Commits

Branch selector

 038d142

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 2, 2024

• Add gemv dispatch for fp16 (#27)

  vinx13

  authored
  038d142

  Copy full SHA for 038d142

Commits on Feb 1, 2024

• Add tvm bindings for MoE gemm (#26)

  vinx13

  authored
  159cf8d

  Copy full SHA for 159cf8d

Commits on Jan 23, 2024

• Use col major for f16 MoE gemm and add small batch specialization (#25)

  vinx13

  authored
  de67b3f

  Copy full SHA for de67b3f

Commits on Jan 21, 2024

• fix incorrect result of gemm_residual (#24)

  jinzhen-lin

  authored
  0e39840

  Copy full SHA for 0e39840

Commits on Jan 11, 2024

• Use sm80 for Hopper

  vinx13

  committed
  6f05dd4

  Copy full SHA for 6f05dd4

Commits on Jan 8, 2024

• Add tvm binding with residual

  

  cyx-6

  authored and

  junrushao

  committed
  6c05e0e

  Copy full SHA for 6c05e0e

• Revert "Add tvm binding for moe gemm kernels (#20)"

  Show description for beda1ee
  

  cyx-6

  authored and

  junrushao

  committed
  beda1ee

  Copy full SHA for beda1ee

Commits on Jan 5, 2024

• Remove duplicated function declaration

  vinx13

  committed
  37d957f

  Copy full SHA for 37d957f

• Add tvm binding for moe gemm kernels (#20)

  cyx-6

  authored
  d79cb6a

  Copy full SHA for d79cb6a

Commits on Jan 2, 2024

• Remove Hardcoded CMAKE_CUDA_FLAGS (#19)

  Show description for 12e0cc7
  junrushao

  authored
  12e0cc7

  Copy full SHA for 12e0cc7

Commits on Dec 30, 2023

• Add tvm binding

  

  cyx-6

  authored and

  junrushao

  committed
  1243f91

  Copy full SHA for 1243f91

Commits on Dec 17, 2023

• Remove invalid use of __CUDA_ARCH__

  Show description for 74ee6cb
  vinx13

  committed
  74ee6cb

  Copy full SHA for 74ee6cb

Commits on Dec 15, 2023

• Add moe gemv kernel (#16)

  vinx13

  authored
  ad22779

  Copy full SHA for ad22779

• Fix compilation warnings

  vinx13

  committed
  245cbb3

  Copy full SHA for 245cbb3

Commits on Dec 14, 2023

• Fix template instantiation

  vinx13

  committed
  d489188

  Copy full SHA for d489188

Commits on Dec 13, 2023

• Fix preprocession and compilation of moe kernels (#15)

  vinx13

  authored
  7191fbd

  Copy full SHA for 7191fbd

Commits on Dec 9, 2023

• Update preprocessor for moe kernels (#14)

  vinx13

  authored
  882c667

  Copy full SHA for 882c667

Commits on Dec 8, 2023

• Add moe gemm kernel (#13)

  vinx13

  authored
  fb25693

  Copy full SHA for fb25693

Commits on Dec 6, 2023

• Add guard for old cuda arch

  vinx13

  committed
  ed951b0

  Copy full SHA for ed951b0

Commits on Dec 5, 2023

• Add weight only batched gemv kernels for small batch size (#11)

  Show description for 3922170
  vinx13

  authored
  3922170

  Copy full SHA for 3922170

Commits on Dec 2, 2023

• Let sm90 use sm80 kernel

  masahi

  committed
  6b744ba

  Copy full SHA for 6b744ba

Commits on Nov 21, 2023

• Inline functions defined in the header

  vinx13

  committed
  942c01d

  Copy full SHA for 942c01d

• Split template instantiation to speed up compilation (#10)

  vinx13

  authored
  c6b6cc8

  Copy full SHA for c6b6cc8

Commits on Nov 18, 2023

• Fix compilation on old cuda arch (#9)

  vinx13

  authored
  5dac670

  Copy full SHA for 5dac670

Commits on Nov 16, 2023

• Add fine grained quantization kernel (#8)

  Show description for a4ee22b
  vinx13

  authored
  a4ee22b

  Copy full SHA for a4ee22b

• Integrate upstream cutlass_extensions changes (#7)

  Show description for 9e34a94
  vinx13

  authored
  9e34a94

  Copy full SHA for 9e34a94

• Add clang-format config and apply formatter (#6)

  vinx13

  authored
  145bee4

  Copy full SHA for 145bee4

Commits on Jul 14, 2023

• disable strict-aliasing warning for GCC

  masahi

  committed
  390e821

  Copy full SHA for 390e821

Commits on Jul 13, 2023

• Add bias stride parameter (#4)

  masahi

  committed
  a864a06

  Copy full SHA for a864a06

Commits on Jul 12, 2023

• Add int8 support and clean up API (#3)

  masahi

  authored
  2c71bdc

  Copy full SHA for 2c71bdc

Commits on Jun 22, 2023

• Volta and Turing support (#2)

  Show description for c9f37e3
  masahi

  authored
  c9f37e3

  Copy full SHA for c9f37e3

Commits on Jun 14, 2023

• fix

  masahi

  committed
  634184f

  Copy full SHA for 634184f

• Merge pull request #1 from tlc-pack/residual-fusion

  Show description for 323dd4b
  masahi

  authored
  323dd4b

  Copy full SHA for 323dd4b

• more act

  masahi

  committed
  2356986

  Copy full SHA for 2356986

Commits on Jun 9, 2023

• implement residual dispatch

  masahi

  committed
  3f9662f

  Copy full SHA for 3f9662f

Pagination

Previous

Next