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
387 lines
11 KiB
ReStructuredText
387 lines
11 KiB
ReStructuredText
.. meta::
|
||
:description: ROCm Compute Profiler performance model: Pipeline metrics
|
||
:keywords: Omniperf, ROCm Compute Profiler, ROCm, profiler, tool, Instinct, accelerator, pipeline, wavefront, metrics, launch, runtime
|
||
VALU, MFMA, instruction mix, FLOPs, arithmetic, operations
|
||
|
||
****************
|
||
Pipeline metrics
|
||
****************
|
||
|
||
This section details the metrics available in ROCm Compute Profiler to analyze the
|
||
pipelines discussed in the :doc:`pipeline-descriptions`.
|
||
|
||
.. _wavefront:
|
||
|
||
Wavefront
|
||
=========
|
||
|
||
.. _wavefront-launch-stats:
|
||
|
||
Wavefront launch stats
|
||
----------------------
|
||
|
||
The wavefront launch stats panel gives general information about the
|
||
kernel launch:
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA
|
||
|
||
.. jinja:: wavefront-launch-stats-gfx908
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: wavefront-launch-stats-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: wavefront-launch-stats-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: wavefront-launch-stats-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. _wavefront-runtime-stats:
|
||
|
||
Wavefront runtime stats
|
||
-----------------------
|
||
|
||
The wavefront runtime statistics gives a high-level overview of the
|
||
execution of wavefronts in a kernel:
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA
|
||
|
||
.. jinja:: wavefront-runtime-stats-gfx908
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: wavefront-runtime-stats-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: wavefront-runtime-stats-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: wavefront-runtime-stats-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. note::
|
||
|
||
The measurement of kernel cycles and time typically
|
||
cannot be directly compared to other cycles, such as, wave cycles. This is due to two factors:
|
||
* The kernel cycles/timings are measured using a counter that is
|
||
impacted by scheduling overhead, this is particularly noticeable for
|
||
short-running kernels (less than 1ms) where scheduling overhead forms a
|
||
significant portion of the overall kernel runtime.
|
||
* The wave cycles metric is incremented per-wavefront scheduled to a SIMD every cycle whereas
|
||
the kernel cycles counter is incremented only once per-cycle when any
|
||
wavefront is scheduled.
|
||
|
||
.. _instruction-mix:
|
||
|
||
Instruction mix
|
||
===============
|
||
|
||
The instruction mix panel shows a breakdown of the various types of instructions
|
||
executed by the user’s kernel, and which pipelines on the
|
||
:doc:`CU <compute-unit>` they were executed on. In addition, ROCm Compute Profiler reports
|
||
further information about the breakdown of operation types for the
|
||
:ref:`VALU <desc-valu>`, vector-memory, and :ref:`MFMA <desc-mfma>`
|
||
instructions.
|
||
|
||
.. note::
|
||
|
||
All metrics in this section count the instructions issued, and not the total
|
||
number of operations executed. The values reported by these metrics will not
|
||
change regardless of the execution mask of the wavefront. Note that even if
|
||
the execution mask is identically zero (meaning no lanes are active)
|
||
the instruction will still be counted, as CDNA architecture-based GPUs still consider
|
||
these as instructions issued. See
|
||
:mi200-isa-pdf:`EXECute Mask, section 3.3 of the CDNA2 ISA guide<19>` for
|
||
examples and further details.
|
||
|
||
Overall instruction mix
|
||
-----------------------
|
||
|
||
This panel shows the total number of each type of instruction issued to
|
||
the :doc:`various compute pipelines <pipeline-descriptions>` on the
|
||
:doc:`CU <compute-unit>`. These are:
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA
|
||
|
||
.. jinja:: instruction-mix-gfx908
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: instruction-mix-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: instruction-mix-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: instruction-mix-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. note::
|
||
|
||
Branch operations are not used for execution mask updates, but only for whole
|
||
wavefront control flow changes. For more details, see :ref:`desc-branch`.
|
||
|
||
.. _valu-arith-instruction-mix:
|
||
|
||
VALU arithmetic instruction mix
|
||
-------------------------------
|
||
|
||
.. warning::
|
||
|
||
Not all metrics in this section (such as, the floating-point instruction
|
||
breakdowns) are available on CDNA architecture-based GPUs older than the
|
||
:ref:`MI2XX <mixxx-note>` series.
|
||
|
||
This panel details the various types of vector instructions that were
|
||
issued to the :ref:`VALU <desc-valu>`. The metrics in this section do *not*
|
||
include :ref:`MFMA <desc-mfma>` instructions using the same precision; for
|
||
instance, the “F16-ADD” metric does not include any 16-bit floating point
|
||
additions executed as part of an MFMA instruction using the same precision.
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: valu-arith-instruction-mix-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: valu-arith-instruction-mix-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: valu-arith-instruction-mix-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
For an example of these counters in action, refer to
|
||
:ref:`valu-arith-instruction-mix-ex`.
|
||
|
||
.. _vmem-instruction-mix:
|
||
|
||
VMEM instruction mix
|
||
--------------------
|
||
|
||
This section breaks down the types of vector memory (VMEM) instructions
|
||
that were issued. Refer to the
|
||
:ref:`Instruction Counts metrics section <ta-instruction-counts>` under address
|
||
processor front end of the vL1D cache for descriptions of these VMEM
|
||
instructions.
|
||
|
||
.. _matrix-instruction-mix:
|
||
|
||
Matrix instruction mix
|
||
----------------------
|
||
|
||
.. warning::
|
||
|
||
The metrics in this section are only available on CDNA2-CDNA4 architecture-based GPUs
|
||
(:ref:`MI2XX <mixxx-note>`) and newer.
|
||
|
||
This section details the types of Matrix Fused Multiply-Add
|
||
(:ref:`MFMA <desc-mfma>`) instructions that were issued. Note that
|
||
MFMA instructions are classified by the type of input data they operate on, and
|
||
*not* the data type the result is accumulated to.
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: matrix-instruction-mix-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: matrix-instruction-mix-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: matrix-instruction-mix-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
Compute pipeline
|
||
================
|
||
|
||
.. _metrics-flop-count:
|
||
|
||
FLOP counting conventions
|
||
-------------------------
|
||
|
||
ROCm Compute Profiler’s conventions for VALU FLOP counting are as follows:
|
||
|
||
* Addition or multiplication: 1 operation
|
||
|
||
* Transcendentals: 1 operation
|
||
|
||
* Fused multiply-add (FMA): 2 operations
|
||
|
||
Integer operations (IOPs) do not use this convention. They are counted
|
||
as a single operation regardless of the instruction type.
|
||
|
||
.. note::
|
||
|
||
Packed operations which operate on multiple operands in the same instruction
|
||
are counted identically to the underlying instruction type. For example, the
|
||
``v_pk_add_f32`` instruction on :ref:`MI2XX <mixxx-note>`, which performs an
|
||
add operation on two pairs of aligned 32-bit floating-point operands is
|
||
counted only as a single addition -- that is, 1 operation.
|
||
|
||
As discussed in the :ref:`instruction-mix` section, the FLOP/IOP
|
||
metrics in this section do not take into account the execution mask of
|
||
the operation, and will report the same value even if the execution mask
|
||
is identically zero.
|
||
|
||
For example, a FMA instruction operating on 32-bit floating-point
|
||
operands (such as ``v_fma_f32`` on a AMD Instinct :ref:`MI2XX <mixxx-note>` Series GPUs)
|
||
would be counted as 128 total FLOPs: 2 operations (due to the
|
||
instruction type) multiplied by 64 operations (because the wavefront is
|
||
composed of 64 work-items).
|
||
|
||
.. _compute-speed-of-light:
|
||
|
||
Compute Speed-of-Light
|
||
----------------------
|
||
|
||
.. warning::
|
||
|
||
The theoretical maximum throughput for some metrics in this section are
|
||
currently computed with the maximum achievable clock frequency, as reported
|
||
by ``rocminfo``, for a GPU. This may not be realistic for all
|
||
workloads.
|
||
|
||
This section reports the number of floating-point and integer operations
|
||
executed on the :ref:`VALU <desc-valu>` and :ref:`MFMA <desc-mfma>` units in
|
||
various precisions. We note that unlike the
|
||
:ref:`VALU instruction mix <valu-arith-instruction-mix>` and
|
||
:ref:`Matrix instruction mix <matrix-instruction-mix>` sections, the metrics here
|
||
are reported as FLOPs and IOPs, that is, the total number of operations
|
||
executed.
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: compute-speed-of-light-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: compute-speed-of-light-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: compute-speed-of-light-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. _pipeline-stats:
|
||
|
||
Pipeline statistics
|
||
-------------------
|
||
|
||
This section reports a number of key performance characteristics of
|
||
various execution units on the :doc:`CU <compute-unit>`. Refer to
|
||
:ref:`ipc-example` for a detailed dive into these metrics, and the
|
||
:ref:`scheduler <desc-scheduler>` the for a high-level overview of execution
|
||
units and instruction issue.
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA
|
||
|
||
.. jinja:: pipeline-stats-gfx908
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: pipeline-stats-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: pipeline-stats-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: pipeline-stats-gfx950
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. note::
|
||
|
||
The branch utilization reported in this section also includes time spent in
|
||
other instruction types (namely: ``s_endpgm``) that are *typically* a very
|
||
small percentage of the overall kernel execution. This complication is
|
||
omitted for simplicity, but may result in small amounts of branch utilization
|
||
(typically less than 1%) for otherwise branch-less kernels.
|
||
|
||
.. _arithmetic-operations:
|
||
|
||
Arithmetic operations
|
||
---------------------
|
||
|
||
This section reports the total number of floating-point and integer
|
||
operations executed in various precisions. Unlike the
|
||
:ref:`compute-speed-of-light` panel, this section reports both
|
||
:ref:`VALU <desc-valu>` and :ref:`MFMA <desc-mfma>` operations of the same precision
|
||
(e.g., F32) in the same metric. Additionally, this panel lets the user
|
||
control how the data is normalized (such as, control the
|
||
:ref:`normalization unit <normalization-units>`), while the speed-of-light panel does
|
||
not. For more details on how operations are counted, see :ref:`FLOP counting convention <metrics-flop-count>`.
|
||
|
||
.. warning::
|
||
|
||
The metrics in this section do not
|
||
take into account the execution mask of the operation, and will report the
|
||
same value even if EXEC is identically zero. For more details, see :ref:`instruction-mix`.
|
||
|
||
.. tab-set::
|
||
|
||
.. tab-item:: CDNA 2
|
||
|
||
.. jinja:: arithmetic-operations-gfx90a
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 3
|
||
|
||
.. jinja:: arithmetic-operations-gfx942
|
||
:file: _templates/metrics_table.j2
|
||
|
||
.. tab-item:: CDNA 4
|
||
:selected:
|
||
|
||
.. jinja:: arithmetic-operations-gfx950
|
||
:file: _templates/metrics_table.j2
|