CNugteren/CLBlast

Issues

• Segmentation fault when consecutively calling CLBlast::Max() and CLBlast::Min()

  #569 opened 11 days ago by diverger
  2

• Find min/max in one function

  #568 opened 11 days ago by diverger
  1

• Possibility of using RDNA3 Matrix Multiplication Instructions to Speed up on AMD GPU

  #566 opened 3 months ago by fancyIX
  4

• Option to round INF to large number instead?

  #567 opened 2 months ago by CodingWinner
  0

• how to implement a new routine from the host side?

  #565 opened 2 months ago by sarithpeiris
  9

• Test #47: clblast_test_xconvgemm ...........Subprocess aborted***Exception: 0.45 sec

  #563 opened 4 months ago by vt-alt
  12

• Windows binaries of PHP extensions

  #564 opened 4 months ago by GajowyJ
  1

• Why clblast::Gemv API is slower than clblast_tune_xgemv when m = 2048 n=16384 precision=fp16.

  #560 opened 4 months ago by liangzelang
  10

• Gemv with half precision cannot get correct result.

  #561 opened 6 months ago by liangzelang
  1

• Banded matrices required buffer size calculated incorrectly (GBMV, HBMV, SBMV & TBMV)

  #538 opened a year ago by BadgerKing7
  8

• Support for 'size_t' as index type for 'Max()' and 'Min()'

  #558 opened 8 months ago by diverger
  0

• What's the meaning of argument 'imax_offset' in clblast::Max()?

  #557 opened 8 months ago by diverger
  1

• Segmentation fault for "_routine_" tuners

  #556 opened 9 months ago by diverger
  1

• Press a key to 'abort' and 'continue' to next in tuning

  #555 opened 8 months ago by diverger
  1

• "make uninstall" support

  #554 opened 9 months ago by diverger
  1

• Tuner stuck in 'dead lock' and never completes

  #546 opened 9 months ago by diverger
  3

• About the performance in different matrix layouts

  #553 opened 9 months ago by diverger
  1

• About the arguments meaning of the matrix operation functions

  #552 opened 9 months ago by diverger
  3

• Link error when call "GemmStridedBatched<cl_float2>"

  #551 opened 9 months ago by diverger
  1

• when i tune GEMM kernel in clblast, i encountered l2 error

  #545 opened 9 months ago by HirataYurina
  2

• ERROR IN ROCK5b

  #547 opened 9 months ago by superHappy-yo
  3

• How to use 'CLBlastSgemmBatched'?

  #550 opened 9 months ago by diverger
  1

• Routines to simply transpose a matrix

  #548 opened 9 months ago by diverger
  1

• 'cublasSdgmm' equivalent support

  #549 opened 9 months ago by diverger
  1

• Segmentation fault with OpenCL 3.0 CUDA (CUDA 12.3)

  #521 opened 9 months ago by fengyuentau
  3

• Accuracy problem on Apple M1 and Intel(R) UHD Graphics 770

  #542 opened 10 months ago by fengyuentau
  12

• SGEMM broken with 1.6.2 on Intel ARC

  #533 opened 10 months ago by 0cc4m
  24

• Android compilation failing

  #537 opened a year ago by samolego
  2

• Tests don't run on Intel Xe/ARC GPU

  #534 opened a year ago by tangjinchuan
  1

• ruby numo-linalg + clblast: OpenCL error: clCreateContext: -6

  #524 opened a year ago by dpblnt
  1

• Version in Python module wrong

  #531 opened a year ago by gspr
  2

• tunner transpose fails on various specific sizes

  #532 opened a year ago by baryluk
  1

• Unparsed options to tunner are ignored, and better handling of platform/device options

  #528 opened a year ago by baryluk
  1

• Build fails with -Werror=format-security

  #529 opened a year ago by gspr
  0

• Consider add SVM Buffer interface support?

  #523 opened a year ago by engineer1109
  3

• gemm performance downgrade for small size M and big size N&K

  #520 opened a year ago by abcdrm
  1

• Binary releases on github are not valid tar.gz files

  #507 opened a year ago by angrave
  0

• USing GPU for CLBLAST (need a tutorial)

  #516 opened a year ago by naturalmechanics
  2

• Consider ad

  #522 opened a year ago by engineer1109
  1

• Pyclblast float16 scalar conversion

  #509 opened a year ago by vathomass
  4

• Do I have to cross-compile both opencl and clblast for android?

  #511 opened a year ago by DavdGao
  2

• need a tutorial on clblast::copy

  #517 opened a year ago by naturalmechanics
  2

• Compilining and running SGMM freezes

  #514 opened a year ago by naturalmechanics
  4

• HGEMM performance in Adreno(tm) 740 is not faster than SGEMM

  #513 opened a year ago by CunyangWei
  1

• Is it a good idea to use GCN cross lane instruction for optimization?

  #510 opened a year ago by fancyIX
  15

• CMake find package paths broken in MSYS2

  #502 opened 2 years ago by bandoti
  3

• [implement details] usm beheavior

  #501 opened 2 years ago by congzhangzh
  2

• Cuda execution failed，when running clblast_sample_sgemm_cuda, "CUDA NVRTC error: nvrtcCompileProgram: NVRTC_ERROR_INVALID_OPTION"

  #500 opened 2 years ago by luoyin500
  2

• [Question] How to Install on Windows?

  #499 opened 2 years ago by ilovefreesw
  2

• CL kernel preprocess cause compilation error

  #497 opened 2 years ago by handsome-fu
  2