| Summary: | Kaveri no fix perfomance | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Dmitry <terapy-session> |
| Component: | Drivers/Vulkan/radeon | Assignee: | mesa-dev |
| Status: | RESOLVED NOTABUG | QA Contact: | mesa-dev |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | unspecified | ||
| Hardware: | x86-64 (AMD64) | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
|
Description
Dmitry
2018-12-12 01:50:32 UTC
I'm sorry but I don't understand what your report means. What are you talking about? And what fix? Have 7850К earlier, it was stated the wrong number of render backends. https://www.phoronix.com/scan.php?page=news_item&px=AMD-Kaveri-Mesa-18.2-Boost This should have been fixed in the new kernel version and Mesa 18.2, respectively. But in RADV, I'm seeing old performance. This code is shared by RadeonSI and RADV, so I assume that commit should also help RADV. I also think. But I measured the FPS before and after the Mesa and kernel update. Note that the patch is a theoretical 2x benefit in a very specific case. Do you have a GL game where there is a significant difference? Otherwise I'd assume this is just a specific bottleneck that is not hit a lot in practice and hence would not typically result in significant performance changes. GL games are not compared, only Vulkan. Because there is a suspicion that the patch is not touched RADV. Unless you have numbers that for GL it actually improves performance for some implications I'm not going to spend time on this. OK. Then close. Hopefully pay attention to another important bug report. https://bugs.freedesktop.org/show_bug.cgi?id=109015 More precisely this, sorry. https://bugs.freedesktop.org/show_bug.cgi?id=109017 |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.