Remove Critical Trace Support (#327)
* Delete core critical-trace files
* Update docs and README
* Update workflows
* Update testing
* Update cmake
* Remove critical trace usage in source code
* Update source/docs/critical_trace.md
- fix spelling
* Formatting
* Update bin/omnitrace-avail/avail.cpp
- statically allocate shared pointers for timemory manager and hash id/aliases to prevent use-after-free errors
[ROCm/rocprofiler-systems commit: 9499e2f521]
This commit is contained in:
Jonathan R. Madsen
کامیت شده توسط
GitHub
GitHub
والد
2a42e3abf0
کامیت
ab551e7eb7
@@ -1,4 +1,4 @@
|
||||
# Generating a Critical Trace
|
||||
# Critical Trace Support
|
||||
|
||||
```eval_rst
|
||||
.. toctree::
|
||||
@@ -6,24 +6,5 @@
|
||||
:maxdepth: 4
|
||||
```
|
||||
|
||||
## Overview
|
||||
|
||||
A critical trace is defined in omnitrace as the most time-consuming path through a parallelized code.
|
||||
The steps for generating a critical trace are:
|
||||
|
||||
1. Enable the `OMNITRACE_CRITICAL_TRACE` setting
|
||||
2. Configure any other relevant critical-trace settings, as needed
|
||||
- `omnitrace-avail --categories settings::critical_trace`
|
||||
3. Execute application
|
||||
4. Locate the JSON files with `call-chain` in their name
|
||||
5. Provide these files to the `omnitrace-critical-trace` executable
|
||||
6. Open generated perfetto file in [ui.perfetto.dev](https://ui.perfetto.dev/)
|
||||
|
||||
## omnitrace-critical-trace Executable
|
||||
|
||||
The `omnitrace-critical-trace` executable post-processes one or more `call-chain` JSON files and generates a perfetto output
|
||||
for visualizing the critical trace.
|
||||
|
||||
**INCOMPLETE**
|
||||
|
||||
This executable is still under-development.
|
||||
Critical trace support has been superseded by causal profiling support.
|
||||
Critical trace support was removed in Omnitrace v1.11.0 due to incomplete implementation.
|
||||
|
||||
@@ -50,10 +50,6 @@ for each variant:
|
||||
- For a binary rewrite: outputs new instrumented binary and exits
|
||||
- For runtime instrumentation or attaching to a process: instructs the application to resume executing and then waits for the application to exit
|
||||
|
||||
### omnitrace-critical-trace: [source/bin/omnitrace-critical-trace](https://github.com/ROCm/omnitrace/tree/main/source/bin/omnitrace-critical-trace)
|
||||
|
||||
Post-processing tool for critical-trace data output by omnitrace.
|
||||
|
||||
## Libraries
|
||||
|
||||
### Common Library: [source/lib/common](https://github.com/ROCm/omnitrace/tree/main/source/lib/common)
|
||||
|
||||
@@ -25,7 +25,6 @@ manage extensions, resources, data, etc.
|
||||
- Background thread records process-, system- and device-level metrics while the application executes
|
||||
- Causal profiling
|
||||
- Quantifies the potential impact of optimizations in parallel codes
|
||||
- Critical trace generation
|
||||
|
||||
### Data Analysis
|
||||
|
||||
@@ -35,7 +34,6 @@ manage extensions, resources, data, etc.
|
||||
- Comprehensive traces
|
||||
- Every individual event/measurement
|
||||
- Application speedup predictions resulting from potential optimizations in functions and lines of code (causal profiling)
|
||||
- Critical trace analysis (alpha)
|
||||
|
||||
### Parallelism API Support
|
||||
|
||||
|
||||
@@ -191,13 +191,7 @@ OMNITRACE_USE_PID = true
|
||||
OMNITRACE_OUTPUT_PATH = omnitrace-%tag%-output
|
||||
OMNITRACE_OUTPUT_PREFIX =
|
||||
OMNITRACE_CI = false
|
||||
OMNITRACE_CRITICAL_TRACE = false
|
||||
OMNITRACE_CRITICAL_TRACE_BUFFER_COUNT = 2000
|
||||
OMNITRACE_CRITICAL_TRACE_COUNT = 0
|
||||
OMNITRACE_CRITICAL_TRACE_DEBUG = false
|
||||
OMNITRACE_THREAD_POOL_SIZE = 8
|
||||
OMNITRACE_CRITICAL_TRACE_PER_ROW = 0
|
||||
OMNITRACE_CRITICAL_TRACE_SERIALIZE_NAMES = false
|
||||
OMNITRACE_DEBUG = false
|
||||
OMNITRACE_DL_VERBOSE = 0
|
||||
OMNITRACE_INSTRUMENTATION_INTERVAL = 1
|
||||
@@ -283,13 +277,7 @@ $ omnitrace-avail -S -bd
|
||||
| OMNITRACE_CONFIG_FILE | Configuration file for omnitrace |
|
||||
| OMNITRACE_COUT_OUTPUT | Write output to stdout |
|
||||
| OMNITRACE_CPU_AFFINITY | Enable pinning threads to CPUs (Linu... |
|
||||
| OMNITRACE_CRITICAL_TRACE | Enable generation of the critical trace |
|
||||
| OMNITRACE_CRITICAL_TRACE_BUFFER_COUNT | Number of critical trace records to ... |
|
||||
| OMNITRACE_CRITICAL_TRACE_COUNT | Number of critical trace to export (... |
|
||||
| OMNITRACE_CRITICAL_TRACE_DEBUG | Enable debugging for critical trace |
|
||||
| OMNITRACE_THREAD_POOL_SIZE | Number of threads to use when genera... |
|
||||
| OMNITRACE_CRITICAL_TRACE_PER_ROW | How many critical traces per row in ... |
|
||||
| OMNITRACE_CRITICAL_TRACE_SERIALIZE_N... | Include names in serialization of cr... |
|
||||
| OMNITRACE_DEBUG | Enable debug output |
|
||||
| OMNITRACE_DIFF_OUTPUT | Generate a difference output vs. a p... |
|
||||
| OMNITRACE_DL_VERBOSE | Verbosity within the omnitrace-dl li... |
|
||||
|
||||
@@ -209,7 +209,6 @@ $ omnitrace-sample -- ./parallel-overhead-locks 30 4 100
|
||||
HSA_TOOLS_LIB=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
HSA_TOOLS_REPORT_LOAD_FAILURE=1
|
||||
LD_PRELOAD=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
OMNITRACE_CRITICAL_TRACE=false
|
||||
OMNITRACE_USE_PROCESS_SAMPLING=false
|
||||
OMNITRACE_USE_SAMPLING=true
|
||||
OMP_TOOL_LIBRARIES=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
@@ -228,7 +227,6 @@ HSA_TOOLS_REPORT_LOAD_FAILURE=1
|
||||
KOKKOS_PROFILE_LIBRARY=/opt/omnitrace/lib/libomnitrace.so.1.7.1
|
||||
LD_PRELOAD=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
OMNITRACE_CPU_FREQ_ENABLED=true
|
||||
OMNITRACE_CRITICAL_TRACE=false
|
||||
OMNITRACE_TRACE_THREAD_LOCKS=true
|
||||
OMNITRACE_TRACE_THREAD_RW_LOCKS=true
|
||||
OMNITRACE_TRACE_THREAD_SPIN_LOCKS=true
|
||||
@@ -258,7 +256,6 @@ $ omnitrace-sample -PTDH -E all -o omnitrace-output %tag% -- ./parallel-overhead
|
||||
|
||||
LD_PRELOAD=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
OMNITRACE_CPU_FREQ_ENABLED=true
|
||||
OMNITRACE_CRITICAL_TRACE=false
|
||||
OMNITRACE_OUTPUT_PATH=omnitrace-output
|
||||
OMNITRACE_OUTPUT_PREFIX=%tag%
|
||||
OMNITRACE_TRACE_THREAD_LOCKS=false
|
||||
@@ -288,7 +285,6 @@ $ omnitrace-sample -PTDH -E all -o omnitrace-output %tag% -c -- ./parallel-overh
|
||||
LD_PRELOAD=/opt/omnitrace/lib/libomnitrace-dl.so.1.7.1
|
||||
OMNITRACE_CONFIG_FILE=
|
||||
OMNITRACE_CPU_FREQ_ENABLED=true
|
||||
OMNITRACE_CRITICAL_TRACE=false
|
||||
OMNITRACE_OUTPUT_PATH=omnitrace-output
|
||||
OMNITRACE_OUTPUT_PREFIX=%tag%
|
||||
OMNITRACE_TRACE_THREAD_LOCKS=false
|
||||
|
||||
مرجع در شماره جدید
Block a user