Rage only using 35% of my Q6600 = wtf?

Post » Tue May 15, 2012 9:10 pm

Title said it all really, rage is only using 35% of my Q6600 @ 3600mhz This is with/without any of the various tweeks to manage multi core performance, plus the fact that rage only uses around 360MB of my 4GB of memory. Tried running the game is safemode too, with the same result. I know this game was coded for running on consoles but this is crazy as I'm certain that its not down to any tweeks left behind as I have removed them all in the process of testing and reinstalled the game.

Even with a very generous 10% for OS overheads the game isnt even taxing my cpu AT ALL - I would love to find out how much it is utilizing my 5830 gpu. Talking about my gpu the only driver version I have found that works is 6.14.10.10834 every other driver I have tried causes the game to run @ 0.5 fps and/or causes the blue screen issue.

In the end I can get the game to run but it feels very laggy due to the fact it cannot use my cpu in the correct manner and the drivers have an even bigger problem with the 5830 cards.

Sorry for the doom and gloom, but please fix this pile of poo!
User avatar
Ashley Campos
 
Posts: 3415
Joined: Fri Sep 22, 2006 9:03 pm

Post » Tue May 15, 2012 11:51 pm

The game may not actually need to use all of your CPU. One possible reason for that is that your GPU may be your bottleneck; the game will only run as fast as the fastest component in your system, and if that fastest component is slower than your CPU, then it's perfectly natural and normal to see reduced CPU usage.

You can check GPU usage with a tool like http://www.geeks3d.com/20100528/gpu-tool-gpu-shark-0-2-0-available/.

Even that may not show 100% GPU usage, as there may very well be other bottlenecks in your system. Disk I/O is one that springs to mind and that would be relevant for Rage. Memory I/O is another.
User avatar
Shirley BEltran
 
Posts: 3450
Joined: Wed Jul 26, 2006 4:14 pm

Post » Wed May 16, 2012 3:30 am

You can test cpu/gpu bottlenecks quite easily with no extra software, already done that. My rig is running on a 128GB M4 ssd and OCed Pc2-9200 so it won't not be an IO / bandwith problem. I would not have posted if it was something to do with my rig, I've seen consistantly better results with less powerful hardware. I'm not saying my rig is the monster that it once was but it punches far above its weight even today and it should not have the problems it does with Rage - with all of its 'fancy' coding and tricks.

I even went as far as installing a fresh OS with only the drivers and rage installed and I still got the same performance. The problem is the same as everyone elses - crappy coding and crappy threading which is made worse because of the special/broken relationship the amd drivers have with the 5830
User avatar
Big mike
 
Posts: 3423
Joined: Fri Sep 21, 2007 6:38 pm

Post » Tue May 15, 2012 5:54 pm

Still other possibilities. Dynamic texture updating, for example, is one thing I've benchmarked and found to run quite poorly on AMD hardware. Batch updates of a 256x256 32-bit texture with an optimal pixel format give a result of 3ms on NVIDIA, 12ms on Intel and 19ms on AMD. Since Rage depends quite heavily on dynamic texture updates, there's one other bottleneck for you right away.
User avatar
Chloe :)
 
Posts: 3386
Joined: Tue Jun 13, 2006 10:00 am

Post » Wed May 16, 2012 3:13 am

I appreciate your comments but its got nothing to do with DT I've done many tests myself and the numbers/scaling just dont add up regarding OC/non OC and resoloution/performance rates. There is something fishy going on with the 5830/drivers - I dont know if thats the whole problem or its just compounding all the other crap the engine is doing.

The 5830 is a different kettle of fish in regards to this game.
User avatar
Daniel Brown
 
Posts: 3463
Joined: Fri May 04, 2007 11:21 am

Post » Wed May 16, 2012 3:54 am

If you'd care to try an experiment...

The link below is to a test program that just benchmarks dynamic texture updates. It uses Direct3D so OpenGL driver quality (...or lack thereof...) is ruled out as a factor. I'd be interested in seeing what kind of performance you (and anyone else who is interested) gets with this (I'm at ~2200 FPS).

http://www.sendspace.com/file/au4jlz

I've included the full source code so you can compile yourself and ensure that it comes with no nasties.
User avatar
Cagla Cali
 
Posts: 3431
Joined: Tue Apr 10, 2007 8:36 am

Post » Wed May 16, 2012 12:46 am

I know your only trying to help, but its got nothing to do with DT. ATI maybe slower yes - but its got squat to do with my specific problem...

I just installed (yet another) driver to see the impact on the game - yet again I have some really funky broken/blue textures though this time I was running a log of my resources and found that cpu usage jumped up to a rock steady 50% which is more in line with a (semi?) threaded game @ low resoloutions. And system memory usage jumped to almost 800MB of data.

These figures are what I would expect from a game such as rage (still a tad on the low side considering the hardware it could utilise) and if the 0.5 fps/texture bug was fixed I'm sure that the performance would scale quite nicely.

I think it boils down to a specific problem in the driver (the only driver?) that will work on a 5830, and the 5830 itself because of the 0.5fps bug which I gather only effects this particular card.
User avatar
Franko AlVarado
 
Posts: 3473
Joined: Sun Nov 18, 2007 7:49 pm

Post » Tue May 15, 2012 6:28 pm

Double post
User avatar
Rob Smith
 
Posts: 3424
Joined: Wed Oct 03, 2007 5:30 pm

Post » Wed May 16, 2012 1:39 am

The game may not actually need to use all of your CPU. One possible reason for that is that your GPU may be your bottleneck; the game will only run as fast as the fastest component in your system, and if that fastest component is slower than your CPU, then it's perfectly natural and normal to see reduced CPU usage.

You can check GPU usage with a tool like http://www.geeks3d.com/20100528/gpu-tool-gpu-shark-0-2-0-available/.

Even that may not show 100% GPU usage, as there may very well be other bottlenecks in your system. Disk I/O is one that springs to mind and that would be relevant for Rage. Memory I/O is another.

WTF?! IF his CPU is at 35%, how could that possibly be a bottle neck? Jesus Mary and Joseph, the hammer legion is strong with this one. Your comment is just full of idiot.
User avatar
Andrea P
 
Posts: 3400
Joined: Mon Feb 12, 2007 7:45 am

Post » Tue May 15, 2012 3:15 pm

WTF?! IF his CPU is at 35%, how could that possibly be a bottle neck? Jesus Mary and Joseph, the hammer legion is strong with this one. Your comment is just full of idiot.

The haters fail basic reading comprehension again (in other news: water is still wet).

GPU is bottleneck, not CPU.

That's "GPU" beginning with 'G', the 7th letter of the alphabet, upper-case 'g', the one that comes before 'H' and after 'F'; I'm sure you must be familiar with it. You did use it in your own post, after all.

You do actually know what a GPU (beginning with 'g') is, don't you?
User avatar
I love YOu
 
Posts: 3505
Joined: Wed Aug 09, 2006 12:05 pm

Post » Wed May 16, 2012 5:30 am

Q6600 here too,

285 GTX SLI & Q6600 @ 3.2GHz - Huge fps drop , impossible to stabilize fps with more than 4xAA !

Single 285 GTX & Q6600 @ 2.4GHz - Damn stable 60fps with 16xAA ! :bowdown:
User avatar
Hayley Bristow
 
Posts: 3467
Joined: Tue Oct 31, 2006 12:24 am

Post » Wed May 16, 2012 2:56 am

ATI 5830 here too,

less than 1 FPS, unplayable :(
User avatar
X(S.a.R.a.H)X
 
Posts: 3413
Joined: Tue Feb 20, 2007 2:38 pm

Post » Wed May 16, 2012 7:30 am

If you'd care to try an experiment...

The link below is to a test program that just benchmarks dynamic texture updates. It uses Direct3D so OpenGL driver quality (...or lack thereof...) is ruled out as a factor. I'd be interested in seeing what kind of performance you (and anyone else who is interested) gets with this (I'm at ~2200 FPS).

http://www.sendspace.com/file/au4jlz

I've included the full source code so you can compile yourself and ensure that it comes with no nasties.

What the hell is that? It was absolutely full of garbage. A toolbar, weatherbug, bflix, etc. There were like 10 things I had to decline in the installation and I'm not even sure it installed the tool.
User avatar
Sara Lee
 
Posts: 3448
Joined: Mon Sep 25, 2006 1:40 pm

Post » Tue May 15, 2012 5:49 pm

What the hell is that? It was absolutely full of garbage. A toolbar, weatherbug, bflix, etc. There were like 10 things I had to decline in the installation and I'm not even sure it installed the tool.

Wow that is an outright lie. The program doesn't even need installing.

I ran it on my i7-950 and GTX460 1GB and get approximately 3700 FPS.
User avatar
Alyna
 
Posts: 3412
Joined: Wed Aug 30, 2006 4:54 am

Post » Tue May 15, 2012 11:59 pm

Wow that is an outright lie. The program doesn't even need installing.

I ran it on my i7-950 and GTX460 1GB and get approximately 3700 FPS.

It wasn't a lie. It was a mistake. The download link was overshadowed by an ad download so I couldn't tell the difference. I hate advertising.

Approximately 2400 for me.
User avatar
Scott Clemmons
 
Posts: 3333
Joined: Sun Sep 16, 2007 5:35 pm

Post » Tue May 15, 2012 5:05 pm

It wasn't a lie. It was a mistake. The download link was overshadowed by an ad download so I couldn't tell the difference. I hate advertising.

Ah okay, apologies then. Glad you sorted it. I use Firefox with adblock so didn't see any of those.
User avatar
Stephanie Nieves
 
Posts: 3407
Joined: Mon Apr 02, 2007 10:52 pm


Return to Othor Games