Skip to content

[SM 6.10][LinAlg] Fixes float truncation and coordinate calculation in LinAlg exec tests #8420

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
tex3d merged 6 commits into from
May 7, 2026
Merged

[SM 6.10][LinAlg] Fixes float truncation and coordinate calculation in LinAlg exec tests #8420

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
106f381
Fixes float truncation and coordinate calculation in LinAlg exec tests
anupamachandra May 4, 2026
e8f988b
Update tools/clang/unittests/HLSLExec/LinAlgTests.cpp
anupamachandra May 6, 2026
b5c6442
Merge branch 'microsoft:main' into anupamac/linalg-exectest-fixes-01
anupamachandra May 6, 2026
cc624bd
Suffix fill value literals with F to ensure correct behavior outside …
anupamachandra May 6, 2026
6c7e20a
Add macro STREAM_FLOAT to stream float values
anupamachandra May 6, 2026
23cc309
clang format failure fix
anupamachandra May 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 15 additions & 11 deletions tools/clang/unittests/HLSLExec/LinAlgTests.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,10 @@
#include <optional>
#include <sstream>
#include <string>

#define STREAM_FLOAT(stream, name, value) \
stream << std::showpoint << " -D" << name << "=" << value << "F" \
<< std::noshowpoint
Copy link
Copy Markdown
Contributor

@tex3d tex3d May 7, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

While this is fine, I feel like the name makes it sound more generic, as if it might be defined like so:

#define STREAM_FLOAT(value) std::showpoint << value << "F" << std::noshowpoint

And used like so:

ExtraDefs << "-DFILL_VALUE=" << STREAM_FLOAT(FillValue);

#include <variant>
#include <vector>

Expand Down Expand Up @@ -505,7 +509,7 @@ static void runSplatStore(ID3D12Device *Device,
const size_t BufferSize = Params.totalBytes();

std::stringstream ExtraDefs;
ExtraDefs << "-DFILL_VALUE=" << FillValue;
STREAM_FLOAT(ExtraDefs, "FILL_VALUE", FillValue);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down Expand Up @@ -624,7 +628,7 @@ static const char ElementAccessShader[] = R"(
// flatten the 2D index into a 1D index then scale by element size
// Always store row-major and work it out in the test runner
uint coordToByteOffset(uint2 coord) {
return (coord.y * M_DIM + coord.x) * ELEM_SIZE;
return (coord.x * N_DIM + coord.y) * ELEM_SIZE;
Copy link
Copy Markdown
Contributor

@tex3d tex3d May 4, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. It sure can be confusing for x to be row and y to be column, but makes most sense in context.

Copy link
Copy Markdown
Collaborator

@V-FEXrt V-FEXrt May 6, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was also vague in the spec until very recently. Thanks for the fix!

}

[WaveSize(4, 64)]
Expand Down Expand Up @@ -936,8 +940,8 @@ static void runMatMatMul(ID3D12Device *Device,

std::stringstream ExtraDefs;
ExtraDefs << " -DK_DIM=" << K;
ExtraDefs << " -DA_FILL=" << AFill;
ExtraDefs << " -DB_FILL=" << BFill;
STREAM_FLOAT(ExtraDefs, "A_FILL", AFill);
STREAM_FLOAT(ExtraDefs, "B_FILL", BFill);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down Expand Up @@ -1018,9 +1022,9 @@ static void runMatMatMulAccum(ID3D12Device *Device,

std::stringstream ExtraDefs;
ExtraDefs << " -DK_DIM=" << K;
ExtraDefs << " -DA_FILL=" << AFill;
ExtraDefs << " -DB_FILL=" << BFill;
ExtraDefs << " -DC_FILL=" << CFill;
STREAM_FLOAT(ExtraDefs, "A_FILL", AFill);
STREAM_FLOAT(ExtraDefs, "B_FILL", BFill);
STREAM_FLOAT(ExtraDefs, "C_FILL", CFill);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down Expand Up @@ -1094,8 +1098,8 @@ static void runMatAccum(ID3D12Device *Device,
const size_t BufferSize = Params.totalBytes();

std::stringstream ExtraDefs;
ExtraDefs << " -DLHS_FILL=" << LHSFill;
ExtraDefs << " -DRHS_FILL=" << RHSFill;
STREAM_FLOAT(ExtraDefs, "LHS_FILL", LHSFill);
STREAM_FLOAT(ExtraDefs, "RHS_FILL", RHSFill);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down Expand Up @@ -1552,7 +1556,7 @@ static void runStoreMemory(ID3D12Device *Device,

std::stringstream ExtraDefs;
ExtraDefs << " -DOFFSET=" << 0;
ExtraDefs << " -DFILL_VALUE=" << FillValue;
STREAM_FLOAT(ExtraDefs, "FILL_VALUE", FillValue);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down Expand Up @@ -1632,7 +1636,7 @@ static void runAccumulateMemory(ID3D12Device *Device,

std::stringstream ExtraDefs;
ExtraDefs << " -DOFFSET=" << 0;
ExtraDefs << " -DFILL_VALUE=" << FillValue;
STREAM_FLOAT(ExtraDefs, "FILL_VALUE", FillValue);

std::string Args = buildCompilerArgs(Params, ExtraDefs.str().c_str());

Expand Down
Loading