Google Git
Sign in
opensecura / 3p / openxla / iree / 7ed278e2d91b1a70187e5414ecb6e736444f9279 / runtime
  1. 31162eb Adding iree_status_to_string comment from review feedback. by Ben Vanik · 2 years, 4 months ago
  2. 34a5166 Exporting iree_status_to_string. (#11235) by Ben Vanik · 2 years, 4 months ago
  3. 2deecf3 Initial iree_loop_t implementation for web browsers. (#11204) by Scott Todd · 2 years, 4 months ago
  4. 4fd7a1d [docker] More code sharing and use development clang for Bazel builds (#11108) by Geoffrey Martin-Noble · 2 years, 4 months ago
  5. acb2688 Fix layering issues in Bazel build (#11166) by Geoffrey Martin-Noble · 2 years, 5 months ago
  6. 03190c6 initial `unpack` microkernel (#11157) by bjacob · 2 years, 5 months ago
  7. 691ab41 pack ukernel: optimized ARM64 code, benchmark,+refactorings (#11134) by bjacob · 2 years, 5 months ago
  8. a389b9a Some strictness to bring Clang, GCC and MSVC more in line with each other (#11140) by bjacob · 2 years, 5 months ago
  9. 0a6cdf0 Add support for GEMM e2e Test For CUDA backend on F16 input (#10842) by Manish Gupta · 2 years, 5 months ago
  10. c074522 Create Docker images for the latest software (#11083) by Geoffrey Martin-Noble · 2 years, 5 months ago
  11. 8ee5738 Prevent inlining of ukernel elementwise ops into shims. (#11091) by Ben Vanik · 2 years, 5 months ago
  12. c149d61 Adding vm::ref<T> support for non-iree_vm_ref_t ref types. (#11065) by Ben Vanik · 2 years, 5 months ago
  13. da03073 Changing default bytecode dispatch away from computed goto. (#11090) by Ben Vanik · 2 years, 5 months ago
  14. a29e67a Assorted fixes for tests on Windows. (#11079) by Scott Todd · 2 years, 5 months ago
  15. c9a6fa0 Add a test for the pack microkernel (#11041) by bjacob · 2 years, 5 months ago
  16. 05b7559 Fixing iree_hal_vulkan_driver_create_using_instance ownership. (#11066) by Ben Vanik · 2 years, 5 months ago
  17. e64babe Optimize code size of the mmt4d ukernel. (#11026) by bjacob · 2 years, 5 months ago
  18. e801a63 Minor ukernel cleanups (#11025) by bjacob · 2 years, 5 months ago
  19. 7aab552 Shorten iree_ukernel_ to iree_uk_ 🇬🇧 (#11024) by bjacob · 2 years, 5 months ago
  20. 0be43fc Introduce iree_ukernel_type_t to track data types in ukernels. (#11023) by bjacob · 2 years, 5 months ago
  21. b28b2c2 Fix several tests on Windows. (#11048) by Scott Todd · 2 years, 5 months ago
  22. fe3767d Adding flag for setting VMA preferredLargeHeapBlockSize. (#10983) by Ben Vanik · 2 years, 5 months ago
  23. beeb0e8 Fixing MSVC breakage from #10969. by Ben Vanik · 2 years, 5 months ago
  24. f7c9bc8 Tightening ukernel common definitions (#10950) by bjacob · 2 years, 5 months ago
  25. 36f6355 Microkernel for `linalg_ext.pack`, take 2 (#10969) by bjacob · 2 years, 5 months ago
  26. 86f1855 Fix GCC build error with IREE_MUST_USE_RESULT attribute placement (#10967) by bjacob · 2 years, 5 months ago
  27. ab4a82d Revert "microkernel for `linalg_ext.pack`" (#10951) by bjacob · 2 years, 5 months ago
  28. 3072817 Making ref_cc.h an implementation detail of iree/vm/ref.h. (#10948) by Ben Vanik · 2 years, 5 months ago
  29. 7a9b84c Making status_cc.h an implementation detail of status.h. (#10946) by Ben Vanik · 2 years, 5 months ago
  30. 9e3b638 microkernel for `linalg_ext.pack` (#10927) by bjacob · 2 years, 5 months ago
  31. decb765 Removing IREE_API_VERSION. (#10937) by Ben Vanik · 2 years, 5 months ago
  32. 7228438 Setting Vulkan pipeline names from entry point names. (#10935) by Ben Vanik · 2 years, 5 months ago
  33. d8ee3d7 Defaulting Vulkan dispatch workloads to graphics+compute queues. (#10926) by Ben Vanik · 2 years, 5 months ago
  34. 4e305e6 Adding AMD-specific frame delineation markers. (#10925) by Ben Vanik · 2 years, 5 months ago
  35. 401e209 Add requests for PyProject.toml (#10916) by powderluv · 2 years, 5 months ago
  36. 96ea431 Add requests to pip requirements for CUDA download (#10910) by powderluv · 2 years, 5 months ago
  37. 0ef1e5e Marking win32 vulkan symbols as optional. by Ben Vanik · 2 years, 5 months ago
  38. bfc32b6 Fixing platform-specific Vulkan symbol resolution. (#10904) by Ben Vanik · 2 years, 5 months ago
  39. 97c1ada Adding python API for device profiling. (#10899) by Ben Vanik · 2 years, 5 months ago
  40. fc631c1 Fix fp16 conversion from and to fp32 for Nan/Inf kind of values. (#10875) by Thomas · 2 years, 5 months ago
  41. 2d79c20 Adding support for the RenderDoc capture API for Vulkan devices. by Ben Vanik · 2 years, 5 months ago
  42. 7859d63 Updating the core tools to support --device_profiling_mode= flag. by Ben Vanik · 2 years, 5 months ago
  43. ec8dc02 Adding iree_hal_device_profiling_begin/end API. by Ben Vanik · 2 years, 5 months ago
  44. 85b1f82 Renaming util.do_not_optimize to util.optimization_barrier. (#10884) by Ben Vanik · 2 years, 5 months ago
  45. f75192a Adding CPU import conversion and ordinal assignment machinery. (#10865) by Ben Vanik · 2 years, 5 months ago
  46. bcf9da7 Fix missing dep with next stricter layering check (#10851) by Jacques Pienaar · 2 years, 5 months ago
  47. cbe9dcc Adding support for multiple queues on the local-task device. (#10817) by Ben Vanik · 2 years, 5 months ago
  48. f135fee Update for pending strictness change in layering check (#10828) by Jacques Pienaar · 2 years, 5 months ago
  49. bed9311 Handle large file sizes on Windows in file_io utils. (#10815) by Ben Vanik · 2 years, 6 months ago
  50. 8f39d27 Integrate llvm-project at b9898e7ed1ce and bump dependencies (#10740) by Thomas · 2 years, 6 months ago
  51. e9ae963 Implementing basic `--iree-execution-model=async-external` support. by Ben Vanik · 2 years, 6 months ago
  52. 7a27de5 Enabling CPU async execution by fixing task system cmd buffer retains. by Ben Vanik · 2 years, 6 months ago
  53. 888843c Adding iree_abort() wrapper. by Ben Vanik · 2 years, 6 months ago
  54. 9392122 Fix vm_util.c build on MSVC. (#10625) by Scott Todd · 2 years, 6 months ago
  55. 9a0058e Extending imports to have a context pointer. (#10580) by Ben Vanik · 2 years, 6 months ago
  56. 3c2d5d7 Enable disabling UBSAN in VMVX kernels (#10522) by Jacques Pienaar · 2 years, 6 months ago
  57. 5bf7ae9 Let vm_util be C-only, split C++ part into vm_util_cc (#10563) by bjacob · 2 years, 6 months ago
  58. a2c9f8c Fixing MSVC breakages and missing status checks in #10530. (#10559) by Ben Vanik · 2 years, 6 months ago
  59. 0bc1a79 Adding missing status checks omitted in #10530. by Ben Vanik · 2 years, 6 months ago
  60. b22698f Fixing MSVC breakage due to use of a C++20 feature. by Ben Vanik · 2 years, 6 months ago
  61. 8916027 Adds `IREE_HAL_EXECUTABLE_LOADER_DEPS` cmake var. (#10551) by Ben Vanik · 2 years, 6 months ago
  62. b1cc1a7 simple streaming kernel for the i8mm case (#10552) by bjacob · 2 years, 6 months ago
  63. 1842210 Touchups for mmt4d early return paths for trivial cases (#10550) by bjacob · 2 years, 6 months ago
  64. 1cb92dd Don't `#include <ostream>` outside of test-only code. (#10530) by bjacob · 2 years, 6 months ago
  65. c1e1d89 Making >1B wide VM buffer load/stores implicitly scale. (#10513) by Ben Vanik · 2 years, 6 months ago
  66. d3527a0 mmt4d ukernel: early return paths for degenerate cases (#10533) by bjacob · 2 years, 6 months ago
  67. f5b34b3 Silence unused-var warning-error when IREE_STATUS_MODE=0 (#10542) by bjacob · 2 years, 6 months ago
  68. e98a03b Adding `--expected_output=` flag to iree-run-module. (#10433) by Ben Vanik · 2 years, 6 months ago
  69. 688497c Accept EINTR (interrupted by signal) futex return code (#10517) by bjacob · 2 years, 6 months ago
  70. c427dff Adding IREE_HAL_EXECUTABLE_IMPORT_PROVIDER_DEFAULT_FN config. (#10507) by Ben Vanik · 2 years, 6 months ago
  71. eae331b Fix mmt4d_test: actually dont use unsupported instructions. (#10497) by bjacob · 2 years, 6 months ago
  72. f1fd8d1 Pass the cpu_data pointer to the ukernel, not just the value of field 0. (#10485) by bjacob · 2 years, 6 months ago
  73. 938c32b rename iree_ukernel_size_t to iree_ukernel_ssize_t (#10484) by bjacob · 2 years, 6 months ago
  74. 9b97d80 mmt4d i8mm kernel register fix (#10476) by bjacob · 2 years, 6 months ago
  75. e58f407 [vulkan] Create zero binding for dummy sets in BuiltinExecutables (#10448) by Lei Zhang · 2 years, 6 months ago
  76. 165949e Rationalize architecture checks in ukernels (#10475) by bjacob · 2 years, 6 months ago
  77. 4a00384 detect cpu features on Apple/ARM64 (#10469) by bjacob · 2 years, 6 months ago
  78. 77e7cac mmt4d builtin ukernels: baseline ARM64 kernels for f32 and i8 (#10440) by bjacob · 2 years, 7 months ago
  79. 250725b Disable ARM64 ukernels on APPLE for now (#10449) by bjacob · 2 years, 7 months ago
  80. 69afc49 fix mac arm64 build (#10444) by bjacob · 2 years, 7 months ago
  81. 3c332d0 Mmt4d builtin ukernel test/benchmark (#10389) by bjacob · 2 years, 7 months ago
  82. a96dd9e New mmt4d builtin with general code and some assembly kernels (i8mm, dotprod) (#10339) by bjacob · 2 years, 7 months ago
  83. 6fd42a4 fix `minimum_duration_ns` in `iree_benchmark_register` (#10424) by bjacob · 2 years, 7 months ago
  84. 2185211 Adding subspan offset/length to hal.buffer_view.create. (#10378) by Ben Vanik · 2 years, 7 months ago
  85. b86ef09 Add RISC-V 64-bit Linux unit tests (#10346) by CindyLiu · 2 years, 7 months ago
  86. 9ef8abf Hiding semaphores from the compiler. (#10374) by Ben Vanik · 2 years, 7 months ago
  87. 1abccdf Adding stream.async.load/store folders for bitcast ops. (#10373) by Ben Vanik · 2 years, 7 months ago
  88. 83b0103 NFC: Rename `DispatchDefaultWorkgroupCountOp` to `DispatchWorkgroupCountFromDagRootOp` (#10344) by MaheshRavishankar · 2 years, 7 months ago
  89. 27ded29 Remove tooling/cpu_features, switch to base/internal/cpu. (#10303) by bjacob · 2 years, 7 months ago
  90. 5c5f678 Reworking CPU data queries and access for vmvx/builtins/compiler-produced code. (#10270) by Ben Vanik · 2 years, 7 months ago
  91. 853dd6e Adding cpu_data.h and CPU data queries into iree/base/internal/cpu.h. by Ben Vanik · 2 years, 7 months ago
  92. 9336348 Initial mmt4d builtins - reference code move out of vmvx/module.c (#10272) by bjacob · 2 years, 7 months ago
  93. 4b4b854 Updating VMVX to use per-worker contexts and support workgroup state. by Ben Vanik · 2 years, 7 months ago
  94. 62ddd83 Fix mapping of buffers in mmt4d module functions. (#10269) by bjacob · 2 years, 7 months ago
  95. 2897621 Reverts 139619c355ce302b22d9bdd56d0fb764729dc6ef. by Ben Vanik · 2 years, 7 months ago
  96. 5914add Adding --task_worker_spin_us= flag to spin for a bit. by Ben Vanik · 2 years, 7 months ago
  97. ec36a08 Adding iree_processor_yield() for use in spin loops. by Ben Vanik · 2 years, 7 months ago
  98. d19400f Changing iree_futex_wait to use absolute deadlines. by Ben Vanik · 2 years, 7 months ago
  99. ff3359c Moving worker waking and submission outside of the executor lock. by Ben Vanik · 2 years, 7 months ago
  100. ca29c27 Allowing for task system work stealing to be disabled via tuning.h. by Ben Vanik · 2 years, 7 months ago