@ChuckPa I understand that the ratios seem normal, and I do believe that these are normal usage levels when tone-mapping.
However, I’m sure it is not using hardware acceleration:
- For starters, if I disable hardware acceleration, the CPU usage profile is identical to when I have it enabled when tone-mapping.
- Secondly, the performance is equal, and the stream needs to buffer every two seconds. I would expect that if it were using hardware tone-mapping, the performance would result in, at least, a watchable stream.
- Finally, and most importantly, using
intel_gpu_topfromintel-gpu-tools, I can actually confirm that no GPU activity is going on in the alleged hardware transcoded tone-mapping.
Here’s the output for intel_gpu_top when doing the tone-mapping with hardware acceleration disabled on the server, and the same exact output when doing the tone-mapping with hardware acceleration enabled:
intel-gpu-top - 0/ 0 MHz; 100% RC6; ----- (null); 0 irqs/s
IMC reads: ------ (null)/s
IMC writes: ------ (null)/s
ENGINE BUSY MI_SEMA MI_WAIT
Render/3D/0 0.00% | | 0% 0%
Blitter/0 0.00% | | 0% 0%
Video/0 0.00% | | 0% 0%
VideoEnhance/0 0.00% | | 0% 0%
For comparison, here’s the output of the tool when hardware transcoding SDR 4K (HEVC Main 10) to 1080p HD (High):
intel-gpu-top - 678/ 679 MHz; 21% RC6; ----- (null); 554 irqs/s
IMC reads: ------ (null)/s
IMC writes: ------ (null)/s
ENGINE BUSY MI_SEMA MI_WAIT
Render/3D/0 13.10% |█████████▊ | 0% 0%
Blitter/0 0.00% | | 0% 0%
Video/0 73.73% |███████████████████████████████████████████████████████▎ | 0% 0%
VideoEnhance/0 0.00% | | 0% 0%
