derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
167 lines
4.1 KiB
ReStructuredText
167 lines
4.1 KiB
ReStructuredText
.. meta::
|
|
:description: ROCm Compute Profiler - RDNA3.5 WGP / roofline metrics
|
|
:keywords: ROCm Compute Profiler, RDNA, gfx115x, WGP, roofline
|
|
|
|
.. _rdna-wgp:
|
|
|
|
=========================
|
|
Workgroup processor (WGP)
|
|
=========================
|
|
|
|
Within each shader engine, Workgroup Processors (WGPs) pair two Compute Units (CUs) that share resources and execute dispatched waves after the Workgroup Manager (SPI) hands off work.
|
|
On RDNA3.5 architecture-based GPUs/APUs, compute kernels are typically tracked with wave32-oriented waves; the gfx115x WGP panels cover occupancy, dispatch, instruction mix, and local caches at that WGP/CU-pair granularity.
|
|
|
|
The sections below list RDNA3.5 (gfx115x) metric descriptions.
|
|
|
|
.. Note::
|
|
The CDNA architecture-based AMD Instinct GPUs use a different execution hierarchy and panel grouping.
|
|
For Instinct-only pipeline metrics (for example, VALU / VMEM / MFMA-style tables), see :doc:`../cdna/cdna-performance-model`-without assuming RDNA WGPs or CUs map directly to those layouts.
|
|
|
|
Roofline
|
|
========
|
|
|
|
Roofline performance rates
|
|
--------------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-roofline-performance-rates-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Roofline plot points
|
|
--------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-roofline-plot-points-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
WGP block metrics
|
|
=================
|
|
|
|
WGP utilization
|
|
---------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wgp-utilization-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Wavefront launch stats
|
|
----------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wavefront-launch-stats-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Wave dispatch
|
|
-------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wave-dispatch-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Wave life
|
|
---------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wave-life-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Wave instruction mix
|
|
--------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wave-instruction-mix-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
.. note::
|
|
|
|
The **Instructions - Dual VALU (VOPD)** row counts ``SQ_INSTS_DUAL_VALU_WAVE32`` —
|
|
the number of VOPD-encoded wave32 instructions issued per normalization unit.
|
|
Each unit corresponds to one dual-issue VALU slot consumed; kernels that use VOPD
|
|
heavily typically show a large value relative to the **Instructions - VALU** row.
|
|
The exact ratio depends on ISA counting semantics for paired operations, so treat
|
|
the two rows as complementary signals rather than a fixed ceiling relationship.
|
|
For peak-rate theory and FLOPs ceilings see :ref:`rdna-dual-issue-valu`.
|
|
|
|
VMEM instruction mix
|
|
--------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-vmem-instruction-mix-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
LDS instruction mix
|
|
-------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-lds-instruction-mix-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
Wait state analysis
|
|
-------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wait-state-analysis-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
WGP instruction cache
|
|
---------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wgp-instruction-cache-gfx115x
|
|
:file: _templates/metrics_table.j2
|
|
|
|
WGP scalar data cache
|
|
---------------------
|
|
|
|
.. tab-set::
|
|
|
|
.. tab-item:: RDNA3.5 (gfx115x)
|
|
:selected:
|
|
|
|
.. jinja:: rdna115x-wgp-scalar-data-cache-gfx115x
|
|
:file: _templates/metrics_table.j2
|