Skip to content
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

Enable GEMM/dot for FP8 using hipblasLT #3577

Merged
merged 20 commits into from
Nov 13, 2024
Merged

Enable GEMM/dot for FP8 using hipblasLT #3577

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
20 commits
Select commit Hold shift + click to select a range
e6e3a5a
Initial
CharlieL7 Oct 3, 2024
dd60b39
Remove conditional for now
CharlieL7 Oct 3, 2024
a3ded15
Merge branch 'develop' of github.com:ROCm/AMDMIGraphX into ocp_fp8_hi…
CharlieL7 Oct 30, 2024
3628c7a
Add check for the hipblas gemm flag
CharlieL7 Oct 30, 2024
f28d40d
Declare env var, disable fp8ocp gemm with no intrinsics
CharlieL7 Oct 30, 2024
8cc9914
Use enabled
CharlieL7 Nov 1, 2024
33a755e
Merge branch 'develop' into ocp_fp8_hipblaslt
CharlieL7 Nov 1, 2024
139d12d
Merge branch 'develop' into ocp_fp8_hipblaslt
causten Nov 4, 2024
856b6cb
Turn off unsupported hipblas types
CharlieL7 Nov 5, 2024
8e00c91
Hipblaslt types check gfx number
CharlieL7 Nov 5, 2024
af9d6e3
revert to previous
CharlieL7 Nov 5, 2024
223ba4d
add preprocessor conditional on hipblaslt version
CharlieL7 Nov 5, 2024
7f5798c
Use CMake macro to check symbols (needs testing)
CharlieL7 Nov 7, 2024
0506abd
Revert " Use CMake macro to check symbols (needs testing)"
CharlieL7 Nov 7, 2024
1c52a93
Merge branch 'develop' into ocp_fp8_hipblaslt
CharlieL7 Nov 7, 2024
d453fdc
Use variable defined by hipblaslt for HIP float8 ocp types
CharlieL7 Nov 9, 2024
6fe6af2
Merge branch 'develop' of github.com:ROCmSoftwarePlatform/AMDMIGraphX…
CharlieL7 Nov 9, 2024
78988d0
Merge branch 'develop' into ocp_fp8_hipblaslt
CharlieL7 Nov 12, 2024
92c1e37
Merge branch 'develop' into ocp_fp8_hipblaslt
causten Nov 12, 2024
519a63b
Merge branch 'develop' into ocp_fp8_hipblaslt
causten Nov 13, 2024
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
6 changes: 6 additions & 0 deletions src/targets/gpu/hip_gemm_impl.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,8 +70,14 @@ hipDataType get_type_hipblas(shape::type_t type)
case shape::int32_type: return HIP_R_32I;
case shape::uint32_type: return HIP_R_32U;
case shape::fp8e4m3fnuz_type: return HIP_R_8F_E4M3_FNUZ;
// TODO can remove this preprocessor conditional when hip verison defaults to have these types
#ifdef ROCM_USE_FLOAT8
case shape::fp8e4m3fn_type: return HIP_R_8F_E4M3;
case shape::fp8e5m2_type: return HIP_R_8F_E5M2;
#else
case shape::fp8e4m3fn_type:
case shape::fp8e5m2_type:
#endif
case shape::tuple_type:
case shape::bool_type:
case shape::uint16_type:
Expand Down
12 changes: 9 additions & 3 deletions src/targets/gpu/target.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,6 +81,7 @@ MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_ENABLE_NHWC)
#ifndef _WIN32
MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_ENABLE_CK)
#endif
MIGRAPHX_DECLARE_ENV_VAR(MIGRAPHX_ENABLE_HIPBLASLT_GEMM)

std::vector<pass> target::get_passes(migraphx::context& gctx, const compile_options& options) const
{
Expand Down Expand Up @@ -129,9 +130,12 @@ std::vector<pass> target::get_passes(migraphx::context& gctx, const compile_opti
unsupported_fp8e4m3fnuz_ops.insert("argmin");

std::set<std::string> unsupported_fp8ocp_ops = {};
// TODO update with hipBLASLt support
unsupported_fp8ocp_ops.insert("dot");
unsupported_fp8ocp_ops.insert("quant_dot");
// TODO: remove this when the flag is removed
if(not enabled(MIGRAPHX_ENABLE_HIPBLASLT_GEMM{}))
{
unsupported_fp8ocp_ops.insert("dot");
unsupported_fp8ocp_ops.insert("quant_dot");
}
#if MIGRAPHX_USE_MIOPEN
// MIOpen doesn't have support for fp8 pooling yet.
unsupported_fp8ocp_ops.insert("pooling");
Expand All @@ -140,6 +144,8 @@ std::vector<pass> target::get_passes(migraphx::context& gctx, const compile_opti
{
unsupported_fp8ocp_ops.insert("convolution");
unsupported_fp8ocp_ops.insert("quant_convolution");
unsupported_fp8ocp_ops.insert("dot");
unsupported_fp8ocp_ops.insert("quant_dot");
}
// add all device kernels
unsupported_fp8ocp_ops.insert("logsoftmax");
Expand Down
Loading