Jump to content
Sign in to follow this  
suma

70256-70313 - which gives you best performance?

Which of the following gives you best performance?  

128 members have voted

  1. 1. Which of the following gives you best performance?

    • 1.05 version
      30
    • 70256 with -exThreads=0 option or with no -exThreads option
      8
    • 70256 with -exThreads=1 option, I have DualCore
      4
    • 70256 with -exThreads=3 option, I have DualCore
      30
    • 70256 with -exThreads=1 option, I have QuadCore (or more)
      1
    • 70256 with -exThreads=3 option, I have QuadCore (or more)
      56


Recommended Posts

Been testing this a whole lot more now and the thread param 3 still holds the best gameplay for me. The game is running real smooth.

@Langgis08:

The FPS test i did with those numbers were without or little AI. I was checking the texture popping only by moving around fast in chernogorsk as i had bad popping before and wanted to check with the new hardware and beta.

All setting to HIGH except the grass wich as i wrote was tested between OFF to NORMAL. Resolution 1280x1024 (3D res 100%), Nvidia CPL set to Performance. Defragged HDD's with O&O v.12 Professional (every week) and no useless programs in the background. RivaTuner only used here to gain 100% GPU fan speed as i current case gets very hot, and if i dont run the fans 100% during ARMA2 i see less FPS after a short while.

This is how i run it and according to FRAPS (if it can be trusted) i have 60-120FPS in chernogorsk depending on area. Normal is 60-70FPS without grass. As soon as i turn the grass to NORMAL i run the same place between 20-50-ish FPS depending on area where normal is 35-45FPS roughly, and 20 where lots of grass and some bushes/trees are. And i seriously think thats a too big hit for grass and i would love at least an option in the game to set it to LOW QUAL/HI QUAL or similar. Its just so boring to turn it off. Especiall since i dont like having advantages like seeing lying down where i shouldnt.

Share this post


Link to post
Share on other sites

Hmm - using my new framerate analyser, I conclude that 1.05 actually had the best performance (although 70313 with exThreads=3 comes close). I have a dual-core 6700. Disclaimer- I should really run these tests a few more times to ensure repeatability - these are just quick one-off results...

fps_beta.jpg

Share this post


Link to post
Share on other sites

lol, everyone's an expert, eh MavericK96? Bottom line is it's what works for you.

It's well documented that ArmA II gives better performance running on XP 32bit, I'd try that rather than fanny about with RAM optimisations. 7's jewellery comes at a price soldier...

Protegimus

True. :) And thanks for the reply/info. I wish I had 12 GB of RAM to mess about with. :cool:

Share this post


Link to post
Share on other sites

For me, using sbsmac benchmark tool the beta 70313 is clearly better that 1.05 on all 3 of the test (coastal, inland, city). I don't get much variation between exthreads 0,1 or 3 though.

I am running Win 7, 64 bit. 8 gig ram. ATI 4870 (2gig) (Cat 10.4) q9650 (quad core o/c to 3.3gig)

Share this post


Link to post
Share on other sites

That's interesting gonk - almost the opposite results to me where exThreads=1 is significantly worse than the other options including 1.05. Having said that. you are on a fast quad-core vs my dual-core. Any chance of posting the graphs ?

Share this post


Link to post
Share on other sites

So I used sbsmac's FPS Analysis Tool (nice work on that, by the way), and here's what I got, running the City Run benchmark mission included with the tool.

Please keep the images <100kb.

FPS Analysis: http://i263.photobucket.com/albums/ii140/AFireEternal/ArmA2FPS.jpg

Notes:

Interestingly, exThreads=1 had more "pulling" stutter (where FPS drops low, 5-10 FPS), and exThreads=3 had a lot more "hard" stutter (where FPS drops to zero for a short time). Overall I think the best option for me is still exThreads=0. exThreads=3 does definitely have faster LOD transitions (still delayed somewhat, but better than 1.05) but the hard stutters are pretty annoying. FPS in general looks to be better with exThreads=0 as well.

Edited by Alex72

Share this post


Link to post
Share on other sites
That's interesting gonk - almost the opposite results to me where exThreads=1 is significantly worse than the other options including 1.05. Having said that. you are on a fast quad-core vs my dual-core. Any chance of posting the graphs ?

ok... hope it is not too large... The slowest blue is 1.05... not much different with threads except for around the city..

Please keep the images <100kb.

FPS Analysis: http://forums.ogn.com.au/attachment.php?attachmentid=45881&d=1274080212

Edited by Alex72

Share this post


Link to post
Share on other sites
Does exThreads= command have anything to do with I7 with HT on?

Good question. Has anyone tested this?

Share this post


Link to post
Share on other sites
Good question. Has anyone tested this?

Yes works for me benson with -exThread=0 being the best for me and i have hyperthread enabled too on an I7 860 quad 2.8ghz.

I am getting error message, when saving a multiplayer or Lan Hosted game.

ErrorMessage: Out of memory (requested 0 KB).

footprint 1093958548 KB.

pages 65536 KB.

Doesnt do it all the time, so far only twice in last 10 games i have tried to save, results in the game CTD, without saving.

System Specs:

I7 860 quad core 2.8ghz

4GB Corsair DDR3 Ram

ATI 1GB HD5770

1TB Serial ATA Hard drive

Share this post


Link to post
Share on other sites

Having a 3.2GHZ OC'ed Q6600, i run the latest beta with "-exThreads=3" & "-cpucount=4" and it gave me 2 good things:

1. Much more fluid feeling overall

2. Viewdistance is now successfully raised from 2000m to 3500m without any noticable lack of performance *wow*

However i have not compared real FPS :-)

Share this post


Link to post
Share on other sites

Well i finally did the test with the new beta + -exThreads=3. To me it seems to be the same as beta 63826, but i notice some "hard" slutter. And still, FRAPS recording don't work. :(

Overall same as Beta 63826 tested on mission Village Sweep, i always use that one for test.

Well i hope newer patch will fix FRAPS recording, it would be nice to record again :)

I'll stick with Beta 63826 again.

Cheers

Share this post


Link to post
Share on other sites

Hey mr.g-c I'm messing with my Q6600 yesterday and today. So far its very stable at 3.0GHz so I'll keep it at that for a couple days to see if anything funny happens but look at the diff.:

beta build70313 = 1828 (-exThreads=3)
beta build70313 = 2056 (-exThreads=3)

beta build70313 = 1983 (-exThreads=1)
beta build70313 = 1998 (-exThreads=1)

beta build70313 = 2168 (-exThreads=0)
beta build70313 = 2026 (-exThreads=0)

beta build70313 = 2319 (-exThreads=3) (Overclock 3GHz)
beta build70313 = 2222 (-exThreads=3) (Overclock 3GHz)

I'm gonna try the latest beta now that it's up. Good stuff! ;)

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×