Execution Tests: Long Vector - Dot fp32 precision#7801
Merged
Conversation
damyanp
approved these changes
Oct 7, 2025
inbelic
approved these changes
Oct 17, 2025
Collaborator
inbelic
left a comment
inbelic
left a comment
There was a problem hiding this comment.
LGTM, can we update the commit description to include "the why"
damyanp
reviewed
Oct 21, 2025
Contributor
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
alsepkow
commented
Oct 23, 2025
damyanp
reviewed
Oct 23, 2025
damyanp
approved these changes
Oct 24, 2025
Member
damyanp
left a comment
damyanp
left a comment
There was a problem hiding this comment.
LGTM, but I haven't deeply reviewed the mathmatics of this, so if you'd like some more review of that you may need to solicit another review.
Contributor
Author
|
Yup, I asked Tex to do a sanity review on that for me.
…________________________________
From: Damyan Pepper ***@***.***>
Sent: Friday, October 24, 2025 9:29 AM
To: microsoft/DirectXShaderCompiler ***@***.***>
Cc: Alex Sepkowski ***@***.***>; Author ***@***.***>
Subject: Re: [microsoft/DirectXShaderCompiler] Execution Tests: Long Vector - Dot fp32 precision (PR #7801)
@damyanp approved this pull request.
LGTM, but I haven't deeply reviewed the mathmatics of this, so if you'd like some more review of that you may need to solicit another review.
—
Reply to this email directly, view it on GitHub<#7801 (review)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABK4EW43ZZHBARQNU4HM2WL3ZJHYDAVCNFSM6AAAAACIRDH2TSVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZTGNZXG44DCNRQGQ>.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
reviewed
Oct 27, 2025
tex3d
approved these changes
Oct 28, 2025
…omething for corner cases like subnormals
Co-authored-by: Tex Riddell <texr@microsoft.com>
Co-authored-by: Tex Riddell <texr@microsoft.com>
alsepkow
force-pushed
the
user/alsepkow/DotPrecision
branch
from
a6ae2ea to
3391ef6
Compare
tex3d
reviewed
Oct 29, 2025
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Update Dot expected values to be computed using doubles to improve precision of the final result.
Additionally, this PR adds logic to compute an absolute epsilon tolerance for the final result of each Dot operation by summing absolute epsilons calculated on each multiply and addition in the dot product. The summation portion is re-ordered to a worst-case ordering when computing the absolute epsilon error tolerance as the DX spec allows for this.