hipBLASLt 0.10.0 for ROCm 6.3.0

Added

• Support the V2 CPP extension API for backward compatibility
• Support for data type Int8 in with Int8 out
• Support for data type FP32/FP64 for gfx110x
• Add the Extension API hipblaslt_ext::matmulIsTuned
• Output atol and rtol for hipblaslt-bench validation
• Output the bench command for hipblaslt CPP ext API path if HIPBLASLT_LOG_MASK=32 is set
• Support odd sizes for FP8/BF8 GEMM

Changed

• Reorganize and add more sample code
• Add a dependency with the hipblas-common package and remove the dependency with the hipblas package

Optimized

• Support fused kernel for HIPBLASLT_MATMUL_DESC_AMAX_D_POINTER for FP8/BF8 data type
• Improve the library loading time
• Improve the overall performance of first returned solution

Upcoming changes

• The V1 CPP extension API will be deprecated in a future release of hipBLASLt

hipBLASLt 0.8.0 for ROCm 6.2.4

hipBLASLt code for ROCm 6.2.4 did not change. The library was rebuilt for the updated ROCm 6.2.4 stack.

hipBLASLt 0.8.0 for ROCm 6.2.2

hipBLASLt code for ROCm 6.2.2 did not change. The library was rebuilt for the updated ROCm 6.2.2 stack.

hipBLASLt 0.8.0 for ROCm 6.2.1

hipBLASLt code for ROCm 6.2.1 did not change. The library was rebuilt for the updated ROCm 6.2.1 stack.

hipBLASLt 0.8.0 for ROCm 6.2.0

hipBLASLt code for ROCm 6.2.0 did not change. The library was rebuilt for the updated ROCm 6.2.0 stack.

hipBLASLt 0.7.0 for ROCm 6.1.2

hipBLASLt code for ROCm 6.1.2 did not change. The library was rebuilt for the updated ROCm 6.1.2 stack.

hipBLASLt 0.7.0 for ROCm 6.1.1

hipBLASLt code for ROCm 6.1.1 did not change. The library was rebuilt for the updated ROCm 6.1.1 stack.

hipBLASLt 0.7.0 for ROCm 6.1.0

Additions

• Added hipblasltExtSoftmax extension API
• Added hipblasltExtLayerNorm extension API
• Added hipblasltExtAMax extension API
• Added GemmTuning extension parameter to set split-k by user
• Support for mix precision datatype: fp16/fp8 in with fp16 out

hipBLASLt 0.6.0 for ROCm 6.0.2

Added

• Add UserArguments for GroupedGemm
• Support datatype: fp16 in with fp32 out
• Add samples
• Support datatype: Int8 in Int32 out
• Support platform gfx94x
• Support fp8/bf8 datatype (only for gfx94x platform)
• Support Scalar A,B,C,D for fp8/bf8 datatype

Changed

• Replace hipblasDatatype_t with hipDataType
• Replace hipblasLtComputeType_t with hipblasComputeType_t
• Deprecate HIPBLASLT_MATMUL_DESC_D_SCALE_VECTOR_POINTER

hipBLASLt 0.6.0 for ROCm 6.0.0

Added

• Add UserArguments for GroupedGemm
• Support datatype: fp16 in with fp32 out
• Add samples
• Support datatype: Int8 in Int32 out
• Support platform gfx94x
• Support fp8/bf8 datatype (only for gfx94x platform)
• Support Scalar A,B,C,D for fp8/bf8 datatype

Changed

• Replace hipblasDatatype_t with hipDataType
• Replace hipblasLtComputeType_t with hipblasComputeType_t
• Deprecate HIPBLASLT_MATMUL_DESC_D_SCALE_VECTOR_POINTER