Page MenuHomePhabricator

[AMDGPU] Add TransVALU to gfx10
ClosedPublic

Authored by sebastian-ne on Apr 8 2021, 9:09 AM.

Details

Summary

Instructions on the transcendental unit are executed in parallel to the
normal VALU, so add this as an extra resource.

This doesn't seem to have any effect, but it should be more correct.

Diff Detail

Event Timeline

sebastian-ne created this revision.Apr 8 2021, 9:09 AM
sebastian-ne requested review of this revision.Apr 8 2021, 9:09 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 8 2021, 9:09 AM
foad added a comment.Apr 8 2021, 9:23 AM

I have no objection to this as a first step in the right direction, but I think it should have no effect at all at the moment. The schedmodel already knows that we can only issue one instruction per cycle, and this patch does not affect that. It only becomes interesting when you model the fact that issuing a trans instruction consumes the trans ALU for (say) 4 cycles, so that on the following 3 cycles you could issue a normal VALU instruction but *not* another trans instruction.

rampitec accepted this revision.Apr 8 2021, 12:06 PM

LGTM. What about other models?

This revision is now accepted and ready to land.Apr 8 2021, 12:06 PM

What about other models?

The parallel execution is only a thing on gfx10+.

This revision was automatically updated to reflect the committed changes.