[Bug 100092] Assertion failed: size_log2 <= ANV_MAX_STATE_SIZE_LOG2 While compiling large compute shader

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Tue Mar 7 03:35:41 UTC 2017


https://bugs.freedesktop.org/show_bug.cgi?id=100092

--- Comment #4 from Josh Gargus <registration-bugzilla at schwa.ca> ---
Thanks Jason!  

In general on NVIDIA, it seems faster to store values in registers than shared
memory, *unless* there is not enough register space and it needs to spill the
data elsewhere (to main GPU memory? to some other cache? I have no idea).

In this case, it seems that making roi shared reduces the register pressure
enough that depth can fit in registers; as a result, making depth shared
reduces performance.

At least, that is my interpretation.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20170307/afd7b5da/attachment.html>


More information about the intel-3d-bugs mailing list