Hey Nerds!
I had read a few articles regarding a massive boost in OpenGL performance in new 22h2 windows and AMD Hardware. I did a quick comparison. Just random information that I found interesting. For what its worth.
Hardware: 5950x, 6900xt x1, 6800xt x2
Comparing Windows 11(22000.00) AMD (22.7.1) VS Windows 11 22H2 (22622.450) AMD (22.7.1)
Note: GPU utilization = 0.16
(Before) Windows 11(22000.00) AMD (22.7.1)
Gamma-ray pulsar binary search #1 on GPUs v1.28 () windows_x86_64
N=130
Run time = 830.94
CPU time = 78.82
(After) Windows 11 22H2 (22622.450) AMD (22.7.1)
N=131
Run time = 836.05
CPU time = 20.09
I'm not smart enough to realize if this is relevant or not, but looks to be a massive decrease in CPU time, but no change in run time. Must be another bottleneck at play?
https://www.amd.com/en/support/kb/release-notes/rn-rad-win-22-7-1
Copyright © 2024 Einstein@Home. All rights reserved.
Please note that the GPU
)
Please note that the GPU calculations here use OpenCL and not OpenGL.
Agreed. The Did mention
)
Well Crap.
They did mention Improved performance in OpenCL in AMD drivers, but must not be relevant to my Cards. I just thought the drop In CPU time was significant, but apperently its not. I am still curious why CPU dropped so much while overall time remained static.
The improvement was aimed
)
The improvement was aimed specifically at Folding@home performance from the other comments I’ve seen about the new driver.
I guess it doesn’t translate to the kinds of calculations being performed by the gamma ray application.
folding does molecular modeling, Einstein is more signals analysis with FFTs.
_________________________________________________________________________