|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
70.181.175.121
In Reply to: RE: Are you Dongle Curious ? posted by AbeCollins on November 26, 2016 at 11:50:10
The Cuda offload contribution is significant. Using Process Lasso, I observed that a 24/192 file converted to DSD256 went from around 34% CPU utilization (ave.) to 24% CPU utilization (ave) with Cuda selected.This is the total CPU including the HQPlayer and Roon.
Edits: 11/26/16Follow Ups:
With AO installed, I am now running at CPU of around 12% when converting to DSD256 with Cuda.
Edits: 11/27/16
That's pretty impressive how much AO and GPU/CUDA offload bring the CPU utilization down. Thanks.
From Miska at CA:
"nVidia just launched their newest Pascal architecture that promises huge performance improvement for HQPlayer-type computing tasks. First products rolling out during the summer and more towards Q1'17. The consumer product like is likely to be called GeForce 1xxx (1000-series)."
I haven't kept up on recent GPU technology but I have a couple customers who use the high-end nVidia Tesla GPUs (Pascal architecture) in High Performance Computing applications. They require lots of power (two to three hundred watts ea) along with a good amount of forced air cooling.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: