View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000681 | madVR | bug | public | 2022-02-23 22:15 | 2022-04-10 21:05 |
Reporter | DMU | Assigned To | madshi | ||
Priority | normal | Severity | minor | Reproducibility | always |
Status | assigned | Resolution | open | ||
Summary | 0000681: The render time in test builds is much higher than in the release version in HDR passthrough mode. | ||||
Description | See FinalStep. | ||||
Tags | No tags attached. | ||||
madVR Version | v0.92.17 | ||||
Media Player (with version info) | mpc-hc 1.9.19 | ||||
Splitter (with version info) | LAV 0.76.0.2 | ||||
Decoder (with version info) | LAV 0.76.0.2 | ||||
Decoding | DXVA2 Copyback | ||||
Deinterlacing | none (progressive) | ||||
DXVA2 Scaling Active | no | ||||
Aero / Desktop Composition | Off | ||||
Problem occurs with mode | windowed mode | ||||
GPU Manufacturer | AMD | ||||
GPU Model | APU R5 4650G | ||||
GPU Driver Version | last | ||||
|
|
|
Is this just debug vs. release build differences? Or are the test builds using typical optimizations? Often test builds include debug symbols and lack some optimizations to make debugging issues easier. Though those wouldn't be for performance testing or validation. |
|
That is surprising. Does it still happen if you turn smoothMotion off? I wonder if it's related. I usually test with it off. |
|
Checked. Disabling SM does not solve the issue. |
|
Just to be safe, can you double check if it's still FinalStep which is consuming the extra time? |
|
Rechecked twice. It's definitely FinalStep. But I noticed one oddity - with the ShowRenderSteps option, the time difference is not so significant, but still there. |
|
Ok, thanks. Which dithering method are you using? Dithering is performed as part of FinalStep, so I wonder if that could be a factor here. If you're using error diffusion, can you try using a different method instead, to see if that removes or reduces the higher rendering time? |
|
In all cases I used Ordered Dithering. Disabling dithering does not solve the issue. 2ms vs 8ms in the FinalStep. |
|
Ok. I'll try to reproduce it here. I'm extremely busy atm, though, so it might take a while. |
Date Modified | Username | Field | Change |
---|---|---|---|
2022-02-23 22:15 | DMU | New Issue | |
2022-02-23 22:15 | DMU | File Added: madVR_FinalStep.png | |
2022-04-09 04:31 | EternalStudent07 | Note Added: 0002872 | |
2022-04-09 09:39 | madshi | Note Added: 0002875 | |
2022-04-10 08:20 | DMU | Note Added: 0002878 | |
2022-04-10 08:20 | DMU | File Added: release_test.png | |
2022-04-10 10:05 | madshi | Note Added: 0002879 | |
2022-04-10 16:25 | DMU | Note Added: 0002880 | |
2022-04-10 16:25 | DMU | File Added: release_VS_test_FinalStep.png | |
2022-04-10 17:15 | madshi | Note Added: 0002881 | |
2022-04-10 18:00 | DMU | Note Added: 0002882 | |
2022-04-10 18:00 | DMU | File Added: dith.png | |
2022-04-10 21:05 | madshi | Note Added: 0002883 | |
2022-04-10 21:05 | madshi | Assigned To | => madshi |
2022-04-10 21:05 | madshi | Status | new => assigned |