AI/rocm-systems
2
0
Fork 0
Code Issues Pull-Requests Actions Pakete Projekte Releases Wiki Aktivität
Dateien
f82845388aab108ed1d1fc404f433a0def391bb3
rocm-systems/source/docs/features.md
T
Jonathan R. Madsen a142b2029d Fix the main library stop routine for timemory (#39) 
* Fix the main library stop routine for timemory

- the main pop_timemory function was popping too many calls
- this primarily affected recursive calls

* Lengthen the timeout for the Configure CMake step

* Fix python tests

- new validate-timemory-json.py script

* Documentation update

- Call-counts in timemory output examples in documentation were affected by the changes

* Fix the per-thread metrics during finalization

- pthread_create_mutex starts/stops the per-thread data
- removed unintentional continue statement

* Docs tweaks

* Fix lap counter on per-thread metrics
2022-06-13 15:57:44 -05:00

1.6 KiB
Originalformat Blame Verlauf

Features

.. toctree::
   :glob:
   :maxdepth: 4

Overview

Omnitrace is designed to be highly extensible. Internally, it leverages the timemory performance analysis toolkit to manage extensions, resources, data, etc.

Data Collection Modes

  • Dynamic instrumentation
    • Runtime instrumentation
      • Instrument executable and shared libraries at runtime
    • Binary rewriting
      • Generate a new executable and/or library with instrumentation built-in
  • Statistical sampling
    • Periodic software interrupts per-thread
  • Background thread sampling
    • Record process and system-level values while an application executes
  • Critical trace generation

Data Analysis

  • Critical trace generation (beta)
  • Support for

Parallelism API Support

  • Built-in MPI support
  • Kokkos-Tools support
  • OpenMP-Tools (OMPT)

GPU Metrics

  • HIP API tracing
  • ROCM HSA API tracing
  • Kernel runtime tracing
  • System-level sampling (via rocm-smi)
    • Memory usage
    • Power usage
    • Temperature
    • Utilization

CPU Metrics

  • CPU hardware counters sampling and profiles
  • CPU frequency sampling
  • Various timing metrics
    • Wall time
    • CPU time (process and/or thread)
    • CPU utilization (process and/or thread)
    • User CPU time
    • Kernel CPU time
  • Various memory metrics
    • High-water mark (sampling and profiles)
    • Memory page allocation
    • Virtual memory usage
  • Network statistics
  • I/O metrics
  • ... many more

Third-party API support

  • TAU
  • LIKWID
  • Caliper
  • CrayPAT
  • VTune
  • NVTX
  • ROCTX
In neuem Issue referenzieren Git Blame ansehen Permalink kopieren