Skip to content

Latest commit

 

History

History
161 lines (137 loc) · 7.94 KB

CHANGELOG.md

File metadata and controls

161 lines (137 loc) · 7.94 KB

Change Log for rocBLAS

Full documentation for rocBLAS is available at rocblas.readthedocs.io.

[rocBLAS 2.40.0 for ROCm 4.4.0]

Optimizations

  • Improved performance of non-batched and batched dot, dotc, and dot_ex for small n. e.g. sdot n <= 31000.
  • Improved performance of non-batched and batched trmv for all sizes and matrix types.
  • Improved performance of non-batched and batched gemv transpose case for all sizes and datatypes.
  • Improved performance of sger and dger for all sizes, in particular the larger dger sizes.
  • Improved performance of syrkx for for large size including those in rocBLAS Issue #1184.

[rocBLAS 2.39.0 for ROCm 4.3.0]

Optimizations

  • Improved performance of non-batched and batched rocblas_Xgemv for gfx908 when m <= 15000 and n <= 15000
  • Improved performance of non-batched and batched rocblas_sgemv and rocblas_dgemv for gfx906 when m <= 6000 and n <= 6000
  • Improved the overall performance of non-batched and batched rocblas_cgemv for gfx906
  • Improved the overall performance of rocblas_Xtrsv

Changed

  • Internal use only APIs prefixed with rocblas_internal_ and deprecated to discourage use

[rocBLAS 2.38.0 for ROCm 4.2.0]

Added

  • Added option to install script to build only rocBLAS clients with a pre-built rocBLAS library
  • Supported gemm ext for unpacked int8 input layout on gfx908 GPUs
    • Added new flags rocblas_gemm_flags::rocblas_gemm_flags_pack_int8x4 to specify if using the packed layout
      • Set the rocblas_gemm_flags_pack_int8x4 when using packed int8x4, this should be always set on GPUs before gfx908.
      • For gfx908 GPUs, unpacked int8 is supported so no need to set this flag.
      • Notice the default flags 0 uses unpacked int8, this somehow changes the behaviour of int8 gemm from ROCm 4.1.0
  • Added a query function rocblas_query_int8_layout_flag to get the preferable layout of int8 for gemm by device

Optimizations

  • Improved performance of single precision copy, swap, and scal when incx == 1 and incy == 1.
  • Improved performance of single precision axpy when incx == 1, incy == 1 and batch_count =< 8192.
  • Improved performance of trmm.

Changed

  • Change cmake_minimum_required to VERSION 3.16.8

[rocBLAS 2.36.0 for ROCm 4.1.0]

Added

  • Added Numerical checking helper function to detect zero/NaN/Inf in the input and the output vectors of rocBLAS level 1 and 2 functions.
  • Added Numerical checking helper function to detect zero/NaN/Inf in the input and the output general matrices of rocBLAS level 2 and 3 functions.

Fixed

  • Fixed complex unit test bug caused by incorrect caxpy and zaxpy function signatures.
  • Make functions compliant with Legacy Blas for special values alpha == 0, k == 0, beta == 1, beta == 0.

Optimizations

  • Improved performance of single precision axpy_batched and axpy_strided_batched: batch_count >= 8192.
  • Improved performance of trmm.

[rocBLAS 2.34.0 for ROCm 4.0.0]

Added

  • Add changelog.
  • Improved performance of gemm_batched for small m, n, k and NT, NC, TN, TT, TC, CN, CT, CC.
  • Improved performance of gemv, gemv_batched, gemv_strided_batched: small n large m.
  • Removed support for legacy hcc compiler.
  • Add rot_ex, rot_batched_ex, and rot_strided_batched_ex.

Fixed

  • Removed -DUSE_TENSILE_HOST from roc::rocblas CMake usage requirements. This is a rocblas internal variable, and does not need to be defined in user code.

[rocBLAS 2.32.0 for ROCm 3.10.0]

Added

  • Improved performance of gemm_batched for NN, general m, n, k, small m, n, k.

[rocBLAS 2.30.0 for ROCm 3.9.0]

Added

  • Slight improvements to FP16 Megatron BERT performance on MI50.
  • Improvements to FP16 Transformer performance on MI50.
  • Slight improvements to FP32 Transformer performance on MI50.
  • Improvements to FP32 DLRM Terabyte performance on gfx908.

[rocBLAS 2.28.0 for ROCm 3.8.0]

Added

  • added two functions:

    • rocblas_status rocblas_set_atomics_mode(rocblas_atomics_mode mode)
    • rocblas_status rocblas_get_atomics_mode(rocblas_atomics_mode mode)

  • added enum rocblas_atomics_mode. It can have two values

    • rocblas_atomics_allowed
    • rocblas_atomics_not_allowed The default is rocblas_atomics_not_allowed

  • function rocblas_Xdgmm algorithm corrected and incx=0 support added

  • dependencies:

    • rocblas-tensile internal component requires msgpack instead of LLVM

  • Moved the following files from /opt/rocm/include to /opt/rocm/include/internal:

    • rocblas-auxillary.h
    • rocblas-complex-types.h
    • rocblas-functions.h
    • rocblas-types.h
    • rocblas-version.h
    • rocblas_bfloat16.h

    These files should NOT be included directly as this may lead to errors. Instead, /opt/rocm/include/rocblas.h should be included directly. /opt/rocm/include/rocblas_module.f90 can also be direcly used.

[rocBLAS 2.26.0 for ROCm 3.7.0]

Added

  • Improvements to rocblas_Xgemm_batched performance for small m, n, k.
  • Improvements to rocblas_Xgemv_batched and rocblas_Xgemv_strided_batched performance for small m (QMCPACK use).
  • Improvements to rocblas_Xdot (batched and non-batched) performance when both incx and incy are 1.
  • Improvements to FP32 ONNX BERT performance for MI50.
  • Significant improvements to FP32 Resnext, Inception Convolution performance for gfx908.
  • Slight improvements to FP32 DLRM Terabyte performance for gfx908.
  • Significant improvements to FP32 BDAS performance for gfx908.
  • Significant improvements to FP32 BDAS performance for MI50 and MI60.
  • Added substitution method for small trsm sizes with m <= 64 && n <= 64. Increases performance drastically for small batched trsm.

[rocBLAS 2.24.0 for ROCm 3.6.0]

Added

  • Improvements to User Guide and Design Document.
  • L1 dot function optimized to utilize shuffle instructions ( improvements on bf16, f16, f32 data types ).
  • L1 dot function added x dot x optimized kernel.
  • Standardization of L1 rocblas-bench to use device pointer mode to focus on GPU memory bandwidth.
  • Adjustments for hipcc (hip-clang) compiler as standard build compiler and Centos8 support.
  • Added Fortran interface for all rocBLAS functions.

[rocBLAS 2.22.0 for ROCm 3.5.0]

Added

  • add geam complex, geam_batched, and geam_strided_batched.
  • add dgmm, dgmm_batched, and dgmm_strided_batched.
  • Optimized performance
    • ger
      • rocblas_sger, rocblas_dger,
      • rocblas_sger_batched, rocblas_dger_batched
      • rocblas_sger_strided_batched, rocblas_dger_strided_batched
    • geru
      • rocblas_cgeru, rocblas_zgeru
      • rocblas_cgeru_batched, rocblas_zgeru_batched
      • rocblas_cgeru_strided_batched, rocblas_zgeru_strided_batched
    • gerc
      • rocblas_cgerc, rocblas_zgerc
      • rocblas_cgerc_batched, rocblas_zgerc_batched
      • rocblas_cgerc_strided_batched, rocblas_zgerc_strided_batched
    • symv
      • rocblas_ssymv, rocblas_dsymv, rocblas_csymv, rocblas_zsymv
      • rocblas_ssymv_batched, rocblas_dsymv_batched, rocblas_csymv_batched, rocblas_zsymv_batched
      • rocblas_ssymv_strided_batched, rocblas_dsymv_strided_batched, rocblas_csymv_strided_batched, rocblas_zsymv_strided_batched
    • sbmv
      • rocblas_ssbmv, rocblas_dsbmv
      • rocblas_ssbmv_batched, rocblas_dsbmv_batched
      • rocblas_ssbmv_strided_batched, rocblas_dsbmv_strided_batched
    • spmv
      • rocblas_sspmv, rocblas_dspmv
      • rocblas_sspmv_batched, rocblas_dspmv_batched
      • rocblas_sspmv_strided_batched, rocblas_dspmv_strided_batched
  • improved documentation.
  • Fix argument checking in functions to match legacy BLAS.
  • Fixed conjugate-transpose version of geam.

Known Issues

  • Compilation for GPU Targets: When using the install.sh script for "all" GPU Targets, which is the default, you must first set an environment variable HCC_AMDGPU_TARGET listing the GPU targets, e.g. HCC_AMDGPU_TARGET=gfx803,gfx900,gfx906,gfx908 If building for a specific architecture(s) using the -a | --architecture flag, you should also set the environment variable HCC_AMDGPU_TARGET to match. Mismatching the environment variable to the -a flag architectures creates builds that may result in SEGFAULTS when running on GPUs which weren't specified.