Skip to content

issues Search Results · repo:uxlfoundation/oneDNN language:C++

Filter by

1k results
 (100 ms)

1k results

inuxlfoundation/oneDNN (press backspace or delete to remove)

After merging #3156 our automated conda build system started showing build issues where __builtin_vsx_lxvp, __builtin_vsx_disassemble_pair and similar instructions are directly used. Some code containing ...
bug
platform:cpu-ppc64
  • spalicki
  • 3
  • Opened 
    2 days ago
  • #3749

Summary We managed to package up the open-source compiler hosted on intel/llvm in nix under https://github.com/blenderfreaky/intel-nix Version commit e5542695163c36a09daa5aa77ce3ba86d21f29b4 Environment ...
sighting
  • izelnakri
  • 3
  • Opened 
    12 days ago
  • #3711

Summary Compile error on windows utils.cpp(63,20) - GetEnvironmentVariable should be GetEnvironmentVariableA utils.cpp(130,18) - GetFileAttributes should be GetFileAttributesA Version GIT TAG v3.9-rc ...
sighting
  • sheldonrobinson
  • Opened 
    14 days ago
  • #3698

Summary Hi, I cloned the main branch and compiled it, then I tried to validate it using ctest, but one of the testcases faild: img width= 807 height= 46 alt= Image src= https://github.com/user-attachments/assets/1433a331-963f-443e-b336-6086e208c05f ...
bug
platform:cpu-aarch64
  • ilotasw
  • 2
  • Opened 
    17 days ago
  • #3676

Summary Hi, I was wondering if there is any any plan to support dynamic batch sizse in LSTM primitives. Problem statement Describe the problem you are trying to solve with reasonable level of details. ...
enhancement
  • ilotasw
  • 2
  • Opened 
    24 days ago
  • #3645

oneDNN enables 4bits for matmul, like u4/s4 and float-4bit. However, there is no document or example about 4-bit data type. Especially, 4bit storage structure and use cases in the real inference. Could ...
platform:cpu-x64
question
  • alanzhai219
  • 10
  • Opened 
    24 days ago
  • #3641

Summary When using SYCL (Intel GPU) to perform matrix multiplication (matmul), the primitive descriptor cannot be created, which triggers an error by llama.cpp+IPEX-LLM. Version - Kernel version: ...
platform:gpu-intel
question
  • zcwang
  • 13
  • Opened 
    27 days ago
  • #3633

Summary In https://github.com/uxlfoundation/oneDNN/blob/main/src/common/type_helpers.hpp#L332. Blocking desc of nchw and nhwc will be equal for some cases. Since blocking_desc_is_equal used in memory_desc_matches_tag, ...
sighting
  • ZlobinGM
  • 1
  • Opened 
    28 days ago
  • #3618

Summary It looks like a bug has been introduced in https://github.com/uxlfoundation/oneDNN/pull/3396, specifically commit 605195441db79d84337c766862c904b285352c96 which is causing the nightly aarch64 ...
bug
  • Ryo-not-rio
  • 7
  • Opened 
    29 days ago
  • #3613

Summary Building oneDNN, and benchdnn in particular, is impossible with SYCL enabled Version tried with v3.8, but problem is universal, as long as you have defined STRINGIFY symbol in src/common/z_magic.hpp ...
not a bug
  • mateuszpn
  • 4
  • Opened 
    on Jul 10
  • #3571
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue search results · GitHub