Ecosyste.ms: Issues

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

GitHub / facebookresearch/HolisticTraceAnalysis issues and pull requests

#206 - Is it possible to categorize the kernels befween forward and backward passes

Issue - State: open - Opened by mpashkovskiy about 17 hours ago
Labels: help wanted, question, needs triage

#205 - idle time breakdown for MTIA

Pull Request - State: closed - Opened by fenypatel99 10 days ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#204 - Validation to pick up new fields in the trace

Pull Request - State: closed - Opened by sanrise 14 days ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#203 - add test for MTIA kernel breakdown and temporal breakdown

Pull Request - State: closed - Opened by fenypatel99 15 days ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#202 - queue length for MTIA

Pull Request - State: closed - Opened by fenypatel99 18 days ago - 8 comments
Labels: CLA Signed, Merged, fb-exported

#201 - kernel launch statistics for MTIA

Pull Request - State: closed - Opened by fenypatel99 18 days ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#200 - categorize "dma_request" as mem kernel

Pull Request - State: closed - Opened by fenypatel99 21 days ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#199 - Add kernel hash parsing and suport local minor updates for fields

Pull Request - State: closed - Opened by briancoutinho 25 days ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#198 - add some debug prints to time at max queue length

Pull Request - State: closed - Opened by briancoutinho about 1 month ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#197 - Fix HTA OSS CI errors

Pull Request - State: closed - Opened by niufei8888 about 1 month ago - 11 comments
Labels: CLA Signed, Merged, fb-exported

#196 - Add pre-commit to the dev requirement

Pull Request - State: open - Opened by niufei8888 about 1 month ago
Labels: CLA Signed

#195 - Unify yaml file reading for both OSS and fbcode

Pull Request - State: closed - Opened by niufei8888 about 1 month ago - 8 comments
Labels: CLA Signed, fb-exported

#194 - Fix test errors and type checking errors induced by a previous PR.

Pull Request - State: open - Opened by fengxizhou about 1 month ago - 3 comments
Labels: CLA Signed, fb-exported

#193 - BUG: Test Failures Caused By Wrong parse_event_args_yaml Path

Issue - State: open - Opened by fengxizhou about 1 month ago - 1 comment
Labels: bug, needs triage

#192 - Memory usage

Pull Request - State: open - Opened by fengxizhou about 1 month ago
Labels: CLA Signed

#191 - Add utility to analyze the memory usage of internal trace representation

Issue - State: open - Opened by fengxizhou about 1 month ago
Labels: feature request, needs triage

#190 - Step 3: Update the memory bandwidth type in the trace format

Pull Request - State: closed - Opened by niufei8888 about 2 months ago - 1 comment
Labels: CLA Signed, fb-exported

#189 - Trace format validation

Pull Request - State: closed - Opened by fengxizhou about 2 months ago - 3 comments
Labels: CLA Signed, Merged

#188 - [Trace format versioning] Step2: Add versioning to parserconfig

Pull Request - State: closed - Opened by niufei8888 about 2 months ago - 10 comments
Labels: CLA Signed, Merged, fb-exported

#187 - [Trace format versioning] Step1: Add yaml event args

Pull Request - State: closed - Opened by niufei8888 about 2 months ago - 14 comments
Labels: CLA Signed, Merged, fb-exported

#186 - Feature Request: Improve Trace Format Validation and Resilience in HTA

Issue - State: open - Opened by fengxizhou about 2 months ago - 1 comment
Labels: feature request, needs triage

#185 - Sync up and add new parser config fields

Pull Request - State: closed - Opened by briancoutinho about 2 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#184 - Avoid circular deps and split out build targets 1/n

Pull Request - State: closed - Opened by briancoutinho about 2 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#183 - replace uses of agg("operation") with aggregate_func()

Pull Request - State: closed - Opened by igorsugak 3 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#182 - replace uses of suffixes as list-of-two to tuple in DataFrame.merge invocations

Pull Request - State: closed - Opened by igorsugak 3 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#181 - add parser config for kernel_backend and test example

Pull Request - State: closed - Opened by briancoutinho 3 months ago - 3 comments
Labels: CLA Signed, Merged

#180 - Kernel Breakdown by Annotation Range

Issue - State: open - Opened by jeromeku 3 months ago - 2 comments
Labels: feature request, needs triage

#179 - add option to pass trace_file list to Trace() object

Pull Request - State: closed - Opened by briancoutinho 3 months ago - 3 comments
Labels: CLA Signed, Merged

#178 - Time blocked on kernel queue full

Pull Request - State: closed - Opened by briancoutinho 3 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#177 - add env options and disable negative weights check

Pull Request - State: closed - Opened by briancoutinho 4 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#176 - Add function level timing measurements

Pull Request - State: closed - Opened by briancoutinho 4 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#175 - [CPA] Additional optimizations to graph constructions

Pull Request - State: closed - Opened by briancoutinho 4 months ago - 2 comments
Labels: CLA Signed, Merged

#174 - add data_provider test utility

Pull Request - State: closed - Opened by fengxizhou 4 months ago - 2 comments
Labels: CLA Signed, Merged

#173 - Add option to filter events in call stack

Pull Request - State: closed - Opened by briancoutinho 4 months ago - 3 comments
Labels: CLA Signed, Merged

#172 - Feature Request: Implement `data_provider` Utility for HTA

Issue - State: open - Opened by fengxizhou 4 months ago
Labels: feature request, needs triage

#171 - [critical path] Optimize graph construction of node events

Pull Request - State: closed - Opened by briancoutinho 4 months ago - 3 comments
Labels: CLA Signed, Merged

#170 - Normalize stream numbers

Pull Request - State: closed - Opened by jj10306 4 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#169 - Add ns rounding to ijson loader too, make it configurable

Pull Request - State: closed - Opened by briancoutinho 5 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#168 - change WPS' HTA parsing backend to illustrate potential differences

Pull Request - State: open - Opened by jj10306 5 months ago - 1 comment
Labels: CLA Signed, fb-exported

#167 - Fix Typo: 'computer' -> 'compute'

Pull Request - State: closed - Opened by mkyybx 5 months ago - 7 comments
Labels: CLA Signed, Merged, fb-exported

#166 - Add option to disable memory profiling when using multiprocessing

Pull Request - State: closed - Opened by pavanky 5 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#165 - Fix Ijson metadata reader corner cases

Pull Request - State: closed - Opened by briancoutinho 5 months ago - 5 comments
Labels: CLA Signed, Merged

#164 - Update parser config to accept and update a parser backend

Pull Request - State: closed - Opened by jj10306 5 months ago - 2 comments
Labels: CLA Signed, fb-exported

#163 - Update et_replay reference to latest module path.

Pull Request - State: closed - Opened by sanrise 5 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#162 - Correct Comm/comp Overlap Calculation

Pull Request - State: closed - Opened by mkyybx 5 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#161 - 160 test failure on test execution trace

Pull Request - State: closed - Opened by fengxizhou 5 months ago - 2 comments
Labels: CLA Signed, Merged

#160 - Test Failure on test_execution_trace

Issue - State: closed - Opened by fengxizhou 5 months ago
Labels: bug, needs triage

#159 - Bump black from 22.8.0 to 24.3.0

Pull Request - State: closed - Opened by dependabot[bot] 5 months ago - 3 comments
Labels: CLA Signed, Merged, dependencies

#158 - Log API Usage Feature

Issue - State: open - Opened by fengxizhou 5 months ago
Labels: feature request, needs triage

#157 - Make HTA compatible with On-demand NCCL traces

Pull Request - State: closed - Opened by sraikund16 5 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#156 - Self-Dependencies in External Event IDs in SYNC_DEPENDENCY

Issue - State: open - Opened by TaekyungHeo 5 months ago
Labels: help wanted, question, needs triage

#155 - Support more memory copy types

Pull Request - State: closed - Opened by shengfukevin 5 months ago - 9 comments
Labels: CLA Signed, Merged, fb-exported

#154 - Conditionally skip unit test in test_execution_trace.py

Pull Request - State: closed - Opened by fengxizhou 5 months ago - 2 comments
Labels: CLA Signed, Merged

#153 - fix submodule dependency for test_execution_trace in .github/workflo…

Pull Request - State: closed - Opened by fengxizhou 5 months ago - 2 comments
Labels: CLA Signed, Merged

#152 - add github/workflows/ci.yml

Pull Request - State: closed - Opened by fengxizhou 5 months ago - 2 comments
Labels: CLA Signed, Merged

#151 - Create a continuous integration (CI) workflow for building and testing HTA

Issue - State: open - Opened by fengxizhou 5 months ago
Labels: feature request, needs triage

#150 - nccl defaults and add env variable to disable call stack depth

Pull Request - State: closed - Opened by briancoutinho 5 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#149 - Fix Idle time breakdown analysis

Pull Request - State: closed - Opened by jj10306 5 months ago - 11 comments
Labels: CLA Signed, Merged, fb-exported

#148 - Analyzer reads the log files correctly but doesn't show data

Issue - State: open - Opened by oabuhamdan 6 months ago
Labels: bug, needs triage

#147 - Migrate memory bandwidth analyzer to HTA

Pull Request - State: closed - Opened by shengfukevin 6 months ago - 10 comments
Labels: CLA Signed, Merged, fb-exported

#146 - CUPTI Counter Analysis empty

Issue - State: open - Opened by jeromeku 6 months ago - 1 comment
Labels: bug

#145 - Is there any visualization tool for HTA, similar to the visualization interface of tensorboard plugin?

Issue - State: closed - Opened by GuWei007 6 months ago - 5 comments
Labels: help wanted, question, needs triage

#144 - [Question] Can HTA work on other trace files (not generated by pytorch) too?

Issue - State: open - Opened by Sarbojit2019 6 months ago - 1 comment
Labels: help wanted, question, needs triage

#143 - add nccl field parser config and test

Pull Request - State: closed - Opened by briancoutinho 6 months ago - 2 comments
Labels: CLA Signed, Merged

#142 - Fix setting negative weight to 0

Pull Request - State: closed - Opened by pavanky 6 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#141 - Why are there call_stack.py and trace_call_stack.py at the same time?

Issue - State: open - Opened by zhouyiyuan-mt 6 months ago
Labels: help wanted, question, needs triage

#140 - Evaluate logger.debug only when needed

Pull Request - State: closed - Opened by pavanky 7 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#139 - Fix CPU kernel start time check when `include_last_profiler_step=True`

Pull Request - State: closed - Opened by jj10306 7 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#138 - [critical path] Add tolerance for negative one weight due to precision issues, improvements

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 4 comments
Labels: CLA Signed, Merged

#137 - Enable comm_replay in PARAM by Integrating and Refactoring Comm Code

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#136 - Add get_stack_of_node to CallGraph

Pull Request - State: closed - Opened by pavanky 7 months ago - 9 comments
Labels: CLA Signed, Merged, fb-exported

#135 - Add parent to dataframe in call_stack

Pull Request - State: closed - Opened by pavanky 7 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#134 - Filter out non kernel events for gpu stacks

Pull Request - State: closed - Opened by pavanky 7 months ago - 8 comments
Labels: CLA Signed, Merged, fb-exported

#133 - Fix sorting of events in call_graph

Pull Request - State: closed - Opened by pavanky 7 months ago - 10 comments
Labels: CLA Signed, Merged, fb-exported

#132 - 0507 add metadata parser ijson

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 4 comments
Labels: CLA Signed, Merged

#131 - Draft refactor of et replay

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#130 - correct typos in docs for function get_queue_length_time_series

Pull Request - State: closed - Opened by staugust 7 months ago - 3 comments
Labels: CLA Signed, Merged

#129 - function get_queue_length_series not found

Issue - State: closed - Opened by staugust 7 months ago
Labels: documentation, needs triage

#128 - 0501 update gpu kernel filtering

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 4 comments
Labels: CLA Signed, Merged

#127 - Performance improvements to call_stack

Pull Request - State: closed - Opened by pavanky 7 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#126 - [critical path] Add save and restore for cp_graph

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 8 comments
Labels: CLA Signed, Merged, fb-exported

#125 - add attribution for kernel-kernel delay and check for sync on same strea

Pull Request - State: closed - Opened by briancoutinho 7 months ago - 3 comments
Labels: CLA Signed, Merged

#124 - Estimate TFLOPS of PyTorch Matrix Multiplication Operators from Kineto Trace

Issue - State: open - Opened by fengxizhou 8 months ago
Labels: feature request, needs triage

#122 - Categorizing ncclDevKernel_AllReduce_Sum_f32_RING_LL as Computation

Issue - State: closed - Opened by OckermanSethGVSU 8 months ago - 1 comment
Labels: bug, needs triage

#121 - [Critical path] determine previous kernel for an event using the stream

Pull Request - State: closed - Opened by briancoutinho 8 months ago - 5 comments
Labels: CLA Signed, Merged

#120 - Fix parsing no fwdbwd, add unit test for ns duration and attempt work around

Pull Request - State: closed - Opened by briancoutinho 8 months ago - 3 comments
Labels: CLA Signed, Merged

#119 - Clarify how traces are collected + Some Minor Documentation Updates

Pull Request - State: open - Opened by wkaisertexas 8 months ago - 2 comments
Labels: CLA Signed

#118 - A faster way to load HTA Trace and create CallGraph

Pull Request - State: closed - Opened by pavanky 8 months ago - 13 comments
Labels: CLA Signed, Merged, fb-exported

#117 - Critical path analysis - matching the kernel related to cudaEventRecord with stream

Issue - State: closed - Opened by briancoutinho 8 months ago - 3 comments
Labels: bug

#116 - [critical path] Add graph validation checks and fix 0 duration stack issue.

Pull Request - State: closed - Opened by briancoutinho 8 months ago - 3 comments
Labels: CLA Signed, Merged, fb-exported

#115 - [1/n] Optimize performance Critical Path Analysis algorithm for CUDA sync events

Pull Request - State: closed - Opened by briancoutinho 8 months ago - 5 comments
Labels: CLA Signed, Merged

#114 - Add an interim fix for stack traversal order in HTA

Pull Request - State: closed - Opened by briancoutinho 9 months ago - 4 comments
Labels: CLA Signed, Merged, fb-exported

#113 - New Trace Callstack processing out of order

Issue - State: open - Opened by briancoutinho 9 months ago
Labels: bug, needs triage

#112 - Fixing Unary Op evaluate issue

Pull Request - State: closed - Opened by amoghavs 9 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#111 - Fix Undefined local_symbol_table if "traceEvents" is not in "trace_record"

Pull Request - State: closed - Opened by mkyybx 9 months ago - 2 comments
Labels: CLA Signed, Merged, fb-exported

#110 - Trace load and json parsing optimizations

Pull Request - State: closed - Opened by briancoutinho 9 months ago - 4 comments
Labels: CLA Signed, Merged

#108 - add knob to turn off causal edges

Pull Request - State: closed - Opened by briancoutinho 9 months ago - 5 comments
Labels: CLA Signed, Merged, fb-exported

#107 - min(arg) is an empty sequence -> issues creating analyzer

Issue - State: closed - Opened by wkaisertexas 9 months ago - 10 comments
Labels: question