[Bug 97035] [regression] WebGL demo iamnop.com/particles is very slow and spends far too long compiling shaders

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=97035

--- Comment #5 from Eero Tamminen <eero.t.tamminen at intel.com> ---
(In reply to Eero Tamminen from comment #3)
> Tested this on SKL GT2 with today's Mesa.
> 
> All the other shapes selectable in the WebGL demo load instantly and run
> well over 60FPS.  However, selecting the default "galaxy" shape takes ~20s
> and runs at ~2FPS.

With Mesa 11.2, compilation takes only few seconds and "galaxy" runs at full
speed.

When comparing at the generated assembly, Mesa 11.2 spills a bit:
SIMD8 shader: 4432 instructions. 0 loops. 35394 cycles. 81:179 spills:fills.
Promoted 18 constants. Compacted 70912 to 48880 bytes (31%)

But the latest Mesa spills a lot more:
SIMD8 shader: 7985 instructions. 0 loops. 57210 cycles. 359:993 spills:fills.
Promoted 19 constants. Compacted 127760 to 86032 bytes (33%)

I'll try whether I can locate the regression in this with eZBench:
  https://cgit.freedesktop.org/ezbench/tree/README


Note: When looking at the SIMD8 shader code generated by the Windows driver, it
doesn't spill at all.

-- 
You are receiving this mail because:
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/20160901/c99711af/attachment.html>