Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / pytorch-labs/tritonbench issues and pull requests
#150 - [run_utils] Add output dir
Pull Request -
State: open - Opened by xuzhao9 7 days ago
Labels: cla signed
#149 - Config cache
Pull Request -
State: closed - Opened by Darkviper7 7 days ago
- 1 comment
#148 - [compile_trace] Add compile time Kineto trace
Pull Request -
State: open - Opened by xuzhao9 8 days ago
- 4 comments
Labels: cla signed
#147 - [compile_time] Use py-spy to peek into the Triton compilation process
Issue -
State: closed - Opened by xuzhao9 8 days ago
- 1 comment
#146 - Add nvidia states to the json
Pull Request -
State: open - Opened by xuzhao9 10 days ago
- 1 comment
Labels: cla signed
#145 - Add variance to the latency metric
Pull Request -
State: closed - Opened by xuzhao9 13 days ago
- 2 comments
Labels: cla signed, Merged
#144 - [ci] Collect more machine status in the CI json
Issue -
State: open - Opened by xuzhao9 13 days ago
#143 - [ci] More aggressively tune the GPU and benchmark
Pull Request -
State: closed - Opened by xuzhao9 13 days ago
- 4 comments
Labels: cla signed, Merged
#142 - [tritonbench] Change default precision for addmm to fp16
Pull Request -
State: closed - Opened by SamGinzburg 13 days ago
- 2 comments
Labels: cla signed, Merged
#141 - [flash_attention] flash_v3 performance is slower than cudnn and triton_tutorial_flash_v2 on certain input and H100
Issue -
State: open - Opened by xuzhao9 15 days ago
- 2 comments
#140 - Update FBGEMM version
Pull Request -
State: closed - Opened by xuzhao9 15 days ago
- 2 comments
Labels: cla signed, Merged
#139 - Allow n_heads specification in fp8 attention bench
Pull Request -
State: closed - Opened by mandroid6 15 days ago
- 4 comments
Labels: cla signed, fb-exported, Merged
#138 - [proton] co-designing proton and tritonbench for a better user experience
Issue -
State: open - Opened by fywkevin 15 days ago
- 4 comments
#137 - [nightly] Run the workflow from non-pull-request
Pull Request -
State: closed - Opened by xuzhao9 15 days ago
- 2 comments
Labels: cla signed, Merged
#136 - Add script to upload to scribe
Pull Request -
State: closed - Opened by xuzhao9 16 days ago
- 2 comments
Labels: cla signed, Merged
#135 - TMA benchmark for fp8 attention
Pull Request -
State: closed - Opened by mandroid6 16 days ago
- 3 comments
Labels: cla signed, fb-exported, Merged
#134 - add support for causal arg in fp8 attention
Pull Request -
State: closed - Opened by mandroid6 16 days ago
- 4 comments
Labels: cla signed, fb-exported, Merged
#133 - [nightly] Deploy nightly workflow
Pull Request -
State: closed - Opened by xuzhao9 16 days ago
- 2 comments
Labels: cla signed, Merged
#132 - [install] Use build constraints to limit the package version
Pull Request -
State: closed - Opened by xuzhao9 20 days ago
- 2 comments
Labels: cla signed, Merged
#131 - Adding support for seq_length in fp8 attention bench
Pull Request -
State: closed - Opened by mandroid6 20 days ago
- 2 comments
Labels: cla signed, fb-exported, Merged
#130 - Fix the docker build
Pull Request -
State: closed - Opened by xuzhao9 21 days ago
- 2 comments
Labels: cla signed, Merged
#129 - fbgemm_gpu_experimental_gen_ai_py.so missing when running fbgemm on AMDGPU
Issue -
State: closed - Opened by htyu 21 days ago
- 11 comments
#128 - Fix installation for amd gpu
Pull Request -
State: closed - Opened by FindHao 27 days ago
- 2 comments
Labels: cla signed, Merged
#127 - Error for installation
Issue -
State: closed - Opened by FindHao 27 days ago
#126 - Support running multiple modes with one command
Issue -
State: closed - Opened by xuzhao9 28 days ago
#125 - Add reset dynamo option
Pull Request -
State: closed - Opened by FindHao about 1 month ago
- 2 comments
Labels: cla signed, Merged
#124 - [CI] Fix the CI failure by skipping xformers
Pull Request -
State: open - Opened by xuzhao9 about 1 month ago
- 2 comments
Labels: cla signed
#123 - Update FBGEMM to 921e3051c0b2b46b81e61104b498c388bc718841
Pull Request -
State: closed - Opened by htyu about 1 month ago
- 2 comments
Labels: cla signed, Merged
#122 - Run warp-specialized FP8 rowsise with --warp_specialization
Pull Request -
State: closed - Opened by htyu about 1 month ago
- 5 comments
Labels: cla signed, fb-exported, Merged
#121 - Need to audit baseline benchmarks
Issue -
State: open - Opened by adamomainz about 2 months ago
- 1 comment
#120 - [docker] Fix the nightly docker build
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 3 comments
Labels: cla signed, Merged
#119 - [metrics] Fix compile_time
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 2 comments
Labels: cla signed, Merged
#118 - [op_collection] Force `--isolate` mode when running multiple ops
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 2 comments
Labels: cla signed, Merged
#117 - Need break-down of compilation time for Triton
Issue -
State: open - Opened by xuzhao9 about 2 months ago
#116 - Use nvtx.range_start
Pull Request -
State: closed - Opened by FindHao about 2 months ago
- 3 comments
Labels: cla signed, Merged
#115 - [UX] Add `--list-metrics` option to list all available metrics
Issue -
State: open - Opened by xuzhao9 about 2 months ago
- 2 comments
#115 - [UX] Add `--list-metrics` option to list all available metrics
Issue -
State: open - Opened by xuzhao9 about 2 months ago
- 2 comments
#114 - [metrics][ncu_rep] Trace entire process on backward pass.
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 3 comments
Labels: cla signed
#114 - [metrics][ncu_rep] Trace entire process on backward pass.
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 3 comments
Labels: cla signed
#113 - Fix donated_buffer issue
Pull Request -
State: closed - Opened by FindHao about 2 months ago
- 2 comments
Labels: cla signed, Merged
#113 - Fix donated_buffer issue
Pull Request -
State: closed - Opened by FindHao about 2 months ago
- 2 comments
Labels: cla signed, Merged
#112 - Fix cudagraph mem
Pull Request -
State: closed - Opened by FindHao about 2 months ago
- 4 comments
Labels: cla signed, Merged
#112 - Fix cudagraph mem
Pull Request -
State: closed - Opened by FindHao about 2 months ago
- 4 comments
Labels: cla signed, Merged
#111 - Support flops metric in proton profiling
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 7 comments
Labels: cla signed, fb-exported, Merged
#111 - Support flops metric in proton profiling
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 7 comments
Labels: cla signed, fb-exported, Merged
#110 - Cudagraph doesn't work anymore
Issue -
State: closed - Opened by FindHao about 2 months ago
- 3 comments
#110 - Cudagraph doesn't work anymore
Issue -
State: closed - Opened by FindHao about 2 months ago
- 3 comments
#109 - remove oss
Pull Request -
State: open - Opened by LinjianMa about 2 months ago
- 1 comment
Labels: cla signed, fb-exported
#109 - remove oss
Pull Request -
State: open - Opened by LinjianMa about 2 months ago
- 1 comment
Labels: cla signed, fb-exported
#108 - Move reduction_gemm
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 2 comments
Labels: cla signed, fb-exported, Merged
#108 - Move reduction_gemm
Pull Request -
State: closed - Opened by xuzhao9 about 2 months ago
- 2 comments
Labels: cla signed, fb-exported, Merged
#107 - Long run issue
Issue -
State: closed - Opened by FindHao about 2 months ago
- 4 comments
#107 - Long run issue
Issue -
State: closed - Opened by FindHao about 2 months ago
- 4 comments
#106 - [metrics] Enable cudagraph mode for kineto_trace
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#105 - [flash_attention] Add pt2_sdpa
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#105 - [flash_attention] Add pt2_sdpa
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#104 - Disable donated_buffer for all ops's backward benchmarking
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 6 comments
Labels: cla signed, Merged
#104 - Disable donated_buffer for all ops's backward benchmarking
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 6 comments
Labels: cla signed, Merged
#103 - `--op-collection` misses `layernorm`
Issue -
State: open - Opened by FindHao 2 months ago
- 1 comment
#102 - [metrics] Add proton profiling
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 3 comments
Labels: cla signed, Merged
#101 - `--output` is not compatible with `--op-collection`
Issue -
State: open - Opened by FindHao 2 months ago
- 2 comments
#101 - `--output` is not compatible with `--op-collection`
Issue -
State: open - Opened by FindHao 2 months ago
- 2 comments
#100 - Specify timeunit for nsys report
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#100 - Specify timeunit for nsys report
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#99 - [decoding_attention] Fix broken flash_attention and xformers
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 5 comments
Labels: cla signed, Merged
#99 - [decoding_attention] Fix broken flash_attention and xformers
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 5 comments
Labels: cla signed, Merged
#98 - Fix embedding accuracy check
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#98 - Fix embedding accuracy check
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#97 - [flash_attention] Bug fix for option `--native-sdpa`
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#97 - [flash_attention] Bug fix for option `--native-sdpa`
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#96 - Enable more kernels in CI after pytorch triton pin update
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 4 comments
Labels: cla signed, Merged
#96 - Enable more kernels in CI after pytorch triton pin update
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 4 comments
Labels: cla signed, Merged
#95 - Add torch_compile_debug to .gitignore
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#95 - Add torch_compile_debug to .gitignore
Pull Request -
State: closed - Opened by FindHao 2 months ago
- 2 comments
Labels: cla signed, Merged
#94 - Naming issue about profile reports
Issue -
State: open - Opened by FindHao 2 months ago
- 1 comment
#94 - Naming issue about profile reports
Issue -
State: open - Opened by FindHao 2 months ago
- 1 comment
#93 - Remove hstu install
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#93 - Remove hstu install
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#92 - Update to generative-recommenders@7906fe79
Pull Request -
State: closed - Opened by bertmaher 2 months ago
- 2 comments
Labels: cla signed, Merged
#92 - Update to generative-recommenders@7906fe79
Pull Request -
State: closed - Opened by bertmaher 2 months ago
- 2 comments
Labels: cla signed, Merged
#91 - [FA] add bwd variants for warp spec
Pull Request -
State: open - Opened by manman-ren 2 months ago
- 3 comments
Labels: cla signed
#91 - [FA] add bwd variants for warp spec
Pull Request -
State: open - Opened by manman-ren 2 months ago
- 3 comments
Labels: cla signed
#90 - embedding test failure
Issue -
State: closed - Opened by FindHao 2 months ago
Labels: pt2
#89 - Align default parameters with typical benchmarks
Pull Request -
State: closed - Opened by bertmaher 2 months ago
- 3 comments
Labels: cla signed, fb-exported, Merged
#89 - Align default parameters with typical benchmarks
Pull Request -
State: closed - Opened by bertmaher 2 months ago
- 3 comments
Labels: cla signed, fb-exported, Merged
#88 - Disable donated buffer when benchmarking layer_norm with backwards
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 4 comments
Labels: cla signed, Merged
#88 - Disable donated buffer when benchmarking layer_norm with backwards
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 4 comments
Labels: cla signed, Merged
#87 - NCU trace doesn't take mode arguments
Issue -
State: closed - Opened by FindHao 2 months ago
- 9 comments
#87 - NCU trace doesn't take mode arguments
Issue -
State: closed - Opened by FindHao 2 months ago
- 9 comments
#86 - Too many repeats in kineto_traces for pt2 compiled ops
Issue -
State: closed - Opened by FindHao 2 months ago
- 6 comments
#85 - Explain the kineto trace with an example
Issue -
State: closed - Opened by xuzhao9 2 months ago
- 1 comment
#85 - Explain the kineto trace with an example
Issue -
State: closed - Opened by xuzhao9 2 months ago
- 1 comment
#84 - Fix backward accuracy
Issue -
State: open - Opened by xuzhao9 2 months ago
#84 - Fix backward accuracy
Issue -
State: open - Opened by xuzhao9 2 months ago
#83 - Add flash_attention_benchmark and gemm_benchmark
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#83 - Add flash_attention_benchmark and gemm_benchmark
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#82 - Skip compile tk and colfax
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
Labels: cla signed
#82 - Skip compile tk and colfax
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
Labels: cla signed
#81 - Install tritonbench as a library
Pull Request -
State: closed - Opened by xuzhao9 2 months ago
- 2 comments
Labels: cla signed, Merged
#80 - changing hw rooflines to match xformers
Pull Request -
State: closed - Opened by adamomainz 2 months ago
- 4 comments
Labels: cla signed, fb-exported, Merged