https://bugs.freedesktop.org/show_bug.cgi?id=100761
--- Comment #6 from Paul Menzel ---
(In reply to Christian König from comment #5)
> What I've forgot: One thing that could possible help would be an HDP flush
> directly before we wait for the fence.
>
> Just
https://bugs.freedesktop.org/show_bug.cgi?id=100761
--- Comment #5 from Christian König ---
What I've forgot: One thing that could possible help would be an HDP flush
directly before we wait for the fence.
Just in case the value written for the fence is stuck in the HDP
https://bugs.freedesktop.org/show_bug.cgi?id=100761
--- Comment #4 from Christian König ---
(In reply to Paul Menzel from comment #3)
> Interesting. What part is responsible for this? AtomBIOS?
No, that is completely unrelated to AtomBIOS.
The PLL just needs awfully
https://bugs.freedesktop.org/show_bug.cgi?id=100761
--- Comment #3 from Paul Menzel ---
(In reply to Christian König from comment #2)
> (In reply to Paul Menzel from comment #0)
> > Looking at the generated graph 540 ms is spent in `radeon_ib_ring_tests`,
> >
https://bugs.freedesktop.org/show_bug.cgi?id=100761
Christian König changed:
What|Removed |Added
Resolution|--- |NOTABUG
https://bugs.freedesktop.org/show_bug.cgi?id=100761
Paul Menzel changed:
What|Removed |Added
CC|
https://bugs.freedesktop.org/show_bug.cgi?id=100761
Bug ID: 100761
Summary: `radeon_ib_ring_tests` takes 0.5 s on resume
Product: DRI
Version: DRI git
Hardware: Other
OS: All
Status: NEW
Severity: normal