AMD 14.7 RC3 Driver Performance Loss vs 14.4

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0
Topic 197681

Has anyone experience performance losses running the latest 14.7 RC3 AMD drivers vs the 14.4 drivers? I have seen an increase in task completion time while running the Perseus Arm Survey of about 300 seconds (5700 vs 6000) also an approximate doubling of reported CPU time per task and a drop of reported GPU utilization (5%) while running 2 task concurrently on two R9 280Xs. This is in a Windows 8.1 OS. If anyone has any suggestion for the best drivers for this particular work unit and OS I would appreciate any input, thanks.

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0

AMD 14.7 RC3 Driver Performance Loss vs 14.4

Also CPU usage has jumped up to from approximately 30% to 80% with just running 2 tasks per GPU with the change of driver from 14.4 to 14.7 RC3.

mikey
mikey
Joined: 22 Jan 05
Posts: 5,317
Credit: 526,073,478
RAC: 43,589

RE: Also CPU usage has

Quote:
Also CPU usage has jumped up to from approximately 30% to 80% with just running 2 tasks per GPU with the change of driver from 14.4 to 14.7 RC3.

If you don't game it looks it's time to revert back to the older ones.

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0

Yes I rolled the driver back

Yes I rolled the driver back to 14.4, I didn't test 14.7rc3 with other GPU tasks but for the Pereus search it isn't ideal. Disappointing as there were improvements to some game performance, mantle seemed smoother in bf4. Maybe the next driver will be better for both functions.

mikey
mikey
Joined: 22 Jan 05
Posts: 5,317
Credit: 526,073,478
RAC: 43,589

RE: Yes I rolled the driver

Quote:
Yes I rolled the driver back to 14.4, I didn't test 14.7rc3 with other GPU tasks but for the Pereus search it isn't ideal. Disappointing as there were improvements to some game performance, mantle seemed smoother in bf4. Maybe the next driver will be better for both functions.

I am still running 14.1 on my AMD cards, I don't game at all so crunching support is all I care about.

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0

Mikey have you found the 14.1

Mikey have you found the 14.1 drivers(http://www.guru3d.com/files-details/amd-catalyst-14-1-beta-1-6-(13-350-1005-january-31).html) to offer better performance/shorter task completion times over the 14.4?

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0

I just realized your citing

I just realized your citing the driver not the catalyst version disregard my last post, we are using the same driver.

Bill592
Bill592
Joined: 25 Feb 05
Posts: 786
Credit: 70,825,065
RAC: 0

RE: I just realized your

Quote:
I just realized your citing the driver not the catalyst version disregard my last post, we are using the same driver.

Howdy,
The catalyst version is the driver version. They are equivalent.
I am still playing BF-2 so, I don't need too new of a driver for that )))

Bill

mikey
mikey
Joined: 22 Jan 05
Posts: 5,317
Credit: 526,073,478
RAC: 43,589

RE: Mikey have you found

Quote:
Mikey have you found the 14.1 drivers(http://www.guru3d.com/files-details/amd-catalyst-14-1-beta-1-6-(13-350-1005-january-31).html) to offer better performance/shorter task completion times over the 14.4?

I have never tried the 14.4 drivers so can't say, I can just say that I am happy with the 14.1 drivers, they work just fine.

Titan
Titan
Joined: 29 Aug 13
Posts: 19
Credit: 25,868,802
RAC: 0

Hey Bill592!, in my previous

Hey Bill592!, in my previous post I was referring the packaged video driver inside of the catalyst bundle for example 14.7 RC3 contains several products, if you will, for example the video driver is labeled as 14.200.1004.0000 vs the 14.4 WHQL catalyst package which has the 14.100.0.0000 video driver. Why they used a different naming convention? I have no idea, however my original post was commenting on the poor performance of the 14.7 RC3 catalyst bundle as the open cl version build seems to have changed vs the 14.4 catalyst bundle, as the performance I have observed was worse while performing the Perseus Arm Survey Tasks.

I am by no means an expert on which driver Mikey is using in his ATI/AMD hosts, as I just looked up his hosts and they display the same driver version that displays with my host, so I assumed we were both using the same driver version that is bundled with 14.4. It is possible that E@H reports invalid information concerning driver version in which case my assumptions concerning driver versions would be false.

Catalyst: 14.7 RC3
DriverVer=08/11/2014, 14.200.1004.0000
CCC: 2014.0811.2303.39561
D3D: 9.14.10.01044
OGL: 6.14.10.12967
OCL: 10.0.1526.3
Mantle: 9.1.10.0020

Catalyst: 14.4 WHQL
DriverVer=04/17/2014, 14.100.0.0000
CCC: 2014.0417.2226.38446
D3D: 9.14.10.01029
OGL: 6.14.10.12847
OCL: 10.0.1445.5
Mantle: 9.1.10.0013

Bill592
Bill592
Joined: 25 Feb 05
Posts: 786
Credit: 70,825,065
RAC: 0

RE: Hey Bill592!, It is

Quote:

Hey Bill592!, It is possible that E@H reports invalid information concerning driver version

Good Mornin Titan

The 'computer' info on Einstein doesn't seem to know Shiʿite about what driver we are using (at least for AMD)
It reads out the same thing 1.4.1848 for all AMD Gpu's (

Take Care,
Bill

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.