Th4d

I agree, someone definitely should, they are in dire need of someone with that skill.

Check out this old article on multithreading on how Valve used different ways of doing multithreading and claimed near linear scalabe results (that we can find in others engines/games today that scale very well with more cores): http://techreport.com/review/11237/valve-source-engine-goes-multi-core About watchdogs, i get 50-60fps with everything on high using mhbo and temporal smaa on an amd x8 and gtx760 2gb 256ssd and fullhd. (my guess is that if i had a 3gb video card i could push it to ultra without adding stutter) Also, today both opengl and mantle pretty much eliminated driver overhead if implemented well enough, dx11 has already made some improvements that can wield up to 10% performance gains when that is the bottleneck.

Could you read back to me the second word on the second vertical bar on this recent roadmap and tell me why it´s there or needed at all? To any sane person this entire thread is riddled with evidence, but this official one should end this discussion about optimizations being needed.

Fanboys can´t help themselves, they ignore everything shoved in their faces and dismiss it as a minority even when people on this forum make a poll and turns out 50% of around 1800 are displeased with the multiplayer performance. That´s the official stance, this developer cannot handle the issue, the only thing left to do is make empty promises that are around for years and belittle anyone who complains. The diagnosis for the mental disorder on such fanboys is cognitive dissonance, because i truly think that he does believe in what he says. http://en.wikipedia.org/wiki/Cognitive_dissonance

http://www.howtogeek.com/126430/htg-explains-what-is-the-windows-page-file-and-should-you-disable-it/ http://www.howtogeek.com/95915/heres-why-disabling-the-windows-pagefile-is-pointless/

I tried starting a couple of single player campaign missions (signal lost and common enemy) at certain spots comparing my fps between freds one and the -system which i simply got used to using for a long time, since i´ve read about windows 7 one being "state of the art" by that guy i posted earlier, and also creating a server with a small tdm one with a friend, making sure videosettings werent too high to make sure the video wouldn´t be the bottleneck. I know it´s somewhat vague and that´s why i asked for the replicatable reliable method he used to claim those gains, but upon going through that entire topic i saw some guy saying he had 9% fps gains and now i´m pretty sure that´s where fred got his number.

I wasn´t being sarcastic or ironic, i truly questioned the fps gains you claimed because i haven´t seen them, to me isn´t 6 or 9%, its zero. Being vague or outright ignoring when people are straightforward and ask both if there is detailed replicatable evidence and as to how one could achieve that claimed clientsided fps gain, barely answering while being sarcastic is pretty much what i would expect from someone that can´t back up those claims with unbiased information, you know, proof that it isn´t bullshit. You pointed me to your topic but seems like everyone there tests in a completely different manner from each other, without any sort of standands. Not to mention they don´t even take into account graphic settings, viewrange, other applications in the background such as antivirus. A lot of things could affect performance, especially if one is using a hard disk drive. So no, no test mentioned there comes close to being done in a scientific manner. My bet is that you throw around this 9% fps gain because someone on that topic said so without providing any evidence for it, which to me is unbelievable. Maybe i should simply add -maxmem 16000 or other placebos to pile up with it so i can feel the game being "smoother" ? /s Somehow this discussion reminds me of this:

Yeah that´s pretty vague and it´s hard to compare 2 unknown variables and it´s impact. But if you guys had tangible gains that we can expect to see in a future update, that´s great! I agree, i think it would be great to see those, making statements like 9% clientsided gains under certain unknown circunstances is pretty easy, but i would love to see them in some replicatable tests. No doubt it would increase our praise for the effort. Yeah i´ve been through that thread, lots of unreliable tests from users without specific parameters to be replicable, results could even be placebo with statements like "performance seems improved, client felt smooth" or "we tested it with single on server loaded = nothing in fps changed. Then a mate with low hardware tested clientside = nothing in fps changed." And tested it for myself with no real gains (i have 16gb), that´s why i expected a more reliable and standard test shown with real world results, since i consider that my perspective alone isn´t enough, after all it´s hard to account for all the variables. 9% fps gain is a big statement, so i would love to see those tests showing that, done in a scientific approach to back it up. I´m not saying that there is no improvement in memory allocation, but i do question if this memory improvement has any influence in fps gains, and compared to what. If you tell me how to test it to achieve those 9%, i will gladly do so. If you want me to continue this inquiry in the other thread, i shall.

6-9% more fps clientside? that´s amazing!, But now, compared to what exactly, the current default intel one, or windows 8 one? and using what tests? are they reliable and by that i mean do they vary between them? Can i see those tests? Because i would love to see 6-9% improvements on a controlled test. Because you know, without well made tests confirmed by third parties one could claim anything right? Meanwhile, thanks for your hard work.

You claim 25% acceleration of all main memory accesses, and claim that this is a big bottleneck for the cpu in ArmA. I´m sorry but unless you get 25% fps increase, one of the two claims there are false. What´s the % gained in fps increase by using your memory allocator, are there any serious, thorough and replicatable benchmarks on it? Compared to what´s used as default in ArmA and system with the windows 7 and 8 ones. If you are able to show me a 25% increase, in read and writes but there are no fps gains to be seen close to that, then there is no cpu memory bottleneck in ArmA, to me that´s a pretty straight forward logic. Don´t get me wrong, i think your exercise to try to improve things is great, but does it wield any real world gains? If so, what are they?

I highly doubt this memory allocator is better than the Windows 8 one. Haven´t seen benchmarks proving otherwise. I remember taking a look at those custom allocators and some friends considered nedmalloc to be faster, but on it´s official site it was said that despite being better than the windows xp one, "windows 7 memory allocator was state of the art". Taking into account that windows 8 is considered to be even better, and now theres windows 8.1, i would like to see a benchmark between this custom allocator and windows 8´s, shown as fps gains. Found the one i was talking about: http://www.nedprod.com/programs/portable/nedmalloc/ "If you're running on an older operating system (e.g. Windows XP, Linux 2.4 series, FreeBSD 6 series, Mac OS X 10.4 or earlier) you will probably find it significantly improves your application's performance (Windows 7, Linux 3.x, FreeBSD 8, Mac OS X 10.6 all contain state-of-the-art allocators and no third party allocator is likely to significantly improve on them in real world results)." ArmA needs to go 64bits, and be optimized for it, not just to use more ram. Does anyone with a 32bit only cpu runs ArmA 3?

Have you ever played DCS? compare them to any vehicle in any version of ArmA, now tell me again how ArmA is a simulator. Not even bis devs call ArmA 3 a simulator, they call it a game with some simulation aspects. To me ArmA is a tactical shooter with strategy aspects, and still somewhat arcady. But a Sim? I laugh. Compared to DCS not even Take On can be called a simulator.

"your" This new driver has great gains only when there´s a cpu bottleneck caused by the driver that slows the game down, but ArmA´s bottleneck isn´t driver related, the bottlenecks are inherent to how the game was designed and lie on it´s own code. Blaming gpu drivers is not feasable as an excuse for the games performance anymore, what saddens me is that this somewhat shows that even if they introduced Mantle or optimized for the upcoming DX12, i fear that it would wield no performance gains, just clearly shows that the engine and the design approach used is simply too outdated for current hardware.

Sure, except others fare better dealing with the bottlenecks on the first core and are a lot better with multithreading If they need to be in order to perform well enough, since most games are GPU limited. "arma has a vast blah blah" and cant be compared, but it´s not like you need to know everything that goes on on the map and like there is no better optimizations to be done, have they fixed the issue with broken building being left under the map? guys over at DayZ know these issues are limiting their design, i remember when dayz was going to be 200 players with 1000 permanent spawned zombies, and are trying to do something about it. Shit, ArmA still claims 60VS60 players on the website giving out very weak machines as minimum/recommended requirenments. And i remember valves hybrid way of multithreading that claimed scalable linear gains on the way they were doing it: http://techreport.com/review/11237/valve-source-engine-goes-multi-core , also, today nvidia released a driver that dealt a lot better with render batch overhead, no improvement here, 28 fps yesterday on a campaign single player mission, 28fps half an hour ago and lowering settings makes no difference. Anyway i think i´ve over extended this off topic cpu core discussion, was just trying to correct the false assumption that the guy there had more than 2 working cores usage at any time in arma. We just need good enough performance no matter how they achieve it, bis has big performance improvement plans for DayZ that do seem very interesting, what are the ones for ArmA? Will ArmA benefit directly from them aswell? If so, great then, just hurry up already it´s been a few years overdue.

Still, usage for arma 3 is shown as 33% on your screenshot, how many cores do you have? 6? 2 cores would be 33% right? Windows spreads usage across cores to even out heat, and it shows as being spread to you since it´s too fast, but it´s just using 2 cores, my bet is that if you disable 4 cores, you will get the same performance with 90-100% usage on the 2 cores left. Arma barely uses more than 2 cores imho, i´ve seen someone at bis saying the same, i believe it was their CEO. Btw, i already tried this on an x6 and x8 cpu, theory held up. Dayz´s biggest statements on optimizations shed a light on Arma´s worse offenders, main 2 being: "we are making a bubble around the player so he won´t have to deal with everything else that is going on on the entire map" and "we are going to make the server multithreaded".