my FR sucks!!!

[mikeWe9a]

In most cases disabling ATI processes will only create problems for you. There's no magic fountain of performance found in disabling processes that the manufacturer saw necessary.

Nowadays I don't even try to tweak my gaming computers. If it doesn't run fast enough at default settings, time to upgrade the hardware. If you can't afford to upgrade your hardware every 5-6 years then its better to ditch the PC and move to consoles.

But according to his hardware specs, he doesn't HAVE an ATI video card- he's got a (nVidia) GTX 970.  He may have had an ATI card at one point, but as of now he's got at least one service running for a video card that is not installed in his computer, and may well be conflicting with the card that he does have installed.

Mike

[MrRiplEy[H]]

But according to his hardware specs, he doesn't HAVE an ATI video card- he's got a (nVidia) GTX 970.  He may have had an ATI card at one point, but as of now he's got at least one service running for a video card that is not installed in his computer, and may well be conflicting with the card that he does have installed.

Mike

In that case shutting down processes is not what he should do, he should uninstall both the ATI and NVIDIA drivers via the control panel and run driver cleaner. Then reinstall the NVidia drivers.

Actually now that I looked again it seems he hasn't installed NVidia display drivers at all! No wonder his performance sucks lol.

[MrRiplEy[H]]

The ATI CCC and MOM processes can cost you, up to, about 10% of your video card's potential performance.

Do you have a link to benchmarks proving this? I personally have not detected any benefit whatsoever in killing these processes, which as you know also run on CPU not on the GPU.

[Skuzzy]

Do you have a link to benchmarks proving this? I personally have not detected any benefit whatsoever in killing these processes, which as you know also run on CPU not on the GPU.

ATI and NVidia both alter how they run when they detect any standard benchmark being run.  They have been doing that for many years.

Yes, they run on the CPU and monitor the video card constantly.  They used to chew up about 15% of the performance, but ATI/AMD has improved that.  Now it is down to about 10%.  Depending on what you are doing you may, or may not, notice the performance loss.  Regardless, any program running on the CPU also impedes video card performance when the CPU is needed to be able to send the data to the video card but is too busy running the Windows message loop for all those programs.  Simple logic should be your guide here.

[MrRiplEy[H]]

ATI and NVidia both alter how they run when they detect any standard benchmark being run.  They have been doing that for many years.

Yes, they run on the CPU and monitor the video card constantly.  They used to chew up about 15% of the performance, but ATI/AMD has improved that.  Now it is down to about 10%.  Depending on what you are doing you may, or may not, notice the performance loss.  Regardless, any program running on the CPU also impedes video card performance when the CPU is needed to be able to send the data to the video card but is too busy running the Windows message loop for all those programs.  Simple logic should be your guide here.

Having 10% overhead on simple monitoring defies my logic Skuzzy. Something must be horribly wrong if this is true. It should count maybe 0.10% of resources and even then be impressively heavy. This should be fairly simple to test. If I have a system that gives me 100 fps with CCC etc. enabled and I get 101-109fps after disabling them, it's clear that those processes were hogging the power. I think I'll perform a test using AH and vsync off using my old AMD card actually.

[Skuzzy]

It was not my intent to infer the Windows message loop would cause the loss in performance.  While it eats away at the CPU cycles, the real loss is due to the CCC/MOM executables continuously poking the video card which inhibits data transfers.

The amount of loss will vary, based on what is running and how it is using the video card.  Standard benchmarks are worthless for this.  Anything from 0% to 10% can be expected.

I expect you to not find any loss, whatsoever.  It will not surprise me.  Leave them running, if you want.  It does not hurt to kill them and it might just help if they are not running.  No downsides.

[Bizman]

Having 10% overhead on simple monitoring defies my logic Skuzzy. Something must be horribly wrong if this is true. It should count maybe 0.10% of resources and even then be impressively heavy. This should be fairly simple to test. If I have a system that gives me 100 fps with CCC etc. enabled and I get 101-109fps after disabling them, it's clear that those processes were hogging the power. I think I'll perform a test using AH and vsync off using my old AMD card actually.

That would be interesting. It's been common knowledge for years that the Catalyst Control Center is a resource hog and you're the first one to ever question and test it.

[Skuzzy]

That would be interesting. It's been common knowledge for years that the Catalyst Control Center is a resource hog and you're the first one to ever question and test it.

The problem with testing it, is you have to have a complete understanding of the process used in the test, and it cannot be a standard benchmark, as the programs are aware of those.

[MrRiplEy[H]]

That would be interesting. It's been common knowledge for years that the Catalyst Control Center is a resource hog and you're the first one to ever question and test it.

This is even more interesting. So everyone just believes it's a resource hog and nobody bothered to even test it? LOL! Now I know why people believe in Santa Claus.... ho ho ho! 

[Skuzzy]

This is even more interesting. So everyone just believes it's a resource hog and nobody bothered to even test it? LOL! Now I know why people believe in Santa Claus.... ho ho ho!  

Back in the day, when CCC was released it created a storm of complaints due to the performance killing aspects of it.  AMD/ATI acknowledged the problems and worked to reduce the performance loss.  No one, "just believes" the problem exists.  It has been a well known issue with CCC since it was released.

Go visit the AMD/ATI forums.  Should be a few hundred threads about it, including AMD/ATI acknowledging the issues.

[Bizman]

This is even more interesting. So everyone just believes it's a resource hog and nobody bothered to even test it? LOL! Now I know why people believe in Santa Claus.... ho ho ho! 

What Skuzzy said. It's been a known resource hog since its release. It would be interesting to see how well AMD/ATI has managed to streamline it.

[MrRiplEy[H]]

What Skuzzy said. It's been a known resource hog since its release. It would be interesting to see how well AMD/ATI has managed to streamline it.

Actually I just realized my son has a computer with a 5850 in it, I can simply try to disable the processes and see what happens with AH framerates/smoothness.

[MrRiplEy[H]]

I sure hope Dimebag uninstalled his ATI drivers and installed the Nvidia drivers by the way.

[SilverZ06]

I sure hope Dimebag uninstalled his ATI drivers and installed the Nvidia drivers by the way.

he never had ati drivers.  I have ati, he has nvidia.

[715]

Actually I just realized my son has a computer with a 5850 in it, I can simply try to disable the processes and see what happens with AH framerates/smoothness.

I have an Intel E8400 @ 3GHz with an ATI 7870 at stock rates under Win XP.  All AH graphics settings maxed except Environment slider, water reflections, and building bump mapping off.  Offline test on CraterMA at A1 in F3 external tower view looking E, vsync off.

MOM and CCC running (total of 33 proc): 63.5 fps  (oscillates between 63 and 64)
MOM and CCC killed    (total of 31 proc): 63.5 fps

However, on my system AH is CPU limited not GPU limited.  AH pegs one core at 100% and the other core loafs at about 15% while the GPU utilization is pretty low (can't remember but it's like 25% or less).