Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / oneapi-src/oneDNN issues and pull requests

#2341 - Introduce matmul-based IP for inference

Pull Request - State: closed - Opened by densamoilov about 1 month ago - 2 comments
Labels: platform:cpu-x64

#2340 - github: codeowners: add graph team to review graph doc

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 3 comments
Labels: devops

#2339 - xe: enable source code information for OpenCL kernels

Pull Request - State: closed - Opened by rjoursler about 1 month ago - 2 comments
Labels: platform:gpu-intel, component:build

#2338 - Enabled clang-tidy checks in header files

Pull Request - State: closed - Opened by vpirogov about 1 month ago - 9 comments
Labels: documentation, platform:cpu-x64, platform:cpu-aarch64, platform:gpu-intel, governance, component:api, component:tests, component:build

#2337 - [WIP] src: cpu: x64: jit_brdgmm_dw_conw: make brdgmm support fusing sum op

Pull Request - State: open - Opened by amakarev about 1 month ago - 2 comments
Labels: platform:cpu-x64

#2337 - [WIP] src: cpu: x64: jit_brdgmm_dw_conw: make brdgmm support fusing sum op

Pull Request - State: open - Opened by amakarev about 1 month ago - 2 comments
Labels: platform:cpu-x64

#2336 - copyright: update copyright year 2025

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 2 comments
Labels: component:graph-api, component:tests, component:build

#2336 - copyright: update copyright year 2025

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 2 comments
Labels: component:graph-api, component:tests, component:build

#2335 - benchdnn: introduce --bia-dt option for drivers supporting bias (fixed MFDNN-3031, fixed MFDNN-12936)

Pull Request - State: closed - Opened by dzarukin about 1 month ago - 6 comments
Labels: documentation, platform:cpu-x64, component:graph-api, component:tests

#2334 - build: enabled __cplusplus preprocessor macro for MSVC compiler

Pull Request - State: closed - Opened by vpirogov about 1 month ago
Labels: component:build

#2334 - build: enabled __cplusplus preprocessor macro for MSVC compiler

Pull Request - State: closed - Opened by vpirogov about 1 month ago
Labels: component:build

#2333 - benchdnn: binary: update filling (maybe fixed MFDNN-12857)

Pull Request - State: closed - Opened by dzarukin about 1 month ago - 1 comment
Labels: component:tests

#2332 - [rls-v3.6 backport] cpu: x64: matmul: copy_routines: associate fp16 support with dt, not only isa

Pull Request - State: closed - Opened by amakarev about 1 month ago - 1 comment
Labels: platform:cpu-x64, backport

#2332 - [rls-v3.6 backport] cpu: x64: matmul: copy_routines: associate fp16 support with dt, not only isa

Pull Request - State: closed - Opened by amakarev about 1 month ago - 1 comment
Labels: platform:cpu-x64, backport

#2331 - benchdnn: graph: extend --dt to support specifying tensor id

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 3 comments
Labels: documentation, component:graph-api, component:tests

#2330 - graph, benchdnn: support implicit causal mask in graph API

Pull Request - State: closed - Opened by ElaineBao about 1 month ago - 6 comments
Labels: documentation, component:api, component:graph-api, component:tests

#2329 - Updated spdlog version v1.15.0

Pull Request - State: closed - Opened by vpirogov about 1 month ago - 2 comments
Labels: documentation, component:build

#2329 - Updated spdlog version v1.15.0

Pull Request - State: closed - Opened by vpirogov about 1 month ago - 2 comments
Labels: documentation, component:build

#2328 - [WiP] common: promote sparse functionality

Pull Request - State: open - Opened by spalicki about 1 month ago
Labels: documentation, component:api, component:tests, component:build, component:examples

#2327 - build: cmake: remove old cmake policies for examples and tests

Pull Request - State: closed - Opened by spalicki about 1 month ago
Labels: component:tests, component:examples

#2326 - build: switched from FindPythonInterp to FindPython2

Pull Request - State: closed - Opened by vpirogov about 1 month ago
Labels: component:build

#2325 - github: workflows: add AArch64 nightly runs

Pull Request - State: closed - Opened by theComputeKid about 1 month ago - 1 comment
Labels: devops

#2324 - [ACL] Stateless feature impacts winograd convolution performance

Issue - State: closed - Opened by alvoron about 1 month ago - 5 comments
Labels: sighting, platform:cpu-aarch64

#2323 - Fix for issue with generated CMake config file

Pull Request - State: closed - Opened by vpirogov about 1 month ago - 1 comment
Labels: component:build

#2322 - xe: sdpa: Improve performance of quantization with better alignment and prefetching

Pull Request - State: closed - Opened by umar456 about 1 month ago - 2 comments
Labels: performance, platform:gpu-intel

#2320 - graph: backend: passes: verbose log enhancement

Pull Request - State: closed - Opened by rongzha1 about 1 month ago - 5 comments
Labels: component:graph-api

#2319 - brgemm: support arbitrary K on AMX

Pull Request - State: closed - Opened by ankalinin about 1 month ago - 10 comments
Labels: platform:cpu-x64, platform:cpu-aarch64, component:tests

#2318 - xe: sdpa: check attn mask presence

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 1 comment
Labels: platform:gpu-intel

#2317 - graph: sdpa: Fix k/v only pattern to allow for optional mask/scale

Pull Request - State: closed - Opened by umar456 about 1 month ago - 2 comments
Labels: component:graph-api

#2316 - graph: backend: dnnl: enhance fp8 matmul patterns

Pull Request - State: closed - Opened by wzt1997 about 1 month ago - 1 comment
Labels: component:graph-api, component:tests

#2315 - xe: increase concat lws[0] size w/new optimal calculation

Pull Request - State: closed - Opened by syurkevi about 1 month ago - 1 comment
Labels: platform:gpu-intel

#2314 - benchdnn: graph: check partition number by default

Pull Request - State: closed - Opened by TaoLv about 1 month ago - 2 comments
Labels: documentation, component:graph-api, component:tests

#2313 - graph: utils: pm: fix fill_local_in_map

Pull Request - State: closed - Opened by ElaineBao about 1 month ago - 2 comments
Labels: component:graph-api, component:tests

#2312 - graph: backend: dnnl: encode mem address into constant cache key

Pull Request - State: closed - Opened by xiang1guo about 1 month ago - 4 comments
Labels: component:graph-api, component:tests

#2311 - xe: conv: apply skip-w optimization for large strides

Pull Request - State: closed - Opened by echeresh about 2 months ago - 3 comments
Labels: platform:gpu-intel, backport

#2310 - github: workflows: bump lukka/get-cmake from 3.31.2 to 3.31.3

Pull Request - State: closed - Opened by dependabot[bot] about 2 months ago - 2 comments
Labels: devops

#2309 - Backport: cpu: x64: fixes matmul segfault

Pull Request - State: closed - Opened by xuxinzen about 2 months ago - 2 comments
Labels: bug, platform:cpu-x64, backport

#2308 - src: cpu: aarch64: Re-enable matmul static quantisation through ACL.

Pull Request - State: closed - Opened by renato-arantes about 2 months ago - 6 comments
Labels: platform:cpu-aarch64

#2307 - [GPU] Fix jit GEMM debug strategy submission

Pull Request - State: closed - Opened by kealan-barbieri about 2 months ago
Labels: bug, platform:gpu-intel

#2306 - xe: conv: apply skip-w optimization for large strides

Pull Request - State: closed - Opened by echeresh about 2 months ago - 4 comments
Labels: platform:gpu-intel

#2305 - fixups for CI workflows

Pull Request - State: closed - Opened by vpirogov about 2 months ago - 1 comment
Labels: devops

#2304 - Fix sdpa descriptor hashing function

Pull Request - State: closed - Opened by umar456 about 2 months ago - 2 comments
Labels: bug

#2303 - test_benchdnn_modeC_conv_ci_cpu fails on AArch64 CI for c7g instance

Issue - State: open - Opened by renato-arantes about 2 months ago - 5 comments
Labels: bug, platform:cpu-aarch64

#2302 - graph: utils: pm: verify multi-consumers number

Pull Request - State: closed - Opened by ElaineBao about 2 months ago - 2 comments
Labels: component:graph-api, component:tests

#2301 - doc: graph: add document for sdpa with compressed key and value

Pull Request - State: closed - Opened by wzt1997 about 2 months ago - 5 comments
Labels: documentation, component:graph-api

#2300 - src: handle DNNL_VERBOSE=OFF support through an early exit (fixes MFDNN-12862)

Pull Request - State: closed - Opened by dzarukin about 2 months ago - 1 comment
Labels: platform:cpu-x64, platform:gpu-intel, component:graph-api

#2299 - src: gpu: intel: jit: gemm: improve second-token gemm performance on xe2

Pull Request - State: closed - Opened by hidefromkgb about 2 months ago - 2 comments
Labels: platform:gpu-intel

#2298 - common: workaround for GCC 8 format-truncation warning

Pull Request - State: closed - Opened by spalicki about 2 months ago - 1 comment

#2297 - xe: jit: enable kernel debug information

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 2 comments
Labels: platform:gpu-intel

#2296 - build: cmake: unsupported compilers cleanup

Pull Request - State: closed - Opened by spalicki about 2 months ago - 9 comments
Labels: documentation, platform:cpu-x64, platform:gpu-intel, component:graph-api, component:build

#2295 - tests: benchdnn: add experimental support for sycl graph

Pull Request - State: closed - Opened by spalicki about 2 months ago - 5 comments
Labels: documentation, component:tests, component:build

#2294 - AArch64 compilers support

Issue - State: closed - Opened by spalicki about 2 months ago - 3 comments
Labels: question, platform:cpu-aarch64

#2293 - Revert PR for static quantisation for both matmul and convolution

Pull Request - State: closed - Opened by renato-arantes about 2 months ago
Labels: platform:cpu-aarch64

#2292 - x64: brdgmm conv: enable zps per group (backport)

Pull Request - State: open - Opened by tczeszun about 2 months ago
Labels: platform:cpu-x64, backport

#2291 - graph: backend: dnnl: support permute for scale and zps inputs

Pull Request - State: closed - Opened by wzt1997 about 2 months ago - 6 comments
Labels: component:graph-api, component:tests

#2290 - [GPU] Conv and Gemm fix

Pull Request - State: closed - Opened by dyoussif about 2 months ago - 4 comments
Labels: platform:gpu-intel

#2289 - xe: ocl: gemm: fix gemm_with_post_ops accumulator type

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 1 comment
Labels: platform:gpu-intel

#2288 - xe: conv: fix tuning support

Pull Request - State: closed - Opened by echeresh about 2 months ago - 2 comments
Labels: platform:gpu-intel

#2287 - gpu: generic: sycl: fix typos in VDISPATCH_SUM calls

Pull Request - State: closed - Opened by spalicki about 2 months ago - 3 comments
Labels: platform:gpu-generic

#2286 - github: automation: fixed rules for commit scope check

Pull Request - State: closed - Opened by vpirogov about 2 months ago - 2 comments
Labels: devops

#2285 - Fix CI failing tests for matmul and convolution

Pull Request - State: closed - Opened by renato-arantes about 2 months ago - 3 comments
Labels: platform:cpu-aarch64

#2284 - [backport v3.7] SDPA fixes for regular scale tensor and alignment for K tensor

Pull Request - State: closed - Opened by umar456 about 2 months ago - 1 comment
Labels: platform:gpu-intel, backport

#2283 - xe: sdpa: Fix KQ gemm alignment for the K tensor

Pull Request - State: closed - Opened by umar456 about 2 months ago - 3 comments
Labels: bug, platform:gpu-intel

#2282 - [rls-v3.6 backport] cpu: x64: matmul/brgemm: enable f16 dst dt for int8 matmul

Pull Request - State: closed - Opened by amakarev about 2 months ago
Labels: platform:cpu-x64, backport

#2281 - Restricted x64 CI to relevant changes

Pull Request - State: closed - Opened by vpirogov about 2 months ago
Labels: devops

#2280 - rfcs: graph api: Constant Block Weight Mechanism Graph API design

Pull Request - State: open - Opened by xiang1guo about 2 months ago
Labels: RFC

#2279 - Add cpu reference and testing for fp4_e3m0 data type

Pull Request - State: closed - Opened by mgouicem about 2 months ago - 1 comment
Labels: documentation, component:api, component:tests

#2278 - doc: graph: document for complex fusions

Pull Request - State: closed - Opened by TaoLv about 2 months ago - 3 comments
Labels: documentation

#2277 - x64: matmul: workaround large buffer reorder failure

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 1 comment
Labels: platform:cpu-x64

#2276 - xe, benchdnn, x64: Enable large buffer GEMM and Convolution testing for GPU

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 3 comments
Labels: platform:gpu-intel, component:tests

#2275 - xe: sdpa: Allow non-transposed scalars for the KQ matmul

Pull Request - State: closed - Opened by umar456 about 2 months ago
Labels: platform:gpu-intel

#2274 - x64: brdgmm conv: enable zps per group (closes MFDNN-12556)

Pull Request - State: closed - Opened by tczeszun about 2 months ago
Labels: platform:cpu-x64, component:tests

#2273 - graph: backend: kernels: sdp verbose enhancement

Pull Request - State: closed - Opened by rongzha1 about 2 months ago - 12 comments
Labels: component:graph-api

#2272 - cpu: x64: brgemm, matmul: add f32:f16 configuration support on AVX2 and AVX512_CORE (fixes MFDNN-11992)

Pull Request - State: closed - Opened by dzarukin about 2 months ago - 7 comments
Labels: platform:cpu-x64, component:tests

#2271 - cpu: softmax: enable relaxed acc_mode (fixes MFDNN-12822)

Pull Request - State: closed - Opened by dzarukin about 2 months ago - 2 comments
Labels: documentation, platform:cpu-x64, component:tests

#2270 - src: attr: quantization refactor (part 1)

Pull Request - State: closed - Opened by dzarukin about 2 months ago - 9 comments
Labels: platform:cpu-x64, platform:cpu-aarch64, platform:gpu-nvidia, platform:gpu-intel, platform:gpu-generic, component:tests

#2269 - s8 quantization and bias shows unexpected behavior

Issue - State: closed - Opened by lucasreis1 about 2 months ago - 5 comments
Labels: question

#2268 - github: workflows: added labels for common code changes

Pull Request - State: closed - Opened by vpirogov about 2 months ago
Labels: devops

#2267 - src: gpu: intel: jit: conv: add reorder-based precomputed zero points

Pull Request - State: closed - Opened by hidefromkgb about 2 months ago - 4 comments
Labels: platform:cpu-x64, platform:gpu-intel, platform:gpu-generic

#2266 - [GPU] Fp8 compute backports

Pull Request - State: closed - Opened by kealan-barbieri about 2 months ago - 2 comments
Labels: platform:gpu-intel, component:tests

#2265 - xe: jit: gemm: remove unsupported GPU architecture strategies

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 1 comment
Labels: platform:gpu-intel

#2264 - cpu: aarch64: thread_local ThreadpoolScheduler setting

Pull Request - State: closed - Opened by davsva01 about 2 months ago - 3 comments
Labels: platform:cpu-aarch64

#2263 - github: workflows: Streamline AArch64 CI

Pull Request - State: closed - Opened by theComputeKid about 2 months ago - 12 comments
Labels: devops

#2262 - benchdnn: graph: use plain copy and paste for reorder if fail to create pd

Pull Request - State: closed - Opened by wzt1997 about 2 months ago - 7 comments
Labels: component:graph-api, component:tests

#2261 - graph: backend: dnnl: fuse dst scale/zps for layout propagation in sdpa primitive kernel

Pull Request - State: closed - Opened by wzt1997 about 2 months ago - 1 comment

#2260 - xe: conv_v2: introduce Stream-K support

Pull Request - State: closed - Opened by echeresh about 2 months ago - 2 comments
Labels: platform:gpu-intel

#2259 - Restricted AArch64 CI to relevant changes

Pull Request - State: closed - Opened by vpirogov about 2 months ago - 10 comments
Labels: devops

#2258 - xe: jit: ngen: add symbol table

Pull Request - State: closed - Opened by rjoursler about 2 months ago - 1 comment
Labels: platform:gpu-intel

#2257 - xe: jit: use F pipe for zero initialization

Pull Request - State: closed - Opened by echeresh about 2 months ago - 3 comments
Labels: platform:gpu-intel

#2256 - xe: jit: gemm: set lateScale lookahead to match load size

Pull Request - State: closed - Opened by kealan-barbieri about 2 months ago - 1 comment
Labels: bug, platform:gpu-intel, backport

#2255 - xe: jit: gemm: fix ARL m = 1 int4 cwei regression

Pull Request - State: closed - Opened by petercad about 2 months ago - 1 comment
Labels: platform:gpu-intel, backport

#2254 - xe: jit: gemm: fix ARL m = 1 int4 cwei regression

Pull Request - State: closed - Opened by petercad about 2 months ago
Labels: platform:gpu-intel

#2253 - cpu: x64: matmul/brgemm: enable f16 dst dt for int8 matmul

Pull Request - State: closed - Opened by amakarev about 2 months ago - 1 comment
Labels: platform:cpu-x64

#2252 - graph, example, benchdnn: add int4 gated mlp

Pull Request - State: closed - Opened by TaoLv about 2 months ago - 3 comments
Labels: component:graph-api, component:tests, component:examples

#2251 - [rls-v3.7-pc] xe: jit: gemm: set lateScale lookahead to match load size

Pull Request - State: closed - Opened by atkassen about 2 months ago - 2 comments
Labels: platform:gpu-intel, backport

#2250 - generic: sycl: Introduce spec constants for sycl matmul

Pull Request - State: open - Opened by ShanoToni about 2 months ago - 1 comment
Labels: platform:gpu-generic, component:tests

#2249 - gpu: generic: add simple SYCL reduction implementation

Pull Request - State: closed - Opened by sgeor255 about 2 months ago
Labels: platform:gpu-generic

#2248 - generic:sycl: Inner Product FWD

Pull Request - State: closed - Opened by sgeor255 about 2 months ago
Labels: documentation, platform:gpu-nvidia, platform:gpu-generic, component:tests

#2247 - [GPU] Shapeless convolution: groundwork for Stream-K support

Pull Request - State: closed - Opened by echeresh about 2 months ago - 4 comments
Labels: platform:gpu-intel