triton-lang/triton-cpu

Issues

• error: LLVM Translation failed for operation: builtin.unrealized_conversion_cast

  #144 opened 4 days ago
  1

• clang release build error

  #143 opened 3 days ago
  1

• Masked load + integer division issue

  #142 opened 6 days ago
  4

• When I use the command pip install -e python in the triton-cpu folder I get an error

  #134 opened 7 days ago
  3

• when I use cmake/make to compile the Triton source code, I encountered the error "cannot find -lNVGPUIR: No such file or directory."

  #133 opened 7 days ago
  2

• Implement device_assert

  #125 opened 12 days ago
  1

• How should we approach libdevice calls that don't map to a libmvec function?

  #111 opened 21 days ago
  2

• Enable more pytests

  #107 opened a month ago
  6

• OptimizeMask pass fails on following input

  #105 opened a month ago
  1

• Leverage BF16 instructions for AVX512 as well

  #87 opened a month ago
  0

• Compile time explosion with some inputs

  #86 opened 22 days ago
  3

• Support tl.debug_barrier

  #85 opened a month ago
  1

• Fix/support masks in the GEMV example

  #82 opened a month ago
  0

• Pytests are taking too long

  #79 opened a month ago
  3

• Locating libTritonCPURuntime.so only works for Python < 3.12

  #76 opened a month ago
  4

• No debug info on Darwin

  #72 opened 2 months ago
  2

• builtin.unrealized_conversion_cast not properly lowered

  #71 opened 2 months ago
  1

• Explore vector dialect transformation options

  #68 opened 2 months ago
  0

• 保存中间的过程文件

  #63 opened 2 months ago
  1

• Element-wise ops should work with scalars too

  #62 opened 2 months ago
  0

• Block pointer padding_option="nan" not working

  #60 opened 2 months ago
  0

• Round-to-zero float32 conversion not supported

  #58 opened 2 months ago
  1

• Assertion error with trivial kernel

  #57 opened 2 months ago
  1

• Consider to combine enable_fp_fusion and enable_fast_math in upstream triton

  #50 opened 2 months ago
  1

• Utilize vector math functions in generated kernels

  #42 opened a month ago
  1

• Investigate a potential usage of LinearLayout for TritonCPU

  #38 opened 3 months ago
  0

• Supporting mixed precisions in matmul

  #37 opened 2 months ago
  0

• Fully understanding the vector arithmetic performance against Torch

  #36 opened 3 days ago
  0

• device_print

  #35 opened 23 days ago
  2