v0.52.0-rc22

Note

If you are installing from a release, please refer to the README, INSTALLATION instructions, and any other documentation packaged with the release, not on the main branch. There may be differences between the latest main and the previous release.

The changelog will now follow, showing the changes from last release.

This release was generated by the CI workflow https://github.com/tenstorrent/tt-metal/actions/runs/10841480463

📦 Uncategorized

• #5337: Fix Mixtral total number of generated tokens in perf benchmark
  • PR: #11994
• Split up n300 demo tests into functionality and performance
  • PR: #11969
• Updated sfpi compiler to address missing SFPNOP insertion
  • PR: #12061
• Move compute kernel config to TTNN
  • PR: #11801
• Add fold to resnet
  • PR: #11940
• [BugFix] Fixed tensor::is_allocated.
  • PR: #12071
• Revert "[BugFix] Fixed tensor::is_allocated."
  • PR: #12082
• #8598: sinh fix
  • PR: #12056
• #11646: Replace ttnn.experimental.tensor.* to ttnn.* in models/experimental, tests
  • PR: #11821
• #10754: Add data-parallel support for UNet Shallow on N300
  • PR: #12062
• #0: Fixed Conv2dConfig in broken tests
  • PR: #12064
• #0: Falcon40b T3K demo mismatch tokens fixed
  • PR: #12105
• #12069: Add catch and handling for device initialize exception, typic…
  • PR: #12070
• Point metal to new UMD main branch
  • PR: #12097
• Update CODEOWNERS
  • PR: #12112
• #11993: Fix offset calculation for uneven shard in reshard fast path
  • PR: #12083
• Update CODEOWNERS
  • PR: #12114
• #12117: Refactor DeviceMesh->MeshDevice, DeviceGrid->MeshShape
  • PR: #12118
• #11854: Move .umd that houses cluster descriptor to TT_METAL_HOME
  • PR: #12113
• Fused AllGather+Matmul
  • PR: #11760
• #12124: support moreh_nll_loss support large wight
  • PR: #12126
• [Bugfix] Fixed is allocated
  • PR: #12109
• #11990: Replace ttnn.experimental.tensor.* to ttnn.* in ttnn folder
  • PR: #12005
• #11132 Run Post-Commit Python Tests against wheels
  • PR: #11870
• #11109: refactoring moreh getitem forward
  • PR: #12134
• Revert "#11132 Run Post-Commit Python Tests against wheels "
  • PR: #12149
• #8865: Update reference dispatch times. Move tt_lib to ttnn.
  • PR: #11991
• #11809: migrate profiler into ttnn
  • PR: #11942
• Use new resnet test infra for GS
  • PR: #12072
• #12073: assert out on block and width sharded concat
  • PR: #12111
• [Bugfix] Fixed wrong capture counter
  • PR: #12169
• #11717: aligning jit build and collection of srcs used by multiple threads during risc compile
  • PR: #12080
• #11453: Add options to enable compilation with sanitizers
  • PR: #12181
• Revert "#11453: Add options to enable compilation with sanitizers"
  • PR: #12183
• #11453: Add options to enable compilation with sanitizers
  • PR: #12185
• #11623: Adding workaround for ND BH hang for MatmulMultiCoreMultiDRAMIn0MCastIn1MCast
  • PR: #12186
• Add UNet Shallow unit tests to post-commit test suite
  • PR: #12145
• #10808: use sharded concat in yolov4
  • PR: #12182
• #11858 , #11859: Fix Dockerfile 20.04 and 22.04 sequence and requirements
  • PR: #11861
• #11720: Enable ttnn binary to be built using rpath as origin
  • PR: #11721
• add myself to codeowners on docker scripts and installation guide
  • PR: #12202
• Add GH workflow telemetry into prepare metal action
  • PR: #12203
• Complex imag and imag_bw, angle_bw TTNN sweeps
  • PR: #12143
• #11989: Code clean up
  • PR: #12008
• Support arbitrary prefill lengths in Mixtral demo
  • PR: #12002
• #5337: Fix Mixtral unit tests
  • PR: #12213