[Bug 97035] compilation of WebGL demo iamnop.com/particles shader slowed 5x, run-time perf dropped to 1/10th

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Wed May 17 11:32:08 UTC 2017


--- Comment #13 from Eero Tamminen <eero.t.tamminen at intel.com> ---
(In reply to Eero Tamminen from comment #12)
> Tested latest Mesa.  On my SKL GT2 the slow shader still takes ~20s to
> compile and produces following Mesa statistics:
> SIMD8 shader: 7985 instructions. 0 loops. 57210 cycles. 359:993
> spills:fills. Promoted 19 constants. Compacted 127760 to 86032 bytes (33%)

Latest Mesa (i.e. 5 months later) uses 17-18ms to compile on that machine, and
gets following statistics:
SIMD8 shader: 8127 instructions. 0 loops. 57436 cycles. 369:1019 spills:fills.
Promoted 18 constants. Compacted 130032 to 87472 bytes (33%)

I.e. there's been a minor improvement in compile time, but regression from
v11.2 is still valid and spilling & instruction counts have gotten slightly
worse. "perf" output looks same as earlier.

You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20170517/dd2bd5b7/attachment.html>

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