AI/rocm-systems
2
0
Bifurcation 0
Code Tickets Demandes d'ajout Actions Paquets Projets Publications Wiki Activité
Fichiers
aadffbe2b11cd9849a2989b721ac8adca4cee5bc
rocm-systems/source/docs/output.md
T
Jonathan R. Madsen 9618ddefba Causal profiling (#229) 
* Addition of basic structure

* Reworked categories

* More causal integration additions

* Causal implementation

* Update examples

* delete virtual_speedup files

* Update perfetto submodule to v31.0

* Update dyninst submodule

* Update timemory submodule

* ElfUtils build for libdw

* OMNITRACE_LIKELY and OMNITRACE_UNLIKELY

* Update common lib join

* Examples updates for causal profiling

* config updates with causal options

- OMNITRACE_CAUSAL_FIXED_LINE
- OMNITRACE_CAUSAL_FIXED_SPEEDUP
- OMNITRACE_CAUSAL_FILE
- OMNITRACE_CAUSAL_BINARY_SCOPE
- OMNITRACE_CAUSAL_SOURCE_SCOPE
- version info in banner
- support increments in parse_numeric_range
- fix occasional deadlock in first call to get_config

* PTL general task group

* Always include PID in debug/verbose messages

* Add blocking/unblocking gotchas to runtime init bundle

* CausalState

* thread_data updates

- generic component_bundle_cache

* Improve handling of causal in category_region

* components updates

- backtrace_causal component
- backtrace::get_data member func
- decrease ignore_depth in backtrace::sample(int)
- handle "omnitrace_main" in backtrace::filter_and_patch(...)
- tweak internal thread state scope for pthread_mutex_gotcha wrappers

* simplify tracing get_instrumentation_bundles usage

* sampling updates

- include backtrace_causal component
- disable backtrace_metrics if using causal and not using perfetto
- disable backtrace and backtrace_timestamp when using causal
- post_process_causal

* causal updates

- more checks in blocking_gotcha and unblocking_gotcha start/stop
- miscellaneous overhaul of data
- experiment update

* Remove virtual speedup

* libomnitrace code_object

* causal-profiling test

* libomnitrace library.cpp updates

- handle causal profiling
- fini_bundle

* Disable causal profiling by default

* Updated causal code and example

- example: three execution variants: cpu + rng, cpu, rng
- example: three instrumentation variants: none, omni, coz
- fix blocking gotcha credit
- rework perform_experiment_impl
- get_eligible_address_ranges
- compute_eligible_lines
- support fixed lines/speedups/functions
- update selected_entry to support function mode
- fix causal::delay
- experiment updates

* omnitrace_progress / omnitrace_user_progress

- with accompanying omnitrace_annotated_progress / omnitrace_user_annotated_progress

* Update timemory submodule

* CausalMode

- mode indicated whether causal predictions source be at line-level or function-level

* code_object, config, runtime, sampling, thread_data

- code_object: address_range
- code_object: basic::line_info serialize(), name(), hash()
- config updates
- two signals for causal sampling
- thread_data init fixes

* pthread updates

- pthread_create_gotcha processes delays
- pthread_mutex_gotcha does not wrap pthread_join in causal mode

* backtrace_causal update

- dynamic delay period stats

* main wrapper uses basename of argv[0]

* update elfio submodule

* perf support (currently unused)

* Fix experiment JSON serialization

- static_vector.hpp (unused)

* causal executable + config options updates

- omnitrace-causal exe simplifies running multiple causal configs
- changed the causal config option names

* Support both throughput and latency points

* process-causal-json.py script

- will be used later for testing

* stable_vector

* Rework thread_data

* Improve omnitrace-causal exe

- better verbosity handling
- correct diagnosis of status for child process
- execvpe when only one iteration (debugging)

* Update timemory submodule

* exe --version

- omnitrace, omnitrace-avail, and omnitrace-sample all support --version on command-line

* OMNITRACE_INTERNAL_API + OMNITRACE_{LIKELY,UNLIKELY}

* omnitrace-causal cmake format

* omnitrace config update

- OMNITRACE_CAUSAL_FILE_CLOBBER

* custom exception

- wraps STL exception and gets stacktrace during construction

* exit_gotcha supports _Exit

* use global construct_on_init + max threads

- add some safety when exceeding max # of threads

* update code_object binary filter

- exclude dyninst and tbbmalloc library

* containers: c_array, static_vector, stable_vector

- moved utility::c_array to container::c_array
- created static_vector: std::vector bound to std::array
- created stable_vector: vector with stable references

* grow thread_data when new thread created

* causal updates

- data: improve compute_eligible_lines to ignore lambdas
- data: use new thread_data
- delay: use new thread_data
- experiment: properly support latency points
- experiment: support file clobber
- experiment: ensure non-zero experiment time
- progress_point: use new thread_data
- backtrace_causal: use new thread_data

* Update causal-profiling tests

* fix omnitrace-causal backslash escaping

* process-causal-json script

* restructure causal implementation

- update verbose messages for omnitrace-causal diagnose_status
- migrated causal implementation in sampling.cpp to causal/sampling.cpp
- OMNITRACE_USE_CAUSAL does not require OMNITRACE_USE_SAMPLING
- added Mode::Causal
- causal sampling uses same signals as regular sampling
- moved tracing::thread_init to implementation file
- combined tracing::thread_init and tracing::thread_init_sampling
- added causal/components folder
- pthread_create_gotcha::wrapper_config
- omnitrace_preload checks OMNITRACE_USE_CAUSAL
  - updates mode accordingly

* update timemory submodule

* update timemory submodule

* causal example updates

- causal for lulesh

* perf code + utility - helpers

- relocated causal perf code
- placement new when generating unique ptr trait for potentially allocating during sampling
- additions to utility header
- removed previously added helpers.hpp

* update timemory submodule

* Default env variables for omnitrace-causal

- activate OMNITRACE_USE_KOKKOSP, etc.

* update stable_vector and static_vector

- static vector can use atomic for size tracking for thread-safe situations

* update causal example header

- CAUSAL_PROGRESS_NAMED
- use CAUSAL_ prefix for some macros

* Tweak lulesh example

- use CAUSAL_PROGRESS instead of CAUSAL_BEGIN and CAUSAL_END

* omnitrace-sample support for causal mode

- set OMNITRACE_USE_SAMPLING to off when OMNITRACE_MODE=causal

* refactor and cleanup code_object

- scope filter
- fixes to address_range

* overhaul causal data + causal config options

- full support for function and line mode
- support static vector of instruction pointers
- improve line info mapping resolution
- remove thread-locality from miscellanous functions where unnecessary
- causal options for {binary,source,function,fileline} exclusion

* causal experiment, sampling, and backtrace updates

- is_selected + unwind address array
- experiment warning about progress points
- increased buffer size for backtrace_casual sampler
- backtrace_causal only stores IP addresses instead of full unwind info

* category_region updates

- minor refactor
- local_category_region::mark

* Update causal tests

* Bump version to 1.8.0

* omnitrace-causal args + CLOBBER -> RESET

- renamed OMNITRACE_CAUSAL_FILE_CLOBBER to OMNITRACE_CAUSAL_FILE_RESET
- updated omnitrace-causal exe to support recently added configuration options
- other miscellaneous tweaks to data.cpp, experiment.cpp, and sampling.cpp

* Refactor causal and code_object

- code_object.hpp and code_object.cpp moved into binary folder
- causal components namespaced into omnitrace::causal::component
- moved sample_data out of backtrace_causal and into own file
- renamed backtrace_causal to causal::component::backtrace

* preload omnitrace_init + OMNITRACE_DEBUG_MARK

- env OMNITRACE_DEBUG_MARK
- fix omnitrace_init call when LD_PRELOAD-ing omnitrace

* Fix fileline support + line-info output names + experiment log

- line-info log files are prefixed with experiment name
- don't print experiment duration when E2E
- account for fileline scope in analysis

* KokkosP: OMNITRACE_KOKKOSP_NAME_LENGTH_MAX

- config option to limit the name of kokkos tool callbacks
- remove [kokkos] from KokkosP names

* Update causal example

- minor tweaks to decrease probability of overlapping regions in binary

* omnitrace-causal update

- prefix N / Ntot in environment printout

* Miscellaneous updates

- causal::finish_experimenting()
- OMNITRACE_CAUSAL_RANDOM_SEED
- KokkosP causal updates
  - exclude some callbacks, make some callbacks unique, etc.
- address_range::operator+=(address_range)
- combine contiguous ranges in binary/analysis.cpp when file, func, line is same and address range is contiguous
- bfd_line_info reads inline info
- wait for perform_experiment_impl to complete
- causal::delay updates
  - delay::process checks if experiment is active
  - uses threading::get_id()
- experiment scales duration up for larger speedup experiments
- line info samples includes excluded lines
- sampler uses CLOCK_REALTIME
- blocking_gotcha updates
  - is no longer fully static
  - adds audit routine which sets the postblock value to zero if try/timed routine fails
- category::host was added to causal_throughput_categories_t
- pthread_create_gotcha sets new threads local parent delay
  - was using internal value, now uses sequent value

* Causal improvements to KokkosP

* Updates to experiment time scaling

- use stats instead of just max

* binary/link_map.{hpp,cpp}

* update process-causal-json.py

* Folded fileline scope into source scope

* Update documentation

- Add documentation for causal profiling
- Replace 'Omnitrace' with 'OmniTrace' everywhere

* Update causal-helpers.cmake + omnitrace-testing.cmake

- split tests/CMakeLists.txt partially into omnitrace-testing.cmake

* omnitrace/causal.h

- OMNITRACE_CAUSAL_PROGRESS
- OMNITRACE_CAUSAL_PROGRESS_NAMED
- OMNITRACE_CAUSAL_BEGIN
- OMNITRACE_CAUSAL_END

* selected_entry + remove default filters for lambdas and operator()

- selected entry stores range and binary load address

* update process-causal-json.py

* format examples/lulesh/CMakeLists.txt

* causal-helpers find_package(Threads)

* OMNITRACE_KOKKOSP_KERNEL_LOGGER

- was OMNITRACE_KOKKOS_KERNEL_LOGGER

* quiet find of coz-profiler

* Fix rocm_smi exception handling

* Update timemory submodule (binutils)

- fix binutls compile error on some systems
- bump binutils to v2.40

* Fix miscellaneous tests

* OMNITRACE_KOKKOSP_PREFIX

* revert rocm_smi handling

* ElfUtils updates

- default to download version 0.188
- add -Wno-error=null-dereference due to GCC 12 compiler error

* Update causal example

* Remove OMNITRACE_VERBOSE from global workflow envs

* Reliable causal test

* disable compilation of causal perf files

* Remove set_current_selection with unwind stack

* update timemory submodule

* fix for segfault on bionic

- locking in TLS dtor was causing segfault

* remove experiment::is_selected(unwind_stack_t)

* update default init of selected_entry

* Fix for when IP is not offset by load address

* Update CMakeLists.txt

* Miscellaneous updates

- OMNITRACE_WARNING_OR_CI_THROW
- OMNITRACE_REQUIRE
- OMNITRACE_PREFER
- fixed issues with no ASLR
-  added load address variable and ipaddr() func to basic/bfd line info
- removed get_basic() from dwarf_line_info
- TIMEMORY_PREFER -> OMNITRACE_PREFER
- removed previously added binary_address and range variables from selected_entry

* Removed superfluous CausalState

* Additional causal tests (lulesh + kokkos)

* filter, prefer, analysis ASLR handling

- removed default filter on cold functions
- fixed OMNITRACE_PREFER
- fixed analysis ASLR handling

* Tweak line-info output

* Removed some superfluous code

- causal/delay
- causal/selected_entry

* Exclude main.cold in function mode

* Update validate-perfetto-proto.py

- account for occasional http errors

* Add sampling test disabling tmp files

* argparser for process-causal-json

- support validation
- support filtering

* Avoid pthread_{lock,unlock} in sampling offload

- use homemade atomic_mutex/atomic_lock since contention will be low and using pthread tools might trigger our wrappers

* Rename process-causal-json.py

- validate-causal-json.py

* rework omnitrace_add_causal_test

- capable of performing validation
- added validation tests

* Fix kokkosp_begin_deep_copy + causal

* Tweak address range in bfd_line_info::read_pc

* Tweak analysis and data IP handling

- look for gaps

* Disable scaling experiment time by speedup

* Revert change in max threads during CI

* binary updates

- significant overhaul of binary analysis implementation
- removed "basic_line_info" and "bfd_line_info" in lieu of "symbol" class
  - symbol class has basic BFD info + vector of inlines + vector of dwarf info

* Updated causal to use new binary analysis

- Fix symbol.cpp includes

* Updated formatting target

- include *.cmake files

* Updated causal tests

- causal tests should be stable now

* Update timemory and dyninst submodules

- TPLs are stripped + built w/o debug info

* Increase tolerance for causal validation speedups

- higher speedups have more variance (increased to +/- 5 from 3)

* Support causal output for MPI

- i.e. tag with MPI rank

* omnitrace-causal launcher argument

* improve experiment sampling output

* causal data updates

- call compute lines once
- fixed filtered cached binary info
- debugging info when experiment fails to start

* Tweaked causal validation tests

* dwarf_entry ranges

* CI updates

- increase max threads to 64

* Tweak causal E2E validation tests

- more threads
- shorter thread runtime
- more iterations

* Fix shadowed variable

* fix symbol read_bfd last PC calculation

* fix maybe-uninitialized warning

* omnitrace-causal launcher update

- only inject "omnitrace-causal --" once
- throw error if no matches found

* Update causal profiling docs for launcher

* fix address range boundaries
2023-01-24 18:53:23 -06:00

889 lignes
65 KiB
Markdown
Brut Annotations Historique

# OmniTrace Output
```eval_rst
.. toctree::
:glob:
:maxdepth: 3
```
## Overview
The general output form of omnitrace is `<OUTPUT_PATH>[/<TIMESTAMP>]/[<PREFIX>]<DATA_NAME>[-<OUTPUT_SUFFIX>].<EXT>`.
E.g. with the base configuration:
```shell
export OMNITRACE_OUTPUT_PATH=omnitrace-example-output
export OMNITRACE_TIME_OUTPUT=ON
export OMNITRACE_USE_PID=OFF
export OMNITRACE_USE_TIMEMORY=ON
export OMNITRACE_USE_PERFETTO=ON
```
```shell
$ omnitrace -- ./foo
...
[omnitrace] Outputting 'omnitrace-example-output/perfetto-trace.proto'...
[omnitrace] Outputting 'omnitrace-example-output/wall-clock.txt'...
[omnitrace] Outputting 'omnitrace-example-output/wall-clock.json'...
```
If we enable the `OMNITRACE_USE_PID` option, then when our non-MPI executable is executed with a PID of 63453:
```shell
$ export OMNITRACE_USE_PID=ON
$ omnitrace -- ./foo
...
[omnitrace] Outputting 'omnitrace-example-output/perfetto-trace-63453.proto'...
[omnitrace] Outputting 'omnitrace-example-output/wall-clock-63453.txt'...
[omnitrace] Outputting 'omnitrace-example-output/wall-clock-63453.json'...
```
If we enable `OMNITRACE_TIME_OUTPUT`, then a job started on January 31, 2022 at 12:30 PM:
```shell
$ export OMNITRACE_TIME_OUTPUT=ON
$ omnitrace -- ./foo
...
[omnitrace] Outputting 'omnitrace-example-output/2022-01-31_12.30_PM/perfetto-trace-63453.proto'...
[omnitrace] Outputting 'omnitrace-example-output/2022-01-31_12.30_PM/wall-clock-63453.txt'...
[omnitrace] Outputting 'omnitrace-example-output/2022-01-31_12.30_PM/wall-clock-63453.json'...
```
## Metadata
[OmniTrace](https://github.com/AMDResearch/omnitrace) will output a metadata.json file. This metadata file will contain
information about the settings, environment variables, output files, and info about the system and the run:
- Hardware cache sizes
- Physical CPUs
- Hardware concurrency
- CPU model, frequency, vendor, and features
- Launch date and time
- Memory maps (e.g. shared libraries)
- Output files
- Environment Variables
- Configuration Settings
### Metadata JSON Sample
```json
{
"omnitrace": {
"metadata": {
"info": {
"HW_L1_CACHE_SIZE": 32768,
"HW_L2_CACHE_SIZE": 524288,
"HW_L3_CACHE_SIZE": 16777216,
"HW_PHYSICAL_CPU": 12,
"HW_CONCURRENCY": 24,
"LAUNCH_TIME": "02:04",
"LAUNCH_DATE": "05/08/22",
"TIMEMORY_GIT_REVISION": "52e7034fd419ff296506cdef43084f6071dbaba1",
"TIMEMORY_VERSION": "3.3.0rc4",
"TIMEMORY_API": "tim::project::timemory",
"TIMEMORY_GIT_DESCRIBE": "v3.2.0-263-g52e7034f",
"PWD": "/home/jrmadsen/devel/c++/AARInternal/hosttrace-dyninst/build-vscode",
"USER": "jrmadsen",
"HOME": "/home/jrmadsen",
"SHELL": "/bin/bash",
"CPU_MODEL": "AMD Ryzen Threadripper PRO 3945WX 12-Cores",
"CPU_FREQUENCY": 2400,
"CPU_VENDOR": "AuthenticAMD",
"CPU_FEATURES": [
"fpu",
"msr",
"sse",
"sse2",
"constant_tsc",
"ssse3",
"fma",
"sse4_1",
"sse4_2",
"popcnt",
"avx2",
"... etc. ..."
],
"memory_maps": [
{
"end_address": "7f4013797000",
"start_address": "7f4012e58000",
"pathname": "/opt/rocm-5.0.0/hip/lib/libamdhip64.so.5.0.50000",
"offset": "34a000",
"device": "103:05",
"inode": 4331165,
"permissions": "rw-p"
},
{
"end_address": "7f4013902000",
"start_address": "7f4013901000",
"pathname": "/usr/lib/x86_64-linux-gnu/libm-2.31.so",
"offset": "14d000",
"device": "103:05",
"inode": 42078854,
"permissions": "rwxp"
},
{
"end_address": "7f4013919000",
"start_address": "7f4013908000",
"pathname": "/usr/lib/x86_64-linux-gnu/libpthread-2.31.so",
"offset": "6000",
"device": "103:05",
"inode": 42078874,
"permissions": "r-xp"
},
{
"...": "etc."
},
],
"memory_maps_files": [
"/opt/rocm-5.0.0/hip/lib/libamdhip64.so.5.0.50000",
"/opt/rocm-5.0.0/hsa-amd-aqlprofile/lib/libhsa-amd-aqlprofile64.so.1.0.50000",
"/opt/rocm-5.0.0/lib/libamd_comgr.so.2.4.50000",
"/opt/rocm-5.0.0/lib/libhsa-runtime64.so.1.5.50000",
"/opt/rocm-5.0.0/rocm_smi/lib/librocm_smi64.so.5.0.50000",
"/opt/rocm-5.0.0/roctracer/lib/libroctracer64.so.1.0.50000",
"/usr/lib/x86_64-linux-gnu/ld-2.31.so",
"/usr/lib/x86_64-linux-gnu/libc-2.31.so",
"/usr/lib/x86_64-linux-gnu/libdl-2.31.so",
"... etc. ..."
],
},
"output": {
"text": [
{
"value": [
"omnitrace-tests-output/parallel-overhead-binary-rewrite/roctracer.txt"
],
"key": "roctracer"
},
{
"value": [
"omnitrace-tests-output/parallel-overhead-binary-rewrite/wall_clock.txt"
],
"key": "wall_clock"
}
],
"json": [
{
"value": [
"omnitrace-tests-output/parallel-overhead-binary-rewrite/roctracer.json",
"omnitrace-tests-output/parallel-overhead-binary-rewrite/roctracer.tree.json"
],
"key": "roctracer"
},
{
"value": [
"omnitrace-tests-output/parallel-overhead-binary-rewrite/wall_clock.json",
"omnitrace-tests-output/parallel-overhead-binary-rewrite/wall_clock.tree.json"
],
"key": "wall_clock"
}
]
},
"environment": [
{
"value": "/home/jrmadsen",
"key": "HOME"
},
{
"value": "/bin/bash",
"key": "SHELL"
},
{
"value": "jrmadsen",
"key": "USER"
},
{
"value": "true",
"key": "... etc. ..."
}
],
"settings": {
"OMNITRACE_JSON_OUTPUT": {
"count": -1,
"environ_updated": false,
"name": "json_output",
"data_type": "bool",
"initial": true,
"enabled": true,
"value": true,
"max_count": 1,
"cmdline": [
"--omnitrace-json-output"
],
"environ": "OMNITRACE_JSON_OUTPUT",
"config_updated": false,
"categories": [
"io",
"json",
"native"
],
"description": "Write json output files"
},
"... etc. ...": {
"etc.": true
}
}
}
}
}
```
## Configuring Output
### Core Configuration Settings
> ***See also: [Customizing OmniTrace Runtime](runtime.md)***
| Setting | Value | Description |
|---------------------------|--------------------|---------------------------------------------------------------------------------------------------|
| `OMNITRACE_OUTPUT_PATH` | Any valid path | Path to folder where output files should be placed |
| `OMNITRACE_OUTPUT_PREFIX` | String | Useful for multiple runs with different arguments. See [Output Prefix Keys](#output-prefix-keys) |
| `OMNITRACE_OUTPUT_FILE` | Any valid filepath | Specific location for perfetto output file. |
| `OMNITRACE_TIME_OUTPUT` | Boolean | Place all output in a timestamped folder, timestamp format controlled via `OMNITRACE_TIME_FORMAT` |
| `OMNITRACE_TIME_FORMAT` | String | See `strftime` man pages for valid identifiers |
| `OMNITRACE_USE_PID` | Boolean | Append either the PID or the MPI rank to all output files (before the extension) |
#### Output Prefix Keys
Output prefix keys have many uses but most useful when dealing with multiple profiling runs or large MPI jobs.
Their inclusion in omnitrace stems from their introduction into timemory for [compile-time-perf](https://github.com/jrmadsen/compile-time-perf)
which needed to be able to create different output files for a generic wrapper around compilation commands while still
overwriting the output from the last time a file was compiled.
If you are ever doing scaling studies and specifying options via the command line, it is highly recommend to just
use a common `OMNITRACE_OUTPUT_PATH`, disable `OMNITRACE_TIME_OUTPUT`,
set `OMNITRACE_OUTPUT_PREFIX="%argt%-"` and let omnitrace cleanly organize the output.
| String | Encoding |
|-----------------|--------------------------------------------------------------------------------------------------------------------|
| `%argv%` | Entire command-line condensed into a single string |
| `%argt%` | Similar to `%argv%` except basename of first command line argument |
| `%args%` | All command line arguments condensed into a single string |
| `%tag%` | Basename of first command line argument |
| `%arg<N>%` | Command line argument at position `<N>` (zero indexed), e.g. `%arg0%` for first argument. |
| `%argv_hash%` | MD5 sum of `%argv%` |
| `%argt_hash%` | MD5 sum if `%argt%` |
| `%args_hash%` | MD5 sum of `%args%` |
| `%tag_hash%` | MD5 sum of `%tag%` |
| `%arg<N>_hash%` | MD5 sum of `%arg<N>%` |
| `%pid%` | Process identifier (i.e. `getpid()`) |
| `%ppid%` | Parent process identifier (i.e. `getppid()`) |
| `%pgid%` | Process group identifier (i.e. `getpgid(getpid())`) |
| `%psid%` | Process session identifier (i.e. `getsid(getpid())`) |
| `%psize%` | Number of sibling process (from reading `/proc/<PPID>/tasks/<PPID>/children`) |
| `%job%` | Value of `SLURM_JOB_ID` environment variable if exists, else `0` |
| `%rank%` | Value of `SLURM_PROCID` environment variable if exists, else `MPI_Comm_rank` (or `0` non-mpi) |
| `%size%` | `MPI_Comm_size` or `1` if non-mpi |
| `%nid%` | `%rank%` if possible, otherwise `%pid%` |
| `%launch_time%` | Launch date and time (uses `OMNITRACE_TIME_FORMAT`) |
| `%env{NAME}%` | Value of environment variable `NAME` (i.e. `getenv(NAME)`) |
| `%cfg{NAME}%` | Value of configuration variable `NAME` (e.g. `%cfg{OMNITRACE_SAMPLING_FREQ}%` would resolve to sampling frequency) |
| `$env{NAME}` | Alternative syntax to `%env{NAME}%` |
| `$cfg{NAME}` | Alternative syntax to `%cfg{NAME}%` |
| `%m` | Shorthand for `%argt_hash%` |
| `%p` | Shorthand for `%pid%` |
| `%j` | Shorthand for `%job%` |
| `%r` | Shorthand for `%rank%` |
| `%s` | Shorthand for `%size%` |
> ***Any output prefix key which contain a `/` will have the `/` characters***
> ***replaced with `_` and any leading underscores will be stripped, e.g. if `%arg0%` is `/usr/bin/foo`, this***
> ***will translate to `usr_bin_foo`. Additionally, any `%arg<N>%` keys which do not have a command line argument***
> ***at position `<N>` will be ignored.***
## Perfetto Output
Use the `OMNITRACE_OUTPUT_FILE` to specify a specific location. If this is an absolute path, then all `OMNITRACE_OUTPUT_PATH`, etc.
settings will be ignored. Visit [ui.perfetto.dev](https://ui.perfetto.dev) and open this file.
![omnitrace-perfetto](images/omnitrace-perfetto.png)
![omnitrace-rocm](images/omnitrace-rocm.png)
![omnitrace-rocm-flow](images/omnitrace-rocm-flow.png)
![omnitrace-user-api](images/omnitrace-user-api.png)
## Timemory Output
Use `omnitrace-avail --components --filename` to view the base filename for each component. E.g.
```shell
$ ./omnitrace-avail wall_clock -C -f
|---------------------------------|---------------|------------------------|
| COMPONENT | AVAILABLE | FILENAME |
|---------------------------------|---------------|------------------------|
| wall_clock | true | wall_clock |
| sampling_wall_clock | true | sampling_wall_clock |
|---------------------------------|---------------|------------------------|
```
Setting `OMNITRACE_COLLAPSE_THREADS=ON` and/or `OMNITRACE_COLLAPSE_PROCESSES=ON` (only valid with full MPI support) the timemory output
will combine the per-thread and/or per-rank data which have identical call-stacks.
The `OMNITRACE_FLAT_PROFILE` setting will remove all call stack heirarchy. Using `OMNITRACE_FLAT_PROFILE=ON` in combination
with `OMNITRACE_COLLAPSE_THREADS=ON` is a useful configuration for identifying min/max measurements regardless of calling context.
The `OMNITRACE_TIMELINE_PROFILE` setting (with `OMNITRACE_FLAT_PROFILE=OFF`) will effectively generate similar data that can be found
in perfetto. Enabling timeline and flat profiling will effectively generate similar data to `strace`. However, while timemory in general
requires significantly less memory than perfetto, this is not the case in timeline mode so activate this setting with caution.
### Timemory Text Output
> ***Hint: the generation of text output is configurable via `OMNITRACE_TEXT_OUTPUT`***
Timemory text output files are meant for human-consumption (use JSON formats for analysis)
and as such, some fields such as the `LABEL` fields may be truncated for readability.
Modification of the truncation can be changed via the `OMNITRACE_MAX_WIDTH` setting.
#### Timemory Text Output Example
In the below, the `NN` field in `|NN>>>` is the thread ID. If MPI support is enabled, this will be `|MM|NN>>>` and `MM` will be the rank.
If `OMNITRACE_COLLAPSE_THREADS=ON` and `OMNITRACE_COLLAPSE_PROCESSES=ON`, neither the `MM` nor the `NN` will be present unless the
component explicitly sets type-traits which specify that the data is only relevant per-thread or per-process, e.g. the `thread_cpu_clock` clock component.
```console
|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| REAL-CLOCK TIMER (I.E. WALL-CLOCK TIMER) |
|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| LABEL | COUNT | DEPTH | METRIC | UNITS | SUM | MEAN | MIN | MAX | VAR | STDDEV | % SELF |
|--------------------------------------------------------------|--------|--------|------------|--------|-----------|-----------|-----------|-----------|----------|----------|--------|
| |00>>> main | 1 | 0 | wall_clock | sec | 13.360265 | 13.360265 | 13.360265 | 13.360265 | 0.000000 | 0.000000 | 18.2 |
| |00>>> |_ompt_thread_initial | 1 | 1 | wall_clock | sec | 10.924161 | 10.924161 | 10.924161 | 10.924161 | 0.000000 | 0.000000 | 0.0 |
| |00>>> |_ompt_implicit_task | 1 | 2 | wall_clock | sec | 10.923050 | 10.923050 | 10.923050 | 10.923050 | 0.000000 | 0.000000 | 0.1 |
| |00>>> |_ompt_parallel [parallelism=12] | 1 | 3 | wall_clock | sec | 10.915026 | 10.915026 | 10.915026 | 10.915026 | 0.000000 | 0.000000 | 0.0 |
| |00>>> |_ompt_implicit_task | 1 | 4 | wall_clock | sec | 10.647951 | 10.647951 | 10.647951 | 10.647951 | 0.000000 | 0.000000 | 0.0 |
| |00>>> |_ompt_work_loop | 156 | 5 | wall_clock | sec | 0.000812 | 0.000005 | 0.000001 | 0.000212 | 0.000000 | 0.000018 | 100.0 |
| |00>>> |_ompt_work_single_executor | 40 | 5 | wall_clock | sec | 0.000016 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_sync_region_barrier_implicit | 308 | 5 | wall_clock | sec | 0.000629 | 0.000002 | 0.000001 | 0.000017 | 0.000000 | 0.000002 | 100.0 |
| |00>>> |_conj_grad | 76 | 5 | wall_clock | sec | 10.641165 | 0.140015 | 0.131894 | 0.155099 | 0.000017 | 0.004080 | 1.0 |
| |00>>> |_ompt_work_single_executor | 803 | 6 | wall_clock | sec | 0.000292 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_work_loop | 7904 | 6 | wall_clock | sec | 7.420265 | 0.000939 | 0.000005 | 0.006974 | 0.000003 | 0.001613 | 100.0 |
| |00>>> |_ompt_sync_region_barrier_implicit | 6004 | 6 | wall_clock | sec | 0.283160 | 0.000047 | 0.000001 | 0.004087 | 0.000000 | 0.000303 | 100.0 |
| |00>>> |_ompt_sync_region_barrier_implementation | 3952 | 6 | wall_clock | sec | 2.829252 | 0.000716 | 0.000007 | 0.009005 | 0.000001 | 0.000985 | 99.7 |
| |00>>> |_ompt_sync_region_reduction | 15808 | 7 | wall_clock | sec | 0.009142 | 0.000001 | 0.000000 | 0.000007 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_work_single_other | 1249 | 6 | wall_clock | sec | 0.000270 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_work_single_other | 114 | 5 | wall_clock | sec | 0.000024 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_sync_region_barrier_implementation | 76 | 5 | wall_clock | sec | 0.000876 | 0.000012 | 0.000008 | 0.000025 | 0.000000 | 0.000003 | 84.4 |
| |00>>> |_ompt_sync_region_reduction | 304 | 6 | wall_clock | sec | 0.000136 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_ompt_master | 226 | 5 | wall_clock | sec | 0.001978 | 0.000009 | 0.000000 | 0.000038 | 0.000000 | 0.000012 | 100.0 |
| |11>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.656145 | 10.656145 | 10.656145 | 10.656145 | 0.000000 | 0.000000 | 0.1 |
| |11>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649183 | 10.649183 | 10.649183 | 10.649183 | 0.000000 | 0.000000 | 0.0 |
| |11>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000852 | 0.000005 | 0.000002 | 0.000230 | 0.000000 | 0.000019 | 100.0 |
| |11>>> |_ompt_work_single_other | 149 | 6 | wall_clock | sec | 0.000035 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |11>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004135 | 0.000013 | 0.000001 | 0.001233 | 0.000000 | 0.000070 | 100.0 |
| |11>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641302 | 0.140017 | 0.131896 | 0.155102 | 0.000017 | 0.004080 | 0.6 |
| |11>>> |_ompt_work_single_other | 2023 | 7 | wall_clock | sec | 0.000458 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |11>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 8.253555 | 0.001044 | 0.000005 | 0.008021 | 0.000003 | 0.001790 | 100.0 |
| |11>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.263840 | 0.000044 | 0.000001 | 0.004087 | 0.000000 | 0.000297 | 100.0 |
| |11>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.059823 | 0.000521 | 0.000007 | 0.009508 | 0.000001 | 0.000863 | 100.0 |
| |11>>> |_ompt_work_single_executor | 29 | 7 | wall_clock | sec | 0.000011 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |11>>> |_ompt_work_single_executor | 5 | 6 | wall_clock | sec | 0.000002 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |11>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000975 | 0.000013 | 0.000008 | 0.000024 | 0.000000 | 0.000003 | 100.0 |
| |10>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.681664 | 10.681664 | 10.681664 | 10.681664 | 0.000000 | 0.000000 | 0.3 |
| |10>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649158 | 10.649158 | 10.649158 | 10.649158 | 0.000000 | 0.000000 | 0.0 |
| |10>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000863 | 0.000006 | 0.000002 | 0.000231 | 0.000000 | 0.000019 | 100.0 |
| |10>>> |_ompt_work_single_other | 140 | 6 | wall_clock | sec | 0.000037 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |10>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004149 | 0.000013 | 0.000001 | 0.001221 | 0.000000 | 0.000070 | 100.0 |
| |10>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641288 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.7 |
| |10>>> |_ompt_work_single_other | 1883 | 7 | wall_clock | sec | 0.000487 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |10>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 8.174545 | 0.001034 | 0.000005 | 0.006899 | 0.000003 | 0.001766 | 100.0 |
| |10>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.268808 | 0.000045 | 0.000001 | 0.004087 | 0.000000 | 0.000299 | 100.0 |
| |10>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.126988 | 0.000538 | 0.000007 | 0.009843 | 0.000001 | 0.000872 | 99.9 |
| |10>>> |_ompt_sync_region_reduction | 3952 | 8 | wall_clock | sec | 0.002574 | 0.000001 | 0.000000 | 0.000014 | 0.000000 | 0.000000 | 100.0 |
| |10>>> |_ompt_work_single_executor | 169 | 7 | wall_clock | sec | 0.000072 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |10>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000954 | 0.000013 | 0.000009 | 0.000023 | 0.000000 | 0.000003 | 95.9 |
| |10>>> |_ompt_sync_region_reduction | 76 | 7 | wall_clock | sec | 0.000039 | 0.000001 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |10>>> |_ompt_work_single_executor | 14 | 6 | wall_clock | sec | 0.000006 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |09>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.686552 | 10.686552 | 10.686552 | 10.686552 | 0.000000 | 0.000000 | 0.3 |
| |09>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649151 | 10.649151 | 10.649151 | 10.649151 | 0.000000 | 0.000000 | 0.0 |
| |09>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000880 | 0.000006 | 0.000002 | 0.000258 | 0.000000 | 0.000021 | 100.0 |
| |09>>> |_ompt_work_single_other | 148 | 6 | wall_clock | sec | 0.000034 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |09>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004129 | 0.000013 | 0.000001 | 0.001210 | 0.000000 | 0.000069 | 100.0 |
| |09>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641308 | 0.140017 | 0.131895 | 0.155102 | 0.000017 | 0.004080 | 0.7 |
| |09>>> |_ompt_work_single_other | 2043 | 7 | wall_clock | sec | 0.000473 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |09>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.977001 | 0.001009 | 0.000005 | 0.007325 | 0.000003 | 0.001732 | 100.0 |
| |09>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.242996 | 0.000040 | 0.000001 | 0.004087 | 0.000000 | 0.000284 | 100.0 |
| |09>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.350895 | 0.000595 | 0.000007 | 0.008689 | 0.000001 | 0.000926 | 100.0 |
| |09>>> |_ompt_work_single_executor | 9 | 7 | wall_clock | sec | 0.000004 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |09>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000973 | 0.000013 | 0.000008 | 0.000025 | 0.000000 | 0.000003 | 100.0 |
| |09>>> |_ompt_work_single_executor | 6 | 6 | wall_clock | sec | 0.000002 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.721622 | 10.721622 | 10.721622 | 10.721622 | 0.000000 | 0.000000 | 0.7 |
| |08>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649135 | 10.649135 | 10.649135 | 10.649135 | 0.000000 | 0.000000 | 0.0 |
| |08>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000839 | 0.000005 | 0.000001 | 0.000231 | 0.000000 | 0.000019 | 100.0 |
| |08>>> |_ompt_work_single_other | 141 | 6 | wall_clock | sec | 0.000030 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004114 | 0.000013 | 0.000001 | 0.001198 | 0.000000 | 0.000069 | 100.0 |
| |08>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641294 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.6 |
| |08>>> |_ompt_work_single_other | 1742 | 7 | wall_clock | sec | 0.000392 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 8.306388 | 0.001051 | 0.000005 | 0.007886 | 0.000003 | 0.001795 | 100.0 |
| |08>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.274358 | 0.000046 | 0.000001 | 0.004090 | 0.000000 | 0.000302 | 100.0 |
| |08>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 1.991251 | 0.000504 | 0.000007 | 0.008694 | 0.000001 | 0.000844 | 99.8 |
| |08>>> |_ompt_sync_region_reduction | 7904 | 8 | wall_clock | sec | 0.003816 | 0.000000 | 0.000000 | 0.000017 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_work_single_executor | 310 | 7 | wall_clock | sec | 0.000112 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000955 | 0.000013 | 0.000009 | 0.000026 | 0.000000 | 0.000003 | 93.7 |
| |08>>> |_ompt_sync_region_reduction | 152 | 7 | wall_clock | sec | 0.000060 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |08>>> |_ompt_work_single_executor | 13 | 6 | wall_clock | sec | 0.000005 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |07>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.747282 | 10.747282 | 10.747282 | 10.747282 | 0.000000 | 0.000000 | 0.9 |
| |07>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649093 | 10.649093 | 10.649093 | 10.649093 | 0.000000 | 0.000000 | 0.0 |
| |07>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000923 | 0.000006 | 0.000002 | 0.000231 | 0.000000 | 0.000019 | 100.0 |
| |07>>> |_ompt_work_single_other | 152 | 6 | wall_clock | sec | 0.000048 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |07>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.003981 | 0.000013 | 0.000001 | 0.001186 | 0.000000 | 0.000068 | 100.0 |
| |07>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641295 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.7 |
| |07>>> |_ompt_work_single_other | 2043 | 7 | wall_clock | sec | 0.000648 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |07>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.978811 | 0.001009 | 0.000005 | 0.006728 | 0.000003 | 0.001732 | 100.0 |
| |07>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.199939 | 0.000033 | 0.000001 | 0.004086 | 0.000000 | 0.000255 | 100.0 |
| |07>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.385843 | 0.000604 | 0.000009 | 0.009039 | 0.000001 | 0.000938 | 100.0 |
| |07>>> |_ompt_work_single_executor | 9 | 7 | wall_clock | sec | 0.000004 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |07>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000905 | 0.000012 | 0.000010 | 0.000025 | 0.000000 | 0.000003 | 100.0 |
| |07>>> |_ompt_work_single_executor | 2 | 6 | wall_clock | sec | 0.000001 | 0.000001 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |06>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.772278 | 10.772278 | 10.772278 | 10.772278 | 0.000000 | 0.000000 | 1.1 |
| |06>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649092 | 10.649092 | 10.649092 | 10.649092 | 0.000000 | 0.000000 | 0.0 |
| |06>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000888 | 0.000006 | 0.000002 | 0.000236 | 0.000000 | 0.000020 | 100.0 |
| |06>>> |_ompt_work_single_other | 153 | 6 | wall_clock | sec | 0.000037 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |06>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004090 | 0.000013 | 0.000001 | 0.001175 | 0.000000 | 0.000067 | 100.0 |
| |06>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641317 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.8 |
| |06>>> |_ompt_work_single_other | 2041 | 7 | wall_clock | sec | 0.000476 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |06>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.467961 | 0.000945 | 0.000005 | 0.010712 | 0.000003 | 0.001627 | 100.0 |
| |06>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.250883 | 0.000042 | 0.000001 | 0.004087 | 0.000000 | 0.000285 | 100.0 |
| |06>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.838733 | 0.000718 | 0.000009 | 0.009015 | 0.000001 | 0.001015 | 99.9 |
| |06>>> |_ompt_sync_region_reduction | 3952 | 8 | wall_clock | sec | 0.003334 | 0.000001 | 0.000000 | 0.000025 | 0.000000 | 0.000001 | 100.0 |
| |06>>> |_ompt_work_single_executor | 11 | 7 | wall_clock | sec | 0.000005 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |06>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000940 | 0.000012 | 0.000009 | 0.000025 | 0.000000 | 0.000003 | 95.4 |
| |06>>> |_ompt_sync_region_reduction | 76 | 7 | wall_clock | sec | 0.000044 | 0.000001 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |06>>> |_ompt_work_single_executor | 1 | 6 | wall_clock | sec | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |05>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.797950 | 10.797950 | 10.797950 | 10.797950 | 0.000000 | 0.000000 | 1.4 |
| |05>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649072 | 10.649072 | 10.649072 | 10.649072 | 0.000000 | 0.000000 | 0.0 |
| |05>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000879 | 0.000006 | 0.000001 | 0.000248 | 0.000000 | 0.000021 | 100.0 |
| |05>>> |_ompt_work_single_other | 142 | 6 | wall_clock | sec | 0.000034 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |05>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004062 | 0.000013 | 0.000002 | 0.001163 | 0.000000 | 0.000067 | 100.0 |
| |05>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641291 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.7 |
| |05>>> |_ompt_work_single_other | 2038 | 7 | wall_clock | sec | 0.000500 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |05>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 8.279191 | 0.001047 | 0.000005 | 0.006596 | 0.000003 | 0.001792 | 100.0 |
| |05>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.250939 | 0.000042 | 0.000001 | 0.004090 | 0.000000 | 0.000286 | 100.0 |
| |05>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.039013 | 0.000516 | 0.000009 | 0.008689 | 0.000001 | 0.000855 | 100.0 |
| |05>>> |_ompt_work_single_executor | 14 | 7 | wall_clock | sec | 0.000005 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |05>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000926 | 0.000012 | 0.000009 | 0.000023 | 0.000000 | 0.000003 | 100.0 |
| |05>>> |_ompt_work_single_executor | 12 | 6 | wall_clock | sec | 0.000005 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.825935 | 10.825935 | 10.825935 | 10.825935 | 0.000000 | 0.000000 | 1.6 |
| |04>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649068 | 10.649068 | 10.649068 | 10.649068 | 0.000000 | 0.000000 | 0.0 |
| |04>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000884 | 0.000006 | 0.000002 | 0.000245 | 0.000000 | 0.000020 | 100.0 |
| |04>>> |_ompt_work_single_other | 150 | 6 | wall_clock | sec | 0.000034 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004069 | 0.000013 | 0.000001 | 0.001151 | 0.000000 | 0.000066 | 100.0 |
| |04>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641300 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 1.1 |
| |04>>> |_ompt_work_single_other | 2041 | 7 | wall_clock | sec | 0.000448 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.438393 | 0.000941 | 0.000005 | 0.007090 | 0.000003 | 0.001624 | 100.0 |
| |04>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.270654 | 0.000045 | 0.000001 | 0.004090 | 0.000000 | 0.000295 | 100.0 |
| |04>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.819165 | 0.000713 | 0.000009 | 0.008379 | 0.000001 | 0.001013 | 99.9 |
| |04>>> |_ompt_sync_region_reduction | 7904 | 8 | wall_clock | sec | 0.003932 | 0.000000 | 0.000000 | 0.000015 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_work_single_executor | 11 | 7 | wall_clock | sec | 0.000005 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000936 | 0.000012 | 0.000009 | 0.000025 | 0.000000 | 0.000003 | 93.2 |
| |04>>> |_ompt_sync_region_reduction | 152 | 7 | wall_clock | sec | 0.000064 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |04>>> |_ompt_work_single_executor | 4 | 6 | wall_clock | sec | 0.000001 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |03>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.849322 | 10.849322 | 10.849322 | 10.849322 | 0.000000 | 0.000000 | 1.8 |
| |03>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649075 | 10.649075 | 10.649075 | 10.649075 | 0.000000 | 0.000000 | 0.0 |
| |03>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000861 | 0.000006 | 0.000002 | 0.000238 | 0.000000 | 0.000020 | 100.0 |
| |03>>> |_ompt_work_single_other | 120 | 6 | wall_clock | sec | 0.000028 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |03>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.003993 | 0.000013 | 0.000001 | 0.001138 | 0.000000 | 0.000065 | 100.0 |
| |03>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641302 | 0.140017 | 0.131896 | 0.155101 | 0.000017 | 0.004080 | 0.8 |
| |03>>> |_ompt_work_single_other | 1756 | 7 | wall_clock | sec | 0.000426 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |03>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 8.005617 | 0.001013 | 0.000005 | 0.011500 | 0.000003 | 0.001741 | 100.0 |
| |03>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.231485 | 0.000039 | 0.000001 | 0.004086 | 0.000000 | 0.000277 | 100.0 |
| |03>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.320428 | 0.000587 | 0.000009 | 0.010868 | 0.000001 | 0.000912 | 100.0 |
| |03>>> |_ompt_work_single_executor | 296 | 7 | wall_clock | sec | 0.000120 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |03>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000967 | 0.000013 | 0.000010 | 0.000023 | 0.000000 | 0.000003 | 100.0 |
| |03>>> |_ompt_work_single_executor | 34 | 6 | wall_clock | sec | 0.000013 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.876387 | 10.876387 | 10.876387 | 10.876387 | 0.000000 | 0.000000 | 2.1 |
| |02>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649050 | 10.649050 | 10.649050 | 10.649050 | 0.000000 | 0.000000 | 0.0 |
| |02>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000924 | 0.000006 | 0.000001 | 0.000241 | 0.000000 | 0.000020 | 100.0 |
| |02>>> |_ompt_work_single_other | 139 | 6 | wall_clock | sec | 0.000040 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.003972 | 0.000013 | 0.000001 | 0.001127 | 0.000000 | 0.000064 | 100.0 |
| |02>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641287 | 0.140017 | 0.131895 | 0.155101 | 0.000017 | 0.004080 | 0.7 |
| |02>>> |_ompt_work_single_other | 1902 | 7 | wall_clock | sec | 0.000553 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.906688 | 0.001000 | 0.000005 | 0.007068 | 0.000003 | 0.001713 | 100.0 |
| |02>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.261367 | 0.000044 | 0.000001 | 0.004088 | 0.000000 | 0.000295 | 100.0 |
| |02>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.402362 | 0.000608 | 0.000009 | 0.010399 | 0.000001 | 0.000944 | 99.9 |
| |02>>> |_ompt_sync_region_reduction | 3952 | 8 | wall_clock | sec | 0.002937 | 0.000001 | 0.000000 | 0.000021 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_work_single_executor | 150 | 7 | wall_clock | sec | 0.000073 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000895 | 0.000012 | 0.000009 | 0.000026 | 0.000000 | 0.000003 | 95.2 |
| |02>>> |_ompt_sync_region_reduction | 76 | 7 | wall_clock | sec | 0.000043 | 0.000001 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |02>>> |_ompt_work_single_executor | 15 | 6 | wall_clock | sec | 0.000007 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |01>>> |_ompt_thread_worker | 1 | 4 | wall_clock | sec | 10.901650 | 10.901650 | 10.901650 | 10.901650 | 0.000000 | 0.000000 | 2.3 |
| |01>>> |_ompt_implicit_task | 1 | 5 | wall_clock | sec | 10.649017 | 10.649017 | 10.649017 | 10.649017 | 0.000000 | 0.000000 | 0.0 |
| |01>>> |_ompt_work_loop | 156 | 6 | wall_clock | sec | 0.000863 | 0.000006 | 0.000001 | 0.000231 | 0.000000 | 0.000019 | 100.0 |
| |01>>> |_ompt_work_single_other | 146 | 6 | wall_clock | sec | 0.000033 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 100.0 |
| |01>>> |_ompt_sync_region_barrier_implicit | 308 | 6 | wall_clock | sec | 0.004012 | 0.000013 | 0.000001 | 0.001115 | 0.000000 | 0.000064 | 100.0 |
| |01>>> |_conj_grad | 76 | 6 | wall_clock | sec | 10.641316 | 0.140017 | 0.131895 | 0.155101 | 0.000017 | 0.004080 | 0.8 |
| |01>>> |_ompt_work_single_other | 1811 | 7 | wall_clock | sec | 0.000403 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |01>>> |_ompt_work_loop | 7904 | 7 | wall_clock | sec | 7.410337 | 0.000938 | 0.000005 | 0.010556 | 0.000003 | 0.001610 | 100.0 |
| |01>>> |_ompt_sync_region_barrier_implicit | 6004 | 7 | wall_clock | sec | 0.202494 | 0.000034 | 0.000001 | 0.003521 | 0.000000 | 0.000256 | 100.0 |
| |01>>> |_ompt_sync_region_barrier_implementation | 3952 | 7 | wall_clock | sec | 2.943604 | 0.000745 | 0.000008 | 0.009033 | 0.000001 | 0.001024 | 100.0 |
| |01>>> |_ompt_work_single_executor | 241 | 7 | wall_clock | sec | 0.000093 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |01>>> |_ompt_sync_region_barrier_implementation | 76 | 6 | wall_clock | sec | 0.000917 | 0.000012 | 0.000009 | 0.000026 | 0.000000 | 0.000003 | 100.0 |
| |01>>> |_ompt_work_single_executor | 8 | 6 | wall_clock | sec | 0.000004 | 0.000000 | 0.000000 | 0.000001 | 0.000000 | 0.000000 | 100.0 |
| |00>>> |_c_print_results | 1 | 2 | wall_clock | sec | 0.000049 | 0.000049 | 0.000049 | 0.000049 | 0.000000 | 0.000000 | 100.0 |
|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
```
### Timemory JSON Output
> ***Hint: the generation of flat JSON output is configurable via `OMNITRACE_JSON_OUTPUT`.***
> ***The generation of hierarchical JSON data is configurable via `OMNITRACE_TREE_OUTPUT`.***
Timemory represents the data within the JSON output in two forms: a flat structure and a hierarchical structure.
The flat JSON data represents the data similar to the text files: the hierarchical information
is represented by the indentation of the `"prefix"` field and the `"depth"` field.
The hierarchical JSON contains additional information with respect to inclusive and exclusive value, however,
it's structure requires processing through recursion. This section of the JSON supports analysis
by [hatchet](https://github.com/hatchet/hatchet).
All the data entries for the flat structure are in a single JSON array.
This format is easier than the hierarchical format to write a simple Python script for post-processing.
#### Timemory JSON Output Sample
In the JSON below, the flat data starts at `["timemory"]["wall_clock"]["ranks"]`
and the hierarchical data starts at `["timemory"]["wall_clock"]["graph"]`.
E.g., accessing the name (prefix) of the nth entry in the flat data layout is:
`["timemory"]["wall_clock"]["ranks"][0]["graph"][<N>]["prefix"]`. When full MPI
support is enable, the per-rank data in flat layout will be represented
in as an entry in the "ranks" array; in the hierarchical data structure,
the per-rank data is represented as entry in the "mpi" array (but "graph"
is used in lieu of "mpi" when full MPI support is enabled).
In the hierarchical layout, all data for the process is all a child of a (dummy)
root node (which has the name `unknown-hash=0`).
```json
{
"timemory": {
"wall_clock": {
"properties": {
"cereal_class_version": 0,
"value": 78,
"enum": "WALL_CLOCK",
"id": "wall_clock",
"ids": [
"real_clock",
"virtual_clock",
"wall_clock"
]
},
"type": "wall_clock",
"description": "Real-clock timer (i.e. wall-clock timer)",
"unit_value": 1000000000,
"unit_repr": "sec",
"thread_scope_only": false,
"thread_count": 2,
"mpi_size": 1,
"upcxx_size": 1,
"process_count": 1,
"num_ranks": 1,
"concurrency": 2,
"ranks": [
{
"rank": 0,
"graph_size": 112,
"graph": [
{
"hash": 17481650134347108265,
"prefix": "|0>>> main",
"depth": 0,
"entry": {
"cereal_class_version": 0,
"laps": 1,
"value": 894743517,
"accum": 894743517,
"repr_data": 0.894743517,
"repr_display": 0.894743517
},
"stats": {
"cereal_class_version": 0,
"sum": 0.894743517,
"count": 1,
"min": 0.894743517,
"max": 0.894743517,
"sqr": 0.8005659612135293,
"mean": 0.894743517,
"stddev": 0.0
},
"rolling_hash": 17481650134347108265
},
{
"hash": 3455444288293231339,
"prefix": "|0>>> |_read_input",
"depth": 1,
"entry": {
"laps": 1,
"value": 9808,
"accum": 9808,
"repr_data": 9.808e-06,
"repr_display": 9.808e-06
},
"stats": {
"sum": 9.808e-06,
"count": 1,
"min": 9.808e-06,
"max": 9.808e-06,
"sqr": 9.6196864e-11,
"mean": 9.808e-06,
"stddev": 0.0
},
"rolling_hash": 2490350348930787988
},
{
"hash": 8456966793631718807,
"prefix": "|0>>> |_setcoeff",
"depth": 1,
"entry": {
"laps": 1,
"value": 922,
"accum": 922,
"repr_data": 9.22e-07,
"repr_display": 9.22e-07
},
"stats": {
"sum": 9.22e-07,
"count": 1,
"min": 9.22e-07,
"max": 9.22e-07,
"sqr": 8.50084e-13,
"mean": 9.22e-07,
"stddev": 0.0
},
"rolling_hash": 7491872854269275456
},
{
"hash": 6107876127803219007,
"prefix": "|0>>> |_ompt_thread_initial",
"depth": 1,
"entry": {
"laps": 1,
"value": 896506392,
"accum": 896506392,
"repr_data": 0.896506392,
"repr_display": 0.896506392
},
"stats": {
"sum": 0.896506392,
"count": 1,
"min": 0.896506392,
"max": 0.896506392,
"sqr": 0.8037237108968578,
"mean": 0.896506392,
"stddev": 0.0
},
"rolling_hash": 5142782188440775656
},
{
"hash": 15402802091993617561,
"prefix": "|0>>> |_ompt_implicit_task",
"depth": 2,
"entry": {
"laps": 1,
"value": 896479111,
"accum": 896479111,
"repr_data": 0.896479111,
"repr_display": 0.896479111
},
"stats": {
"sum": 0.896479111,
"count": 1,
"min": 0.896479111,
"max": 0.896479111,
"sqr": 0.8036747964593504,
"mean": 0.896479111,
"stddev": 0.0
},
"rolling_hash": 2098840206724841601 },
{
"..." : "... etc. ..."
}
]
}
],
"graph": [
[
{
"cereal_class_version": 0,
"node": {
"hash": 0,
"prefix": "unknown-hash=0",
"tid": [
0
],
"pid": [
2539175
],
"depth": 0,
"is_dummy": false,
"inclusive": {
"entry": {
"laps": 0,
"value": 0,
"accum": 0,
"repr_data": 0.0,
"repr_display": 0.0
},
"stats": {
"sum": 0.0,
"count": 0,
"min": 0.0,
"max": 0.0,
"sqr": 0.0,
"mean": 0.0,
"stddev": 0.0
}
},
"exclusive": {
"entry": {
"laps": 0,
"value": -894743517,
"accum": -894743517,
"repr_data": -0.894743517,
"repr_display": -0.894743517
},
"stats": {
"sum": 0.0,
"count": 0,
"min": 0.0,
"max": 0.0,
"sqr": 0.0,
"mean": 0.0,
"stddev": 0.0
}
}
},
"children": [
{
"node": {
"hash": 17481650134347108265,
"prefix": "main",
"tid": [
0
],
"pid": [
2539175
],
"depth": 1,
"is_dummy": false,
"inclusive": {
"entry": {
"laps": 1,
"value": 894743517,
"accum": 894743517,
"repr_data": 0.894743517,
"repr_display": 0.894743517
},
"stats": {
"sum": 0.894743517,
"count": 1,
"min": 0.894743517,
"max": 0.894743517,
"sqr": 0.8005659612135293,
"mean": 0.894743517,
"stddev": 0.0
}
},
"exclusive": {
"entry": {
"laps": 1,
"value": -1773605,
"accum": -1773605,
"repr_data": -0.001773605,
"repr_display": -0.001773605
},
"stats": {
"sum": -0.001773605,
"count": 1,
"min": 9.22e-07,
"max": 0.896506392,
"sqr": -0.0031577497803754,
"mean": -0.001773605,
"stddev": 0.0
}
}
},
"children": [
{
"..." : "... etc. ..."
}
]
}
]
}
]
]
}
}
}
```
#### Timemory JSON Output Python Post-Processing Example
```python
#!/usr/bin/env python3
import sys
import json
def read_json(inp):
with open(inp, "r") as f:
return json.load(f)
def find_max(data):
"""Find the max for any function called multiple times"""
max_entry = None
for itr in data:
if itr["entry"]["laps"] == 1:
continue
if max_entry is None:
max_entry = itr
else:
if itr["stats"]["mean"] > max_entry["stats"]["mean"]:
max_entry = itr
return max_entry
def strip_name(name):
"""Return everything after |_ if it exists"""
idx = name.index("|_")
return name if idx is None else name[(idx + 2) :]
if __name__ == "__main__":
input_data = [[x, read_json(x)] for x in sys.argv[1:]]
for file, data in input_data:
for metric, metric_data in data["timemory"].items():
print(f"[{file}] Found metric: {metric}")
for n, itr in enumerate(metric_data["ranks"]):
max_entry = find_max(itr["graph"])
print(
"[{}] Maximum value: '{}' at depth {} was called {}x :: {:.3f} {} (mean = {:.3e} {})".format(
file,
strip_name(max_entry["prefix"]),
max_entry["depth"],
max_entry["entry"]["laps"],
max_entry["entry"]["repr_data"],
metric_data["unit_repr"],
max_entry["stats"]["mean"],
metric_data["unit_repr"],
)
)
```
This script applied to the corresponding JSON output from [Text Output Example](#timemory-text-output-example) would be:
```console
[openmp-cg.inst-wall_clock.json] Found metric: wall_clock
[openmp-cg.inst-wall_clock.json] Maximum value: 'conj_grad' at depth 6 was called 76x :: 10.641 sec (mean = 1.400e-01 sec)
```
Référencer dans un nouveau ticket Voir Git Blâme Copier le lien permanent