1 Release published by 1 person

• v3.5.2

  published Jul 26, 2024

2 Pull requests merged by 2 people

• github: workflows: bump ossf/scorecard-action from 2.3.3 to 2.4.0

  #2012 merged Jul 29, 2024

• rfcs: support group normalization in graph api

  #1769 merged Jul 29, 2024

8 Pull requests opened by 4 people

• gpu: sycl: refactor where kernel arguments are constructed

  #2004 opened Jul 23, 2024

• gpu: sycl: refactor post ops to use a union of post op types

  #2005 opened Jul 23, 2024

• gpu: sycl: Ensure dimensions are <= INT32_MAX

  #2006 opened Jul 23, 2024

• cpu: aarch64: Enable BRGEMM Depthwise Forward Convolution.

  #2009 opened Jul 26, 2024

• gpu: sycl: introduce memory objects

  #2010 opened Jul 26, 2024

• gpu: sycl: reorder: bugfix - more checks

  #2013 opened Jul 29, 2024

• Aarch64 jit depthwise convolution kernels

  #2014 opened Jul 29, 2024

• gpu: sycl: lnorm: bugfix scales datatypes

  #2016 opened Jul 29, 2024

3 Issues closed by 2 people

• could not create a primitive descriptor for a reorder primitive

  #1996 closed Jul 23, 2024

• DNNL_TEST_SET NIGHTLY produces duplicate benchdnn tests

  #1993 closed Jul 23, 2024

• python call sycl C++ Error

  #2002 closed Jul 23, 2024

4 Issues opened by 3 people

• Pytorch bias quantisation on aarch64 - how best to proceed?

  #2015 opened Jul 29, 2024

• The testcase test_graph_unit_dnnl_layer_norm_usm_cpu fails

  #2011 opened Jul 27, 2024

• brg:sve_256 fails benchdnn accuracy tests

  #2008 opened Jul 24, 2024

• brgconv:sve_256 uses a lot of memory

  #2007 opened Jul 24, 2024

12 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• gpu: sycl: softmax: Adding post-ops

  #2003 commented on Jul 26, 2024 • 11 new comments

• gpu: sycl: prelu: Fix backward pass

  #1998 commented on Jul 26, 2024 • 2 new comments

• Add support for different src and dst datatypes in the SYCL implementation for pooling

  #1878 commented on Jul 26, 2024 • 1 new comment

• Speed drop when running oneDNN in a subthread

  #1997 commented on Jul 24, 2024 • 0 new comments

• Working with the Reorder

  #1980 commented on Jul 25, 2024 • 0 new comments

• gpu: sycl: sum: implemented

  #1940 commented on Jul 26, 2024 • 0 new comments

• gpu: nvidia: Add support for cublaslt matmul

  #1972 commented on Jul 29, 2024 • 0 new comments

• gpu: sycl: resampling: Enabling different src/dst datatypes

  #1975 commented on Jul 26, 2024 • 0 new comments

• gpu: sycl: convolution: implemented

  #1989 commented on Jul 29, 2024 • 0 new comments

• gpu: sycl: binary: Adding datatype support with int32

  #1995 commented on Jul 26, 2024 • 0 new comments

• rfcs: add proposal for collecting and analyzing synthetic data

  #1999 commented on Jul 29, 2024 • 0 new comments

• Fixed parser of OpenCL 3.0 for NVIDIA GPU

  #2001 commented on Jul 29, 2024 • 0 new comments