Skip to content

[SYCL][E2E][Joint Matrix] Added -ffp-model=precise to some tests #12429

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
merged 1 commit into from
Jan 18, 2024
Merged

[SYCL][E2E][Joint Matrix] Added -ffp-model=precise to some tests #12429

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/SG32/joint_matrix_bf16_fill_k_cache.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,9 +8,11 @@
// REQUIRES: matrix
// REQUIRES-INTEL-DRIVER: lin: 27501, win: 101.4943

// RUN: %{build} -o %t.out
// RUN: %{build} -o %t.out -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>

Expand Down
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/SG32/joint_matrix_bf16_fill_k_cache_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,9 +8,11 @@
// REQUIRES: matrix, gpu
// REQUIRES-INTEL-DRIVER: lin: 27501, win: 101.4943

// RUN: %{build} -o %t.out -DINIT_LIST
// RUN: %{build} -o %t.out -DINIT_LIST -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>

Expand Down
5 changes: 3 additions & 2 deletions sycl/test-e2e/Matrix/SG32/joint_matrix_bf16_fill_k_cache_unroll.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,14 +8,15 @@
// REQUIRES: matrix
// REQUIRES-INTEL-DRIVER: lin: 27501, win: 101.4943

// RUN: %{build} -mllvm -inline-threshold=5000 -o %t_gpu.out -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=5000 -ffp-model=precise -o %t_gpu.out -DMANUAL_UNROLL
// RUN: %if gpu %{ %{run} %t_gpu.out %}

// RUN: %{build} -mllvm -inline-threshold=5000 -o %t_cpu.out -DMANUAL_UNROLL -DtM=16 -DtK=32 -DNCACHE1=32 -DKCACHE1=32
// RUN: %{build} -mllvm -inline-threshold=5000 -ffp-model=precise -o %t_cpu.out -DMANUAL_UNROLL -DtM=16 -DtK=32 -DNCACHE1=32 -DKCACHE1=32
// RUN: %if cpu %{ %{run} %t_cpu.out %}

// -mllvm -inline-threshold added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>
Expand Down
3 changes: 2 additions & 1 deletion sycl/test-e2e/Matrix/SG32/joint_matrix_bf16_fill_k_cache_unroll_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,11 +8,12 @@
// REQUIRES: matrix, gpu
// REQUIRES-INTEL-DRIVER: lin: 27501, win: 101.4943

// RUN: %{build} -mllvm -inline-threshold=5000 -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=5000 -ffp-model=precise -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{run} %t_gpu.out

// -mllvm -inline-threshold added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>
Expand Down
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/XMX8/joint_matrix_bf16_fill_k_cache.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,11 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix-xmx8

// RUN: %{build} -o %t.out
// RUN: %{build} -o %t.out -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>

Expand Down
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/XMX8/joint_matrix_bf16_fill_k_cache_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,11 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix-xmx8

// RUN: %{build} -o %t.out -DINIT_LIST
// RUN: %{build} -o %t.out -DINIT_LIST -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>

Expand Down
3 changes: 2 additions & 1 deletion sycl/test-e2e/Matrix/XMX8/joint_matrix_bf16_fill_k_cache_unroll.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,11 +7,12 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix-xmx8

// RUN: %{build} -mllvm -inline-threshold=2000 -o %t.out -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=2000 -ffp-model=precise -o %t.out -DMANUAL_UNROLL
// RUN: %{run} %t.out

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>
Expand Down
3 changes: 2 additions & 1 deletion sycl/test-e2e/Matrix/XMX8/joint_matrix_bf16_fill_k_cache_unroll_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,11 +7,12 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix-xmx8

// RUN: %{build} -mllvm -inline-threshold=2000 -o %t.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=2000 -ffp-model=precise -o %t.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{run} %t.out

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "../common.hpp"
#include <cstddef>
Expand Down
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,11 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix

// RUN: %{build} -o %t.out
// RUN: %{build} -o %t.out -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
#include <cstddef>

Expand Down
4 changes: 3 additions & 1 deletion sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,11 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix, gpu

// RUN: %{build} -o %t.out -DINIT_LIST
// RUN: %{build} -o %t.out -DINIT_LIST -ffp-model=precise
// RUN: %{run} %t.out

// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
#include <cstddef>
#define SG_SZ 16
Expand Down
5 changes: 3 additions & 2 deletions sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache_unroll.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,14 +7,15 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix

// RUN: %{build} -mllvm -inline-threshold=2000 -o %t_gpu.out -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=2000 -ffp-model=precise -o %t_gpu.out -DMANUAL_UNROLL
// RUN: %if gpu %{ %{run} %t_gpu.out %}

// RUN: %{build} -mllvm -inline-threshold=2000 -o %t_cpu.out -DMANUAL_UNROLL -DtM=16 -DtK=32 -DNCACHE1=32 -DKCACHE1=32
// RUN: %{build} -mllvm -inline-threshold=2000 -ffp-model=precise -o %t_cpu.out -DMANUAL_UNROLL -DtM=16 -DtK=32 -DNCACHE1=32 -DKCACHE1=32
// RUN: %if cpu %{ %{run} %t_cpu.out %}

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
#include <cstddef>
Expand Down
3 changes: 2 additions & 1 deletion sycl/test-e2e/Matrix/joint_matrix_bf16_fill_k_cache_unroll_init.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,11 +7,12 @@
//===----------------------------------------------------------------------===//
// REQUIRES: matrix, gpu

// RUN: %{build} -mllvm -inline-threshold=2000 -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{build} -mllvm -inline-threshold=2000 -ffp-model=precise -o %t_gpu.out -DINIT_LIST -DMANUAL_UNROLL
// RUN: %{run} %t_gpu.out

// -mllvm -inline-threshold=2000 added as a workaround,
// since IGC doesn't support some variants of IR for Joint Matrix currently
// -ffp-model=precise is added to not depend on compiler defaults.

#include "common.hpp"
#include <cstddef>
Expand Down