This is an automated email from the ASF dual-hosted git repository.

github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git


    from 61f80814e6 [TVMScript] Fix PEP 563 closure variable resolution (#18856)
     add 173bb23274 [Runtime][Builtin] Handle mismatched type on argument #0 
when calling Builtin Runtime Operators (#18837)
     add 9ffffae896 fix: Complete ICHECK update across codebase (#18858)
     add 2bab8bc47d fix: Use T.decl_buffer instead of T.Buffer for aliased 
buffers in LongRoPE (#18859)
     add 63f099ad29 [CI][Docs] Update Sphinx dependencies (#18860)
     add e693452ad8 fix: Complete CHECK update across contrib runtime (#18861)

No new revisions were added by this update.

Summary of changes:
 3rdparty/tensorrt_llm/custom_allreduce_kernels.cu  |  8 +--
 apps/ios_rpc/tvmrpc/RPCArgs.mm                     |  2 +-
 apps/ios_rpc/tvmrpc/RPCServer.mm                   |  2 +-
 apps/ios_rpc/tvmrpc/TVMRuntime.mm                  |  2 +-
 docker/install/ubuntu_install_sphinx.sh            | 14 ++---
 docs/arch/pass_infra.rst                           |  6 +-
 docs/contribute/error_handling.rst                 |  4 +-
 docs/errors.rst                                    |  4 +-
 python/tvm/contrib/cutlass/attention_operation.py  | 14 ++---
 .../relax/frontend/nn/llm/position_embedding.py    |  8 +--
 src/arith/iter_affine_map.cc                       |  2 +-
 src/relax/transform/bind_params.cc                 |  2 +-
 src/runtime/contrib/coreml/coreml_runtime.mm       |  7 ++-
 src/runtime/contrib/cutlass/fp16_group_gemm.cuh    | 18 +++---
 .../cutlass/fp16_group_gemm_runner_sm100.cuh       | 14 ++---
 .../cutlass/fp16_group_gemm_runner_sm90.cuh        | 14 ++---
 src/runtime/contrib/cutlass/fp8_gemm.cu            | 19 +++---
 src/runtime/contrib/cutlass/fp8_group_gemm_sm90.cu | 18 +++---
 .../contrib/cutlass/fp8_groupwise_scaled_gemm.cuh  | 73 +++++++++++-----------
 .../fp8_groupwise_scaled_gemm_runner_sm100.cuh     | 12 ++--
 .../fp8_groupwise_scaled_gemm_runner_sm90.cuh      | 12 ++--
 ...p8_groupwise_scaled_group_gemm_runner_sm100.cuh | 12 ++--
 .../fp8_groupwise_scaled_group_gemm_sm100.cu       | 36 +++++------
 src/runtime/contrib/cutlass/gemm_runner.cuh        | 14 ++---
 src/runtime/contrib/nvshmem/kv_transfer.cu         | 71 ++++++++++++---------
 src/runtime/vm/builtin.cc                          | 36 +++++++++--
 src/target/llvm/codegen_hexagon.cc                 |  2 +-
 src/tir/transform/storage_rewrite.cc               |  2 +-
 src/tir/transform/vectorize_loop.cc                |  2 +-
 29 files changed, 235 insertions(+), 195 deletions(-)

Reply via email to