doesn't take advantage of multicore cpu's.


that is true.
the long answer is more complicated: all physics simulation is single threaded.
some graphics output or sound control is executed in a variety of threads, but those usually
account for 10% extra load. if there are too many computations in the sim thread, the game
will slow down. since all things that are not simulation can run on all processors concurrently,
1 core will be slightly slower than 2+ cores.
short answer is: FA can take advantage of 1.1 cores.


Question :
What is required to make FA run well and fast?
Is it CPU only, does graphics card have a role? Does Ram have a role? Perhaps background programs have an impact?


any low range current GPU should be able to run FA on full detail on a big screen.
probably a CPU integrated graphics core will be able to do so too. so the graphics card
is generally not a problem.


CPU integrated graphics will leech of system RAM and decrease available RAM and bandwidth and will also heat the CPU directly.

rootbeer23 wrote:
if you dont have enough RAM for the whole FA process to fit in, you are in a world of pain.
but who doesnt have 2GB ram?


Just after the release of Windows Vista a whole lot computers were sold with 512 MB and 1 GB due to high RAM prices. Needless to say the user experience of Vista on 512 MB was one of the reasons for it being hated.

rootbeer23 wrote:
Is there a risk that showing CPU model will create unfair kicks? Oh yes for sure. The reason? Ignorance.

So, put up some info and educate us please.

Then, even people with poor requirements to run FA (who are currently unaware of this fact) will know. So many times I have been in game that slows down due to bad cpus and I always get targeted as my game ping is usually between 220 and 360 - which doesnt cause lag but because it is higher than the people playing from europe, I am assumed to be the cause of the game lag. People are unaware of the issue of cpu lag. This is just an example of the upset that ignorance can cause.

So, let's inform them. Perhaps we could get speed2 to do a speed school of FAF video on cpu lag to educate.

Just saying.


network latency is of course not relevant in this case.

hardware specs are important for performance though. so a kick would only be unfair
if it was not justified. i mean thats almost a tautology, no?


The story about him being targeted first is a good one, but players from S-A don't seem to be a protected "race" like the aussies around here, so it seems nobody cares. :x :roll:

Statistics: Posted by E8400-CV — 10 Apr 2015, 11:41


]]>
2015-04-10T10:15:43+02:00 2015-04-10T10:15:43+02:00 /viewtopic.php?t=9751&p=98135#p98135 <![CDATA[Re: Show CPU model in lobby]]>
hotfog wrote:
well I have a question.

I have often heard that FA doesn't take advantage of multicore cpu's.


that is true.
the long answer is more complicated: all physics simulation is single threaded.
some graphics output or sound control is executed in a variety of threads, but those usually
account for 10% extra load. if there are too many computations in the sim thread, the game
will slow down. since all things that are not simulation can run on all processors concurrently,
1 core will be slightly slower than 2+ cores.
short answer is: FA can take advantage of 1.1 cores.


Question :
What is required to make FA run well and fast?
Is it CPU only, does graphics card have a role? Does Ram have a role? Perhaps background programs have an impact?


any low range current GPU should be able to run FA on full detail on a big screen.
probably a CPU integrated graphics core will be able to do so too. so the graphics card
is generally not a problem.

if you dont have enough RAM for the whole FA process to fit in, you are in a world of pain.
but who doesnt have 2GB ram?
other applications may or may not be a problem. in general, 2 apps can make full use of
a multicore system. but if you do raytracing on 4 cores and FA on the other 4 cores, i still
would think that is probably some kind of trouble. so, you have to try it out generally.


iow, INFORM the FAF community as there is no clear description anywhere that i can find of an ideal system to run FA.


intel dual core with at least 3ghz.
amd with 3500mhz is ok too.
no problems for 10x10km maps or less, generally acceptable for larger maps.
more than 4GB ram dont do any more good, bigger GPU doesnt help either.
the best computer to run FA is the one with the best single thread performance (you can google for that).


Is there a risk that showing CPU model will create unfair kicks? Oh yes for sure. The reason? Ignorance.

So, put up some info and educate us please.

Then, even people with poor requirements to run FA (who are currently unaware of this fact) will know. So many times I have been in game that slows down due to bad cpus and I always get targeted as my game ping is usually between 220 and 360 - which doesnt cause lag but because it is higher than the people playing from europe, I am assumed to be the cause of the game lag. People are unaware of the issue of cpu lag. This is just an example of the upset that ignorance can cause.

So, let's inform them. Perhaps we could get speed2 to do a speed school of FAF video on cpu lag to educate.

Just saying.


network latency is of course not relevant in this case.

hardware specs are important for performance though. so a kick would only be unfair
if it was not justified. i mean thats almost a tautology, no?

Statistics: Posted by rootbeer23 — 10 Apr 2015, 10:15


]]>
2015-04-10T09:34:34+02:00 2015-04-10T09:34:34+02:00 /viewtopic.php?t=9751&p=98133#p98133 <![CDATA[Re: Show CPU model in lobby]]>
I have often heard that FA doesn't take advantage of multicore cpu's.

Question :
What is required to make FA run well and fast?
Is it CPU only, does graphics card have a role? Does Ram have a role? Perhaps background programs have an impact?

iow, INFORM the FAF community as there is no clear description anywhere that i can find of an ideal system to run FA.

Is there a risk that showing CPU model will create unfair kicks? Oh yes for sure. The reason? Ignorance.

So, put up some info and educate us please.

Then, even people with poor requirements to run FA (who are currently unaware of this fact) will know. So many times I have been in game that slows down due to bad cpus and I always get targeted as my game ping is usually between 220 and 360 - which doesnt cause lag but because it is higher than the people playing from europe, I am assumed to be the cause of the game lag. People are unaware of the issue of cpu lag. This is just an example of the upset that ignorance can cause.

So, let's inform them. Perhaps we could get speed2 to do a speed school of FAF video on cpu lag to educate.

Just saying.

Statistics: Posted by HotFog — 10 Apr 2015, 09:34


]]>
2015-04-10T08:51:08+02:00 2015-04-10T08:51:08+02:00 /viewtopic.php?t=9751&p=98131#p98131 <![CDATA[Re: Show CPU model in lobby]]>
E8400-CV wrote:
No, I would enable the option to kick if it slows down bad and welcome you in.

Haswell CPU's run a bit cooler than previous generations. If I run my i7-2860QM full-bore, it will run a lot hotter.


So what is this function for? You only need the one from your other topic.

Statistics: Posted by Aurion — 10 Apr 2015, 08:51


]]>
2015-04-10T04:49:09+02:00 2015-04-10T04:49:09+02:00 /viewtopic.php?t=9751&p=98123#p98123 <![CDATA[Re: Show CPU model in lobby]]>
E8400-CV wrote:
Yes, but do you see "free" as in system RAM minus what task manager reports as used? Or what resource management shows as "free"?


There's both "free" and "available" (free+filesystem cache - when getting meminfo in python) so in this case "available" would be the thing to use I guess. Make sure that there's enough (500 mb?) of memory available when in the lobby.

The current setons replay i'm running has allocated 700 mb @ 20 min gameplay

Statistics: Posted by Crotalus — 10 Apr 2015, 04:49


]]>
2015-04-10T02:02:39+02:00 2015-04-10T02:02:39+02:00 /viewtopic.php?t=9751&p=98118#p98118 <![CDATA[Re: Show CPU model in lobby]]>
E8400-CV wrote:
Aurion wrote:
E8400-CV wrote:...
Decent laptops for FA gaming are unicorns.


I must be playing with a unicorn.
(the unicorn being an ASUS ROG G75VW for over two years ago (i7 -3610QM @2.30Ghz - 8GB RAM)
Not to say it never lags, as I don't often end up in over 60 minutes seton games, but so far, so good.

Statistics: Posted by Zoram — 10 Apr 2015, 02:02


]]>
2015-04-10T01:25:02+02:00 2015-04-10T01:25:02+02:00 /viewtopic.php?t=9751&p=98114#p98114 <![CDATA[Re: Show CPU model in lobby]]> Statistics: Posted by IceDreamer — 10 Apr 2015, 01:25


]]>
2015-04-10T01:22:12+02:00 2015-04-10T01:22:12+02:00 /viewtopic.php?t=9751&p=98113#p98113 <![CDATA[Re: Show CPU model in lobby]]>
Crotalus wrote:
Something that also would be useful is to show current free memory in the system so you know if someone is going to swap memory during the game. FA.exe usually grows to ~1 GB during a long game so if people are low on memory it will cause stutter.

Was about to patch it in but there's no good way feeding data from FAF-client -> game after it's launched so you can't show memory status during a game, only in the lobby.


Yes, but do you see "free" as in system RAM minus what task manager reports as used? Or what resource management shows as "free"?

You're right about the RAM thing. I recently found out opening a JPEG from my new camera eats up 350 MB of RAM :shock:

IceDreamer wrote:
E8400-CV wrote:Enough of the laptop nonsense slowing down the game to -4 or noobs lying about "the cpu score is bugged for me". I've found 99% (rounded down!) to be pure bullsh*t.

When hovering over the cpu score, let it show the CPU model. Just like it's listed in Windows System properties. For example;

Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz 3.30 GHz



That way it's easier to pick out the laggers. Like the people with hot-head laptop cpu's that will overheat and slow down.


E8400-CV wrote:
No, I want to play without slowdowns. Having a better idea of what someone's system is, will make that easier to achieve.


OK so I join your game, and you see this: "Intel(R) Core(TM) i7-4720HQ CPU @ 2.60GHz".

Do I get to play? If you're going with a CPU models ideology, you must know quite a bit about them, so you know HQ means it's a mobile chip, and I'm therefore on a laptop. 2.6GHz is also fairly slow compared to most desktop CPUs. So what's your verdict?

And yes, this post is a trap, but please give your honest answer: Would you kick that CPU?


No, I would enable the option to kick if it slows down bad and welcome you in.

Haswell CPU's run a bit cooler than previous generations. If I run my i7-2860QM full-bore, it will run a lot hotter.

Statistics: Posted by E8400-CV — 10 Apr 2015, 01:22


]]>
2015-04-10T01:01:10+02:00 2015-04-10T01:01:10+02:00 /viewtopic.php?t=9751&p=98107#p98107 <![CDATA[Re: Show CPU model in lobby]]>
E8400-CV wrote:
Enough of the laptop nonsense slowing down the game to -4 or noobs lying about "the cpu score is bugged for me". I've found 99% (rounded down!) to be pure bullsh*t.

When hovering over the cpu score, let it show the CPU model. Just like it's listed in Windows System properties. For example;

Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz 3.30 GHz



That way it's easier to pick out the laggers. Like the people with hot-head laptop cpu's that will overheat and slow down.


E8400-CV wrote:
No, I want to play without slowdowns. Having a better idea of what someone's system is, will make that easier to achieve.


OK so I join your game, and you see this: "Intel(R) Core(TM) i7-4720HQ CPU @ 2.60GHz".

Do I get to play? If you're going with a CPU models ideology, you must know quite a bit about them, so you know HQ means it's a mobile chip, and I'm therefore on a laptop. 2.6GHz is also fairly slow compared to most desktop CPUs. So what's your verdict?

And yes, this post is a trap, but please give your honest answer: Would you kick that CPU?

Statistics: Posted by IceDreamer — 10 Apr 2015, 01:01


]]>
2015-04-10T00:55:18+02:00 2015-04-10T00:55:18+02:00 /viewtopic.php?t=9751&p=98106#p98106 <![CDATA[Re: Show CPU model in lobby]]>
Was about to patch it in but there's no good way feeding data from FAF-client -> game after it's launched so you can't show memory status during a game, only in the lobby.

Statistics: Posted by Crotalus — 10 Apr 2015, 00:55


]]>
2015-04-10T00:26:59+02:00 2015-04-10T00:26:59+02:00 /viewtopic.php?t=9751&p=98101#p98101 <![CDATA[Re: Show CPU model in lobby]]>
Aurion wrote:
E8400-CV wrote:...

No, I want to play without slowdowns. Having a better idea of what someone's system is, will make that easier to achieve.


And you will also run in false positives, meaning you will limit your options and you will be an asshole (but according to your posts on this forum you don't really care about that). It's simply not fair for the people that have a decent laptop with a decent thermal design that can play FA just fine (compared to the average or above average desktop CPUs).


Not at all. That's why this is to be combined with the suggestion from the other topic; Slow simspeed kick

Decent laptops for FA gaming are unicorns. Basing argumentation against showing CPU model on their existence is rather funny. Show me one laptop that keeps the same Turbo clock as during the benchmark after 60 minutes of FAF (so also GPU load). Even my Thinkpad W520 with it's 170W powerbrick will go down 100-200 W.

Ceneraii wrote:
Well this is setons rage I guess, laptops don't usually do well there ;D


You guessed wrong. I play setons a lot less because it's rare to get 8 players with decent CPU. It was a Rohan, -4 at 35 minutes, sandwiched between short breaks of no lag.

Statistics: Posted by E8400-CV — 10 Apr 2015, 00:26


]]>
2015-04-10T00:26:25+02:00 2015-04-10T00:26:25+02:00 /viewtopic.php?t=9751&p=98100#p98100 <![CDATA[Re: Show CPU model in lobby]]>
However, in the meantime, I strongly suggest you simply kick the ppl with slow cpu scores to be sure.

Just throw out those babies with that bathwater.

:)

Statistics: Posted by HotFog — 10 Apr 2015, 00:26


]]>
2015-04-09T14:50:07+02:00 2015-04-09T14:50:07+02:00 /viewtopic.php?t=9751&p=98058#p98058 <![CDATA[Re: Show CPU model in lobby]]> Statistics: Posted by Ceneraii — 09 Apr 2015, 14:50


]]>