update links to vllm perf validation doc

.azuredevops/components/AMDMIGraphX.yml

@@ -189,9 +189,9 @@ jobs:
         -DMIGRAPHX_ENABLE_C_API_TEST=ON
         ..
   - template: ${{ variables.CI_TEMPLATE_PATH }}/steps/gpu-diagnostics.yml
-  - task: Bash@3
+  - template: ${{ variables.CI_TEMPLATE_PATH }}/steps/test.yml
-    displayName: Build and run MIGraphX tests
+    parameters:
-    inputs:
+      componentName: AMDMIGraphX
-      targetType: inline
+      testExecutable: make
-      workingDirectory: build
+      testParameters: -j$(nproc) check
-      script: make -j$(nproc) check
+      testPublishResults: false

.azuredevops/components/ROCR-Runtime.yml

@@ -107,6 +107,7 @@ jobs:
       runRocminfo: false
   - task: Bash@3
     displayName: Build kfdtest
+    continueOnError: true
     inputs:
       targetType: 'inline'
       workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/kfdtest
@@ -122,6 +123,7 @@ jobs:
       testDir: $(Build.SourcesDirectory)/libhsakmt/tests/kfdtest/scripts
   - task: Bash@3
     displayName: Build rdmatest app
+    continueOnError: true
     inputs:
       targetType: 'inline'
       workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/app
@@ -130,6 +132,7 @@ jobs:
         cmake --build .
   - task: Bash@3
     displayName: Build rdmatest driver
+    continueOnError: true
     inputs:
       targetType: 'inline'
       workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/drv
@@ -139,6 +142,7 @@ jobs:
         RDMA_HEADER_DIR=/usr/src/amdgpu-*/include make all
   - task: Bash@3
     displayName: Install rdmatest driver
+    continueOnError: true
     inputs:
       targetType: 'inline'
       workingDirectory: $(Build.SourcesDirectory)/libhsakmt/tests/rdma/simple/drv
@@ -154,6 +158,7 @@ jobs:
       testPublishResults: false
   - task: Bash@3
     displayName: Build rocrtst
+    continueOnError: true
     inputs:
       targetType: 'inline'
       workingDirectory: $(Build.SourcesDirectory)/rocrtst/suites/test_common

.azuredevops/components/ROCmValidationSuite.yml

@@ -13,6 +13,7 @@ parameters:
     - libyaml-cpp-dev
     - libpci-dev
     - libpci3
+    - libgst-dev
     - libgtest-dev
     - git
 - name: rocmDependencies

.azuredevops/components/Tensile.yml

@@ -105,6 +105,12 @@ jobs:
   - template: ${{ variables.CI_TEMPLATE_PATH }}/steps/checkout.yml
     parameters:
       checkoutRepo: ${{ parameters.checkoutRepo }}
+  - template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-aqlprofile.yml
+    parameters:
+      ${{ if eq(parameters.checkoutRef, '') }}:
+        dependencySource: staging
+      ${{ elseif ne(parameters.checkoutRef, '') }}:
+        dependencySource: tag-builds
   - template: ${{ variables.CI_TEMPLATE_PATH }}/steps/dependencies-rocm.yml
     parameters:
       dependencyList: ${{ parameters.rocmDependencies }}

.azuredevops/components/omnitrace.yml

@@ -40,6 +40,7 @@ parameters:
 - name: rocmDependencies
   type: object
   default:
+    - aomp
     - clr
     - llvm-project
     - rccl

.wordlist.txt

@@ -36,6 +36,7 @@ Bluefield
 Bootloader
 CCD
 CDNA
+CHTML
 CIFAR
 CLI
 CLion
@@ -70,6 +71,7 @@ Concretized
 Conda
 ConnectX
 CuPy
+Dashboarding
 DDR
 DF
 DGEMM
@@ -227,6 +229,7 @@ Mellanox's
 Meta's
 Miniconda
 MirroredStrategy
+Mixtral
 Multicore
 Multithreaded
 MyEnvironment
@@ -294,6 +297,7 @@ PowerShell
 PyPi
 PyTorch
 Qcycles
+Qwen
 RAII
 RAS
 RCCL
@@ -563,6 +567,7 @@ hipfort
 hipify
 hipsolver
 hipsparse
+hlist
 hotspotting
 hpc
 hpp
@@ -586,6 +591,7 @@ intra
 invariants
 invocating
 ipo
+jax
 kdb
 kfd
 latencies
@@ -606,6 +612,7 @@ migraphx
 miopen
 miopengemm
 mivisionx
+mjx
 mkdir
 mlirmiopen
 mtypes

docs/conf.py

@@ -81,6 +81,7 @@ article_pages = [
         "file": "how-to/llm-fine-tuning-optimization/profiling-and-debugging",
         "os": ["linux"],
     },
+    {"file": "how-to/performance-validation/mi300x/vllm-benchmark", "os": ["linux"]},
     {"file": "how-to/system-optimization/index", "os": ["linux"]},
     {"file": "how-to/system-optimization/mi300x", "os": ["linux"]},
     {"file": "how-to/system-optimization/mi200", "os": ["linux"]},

docs/how-to/llm-fine-tuning-optimization/llm-inference-frameworks.rst

@@ -16,7 +16,7 @@ This section discusses how to implement `vLLM <https://docs.vllm.ai/en/latest>`_
 vLLM inference
 ==============
 
-vLLM is renowned for its paged attention algorithm that can reduce memory consumption and increase throughput thanks to
+vLLM is renowned for its PagedAttention algorithm that can reduce memory consumption and increase throughput thanks to
 its paging scheme. Instead of allocating GPU high-bandwidth memory (HBM) for the maximum output token lengths of the
 models, the paged attention of vLLM allocates GPU HBM dynamically for its actual decoding lengths. This paged attention
 is also effective when multiple requests share the same key and value contents for a large value of beam search or
@@ -139,9 +139,7 @@ Refer to :ref:`mi300x-vllm-optimization` for performance optimization tips.
 
 ROCm provides a prebuilt optimized Docker image for validating the performance of LLM inference with vLLM 
 on the MI300X accelerator. The Docker image includes ROCm, vLLM, PyTorch, and tuning files in the CSV 
-format. For more information, see the guide to 
+format. For more information, see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
-`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_ 
-on the ROCm GitHub repository.
 
 .. _fine-tuning-llms-tgi:
 

docs/how-to/performance-validation/mi300x/vllm-benchmark.rst

@@ -0,0 +1,407 @@
+.. meta::
+   :description: Learn how to validate LLM inference performance on MI300X accelerators using AMD MAD and the unified
+                 ROCm Docker image.
+   :keywords: model, MAD, automation, dashboarding, validate
+
+***********************************************************
+LLM inference performance validation on AMD Instinct MI300X
+***********************************************************
+
+.. _vllm-benchmark-unified-docker:
+
+The `ROCm vLLM Docker <https://hub.docker.com/r/rocm/vllm/tags>`_ image offers
+a prebuilt, optimized environment designed for validating large language model
+(LLM) inference performance on the AMD Instinct™ MI300X accelerator. This
+ROCm vLLM Docker image integrates vLLM and PyTorch tailored specifically for the
+MI300X accelerator and includes the following components:
+
+* `ROCm 6.2.1 <https://github.com/ROCm/ROCm>`_
+
+* `vLLM 0.6.4 <https://docs.vllm.ai/en/latest>`_
+
+* `PyTorch 2.5.0 <https://github.com/pytorch/pytorch>`_
+
+* Tuning files (in CSV format)
+
+With this Docker image, you can quickly validate the expected inference
+performance numbers on the MI300X accelerator. This topic also provides tips on
+optimizing performance with popular AI models.
+
+.. hlist::
+   :columns: 6
+
+   * Llama 3.1 8B
+
+   * Llama 3.1 70B
+
+   * Llama 3.1 405B
+
+   * Llama 2 7B
+
+   * Llama 2 70B
+
+   * Mixtral 8x7B
+
+   * Mixtral 8x22B
+
+   * Mixtral 7B
+
+   * Qwen2 7B
+
+   * Qwen2 72B
+
+   * JAIS 13B
+
+   * JAIS 30B
+
+.. _vllm-benchmark-vllm:
+
+.. note::
+
+   vLLM is a toolkit and library for LLM inference and serving. AMD implements
+   high-performance custom kernels and modules in vLLM to enhance performance.
+   See :ref:`fine-tuning-llms-vllm` and :ref:`mi300x-vllm-optimization` for
+   more information.
+
+Getting started
+===============
+
+Use the following procedures to reproduce the benchmark results on an
+MI300X accelerator with the prebuilt vLLM Docker image.
+
+.. _vllm-benchmark-get-started:
+
+1. Disable NUMA auto-balancing.
+
+   To optimize performance, disable automatic NUMA balancing. Otherwise, the GPU
+   might hang until the periodic balancing is finalized. For more information,
+   see :ref:`AMD Instinct MI300X system optimization <mi300x-disable-numa>`.
+
+   .. code-block:: shell
+
+      # disable automatic NUMA balancing
+      sh -c 'echo 0 > /proc/sys/kernel/numa_balancing'
+      # check if NUMA balancing is disabled (returns 0 if disabled)
+      cat /proc/sys/kernel/numa_balancing
+      0
+
+2. Download the :ref:`ROCm vLLM Docker image <vllm-benchmark-unified-docker>`.
+
+   Use the following command to pull the Docker image from Docker Hub.
+
+   .. code-block:: shell
+
+      docker pull rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
+
+Once setup is complete, you can choose between two options to reproduce the
+benchmark results:
+
+-  :ref:`MAD-integrated benchmarking <vllm-benchmark-mad>`
+
+-  :ref:`Standalone benchmarking <vllm-benchmark-standalone>`
+
+.. _vllm-benchmark-mad:
+
+MAD-integrated benchmarking
+===========================
+
+Clone the ROCm Model Automation and Dashboarding (`<https://github.com/ROCm/MAD>`__) repository to a local
+directory and install the required packages on the host machine.
+
+.. code-block:: shell
+
+   git clone https://github.com/ROCm/MAD
+   cd MAD
+   pip install -r requirements.txt
+
+Use this command to run a performance benchmark test of the Llama 3.1 8B model
+on one GPU with ``float16`` data type in the host machine.
+
+.. code-block:: shell
+
+   export MAD_SECRETS_HFTOKEN="your personal Hugging Face token to access gated models"
+   python3 tools/run_models.py --tags pyt_vllm_llama-3.1-8b --keep-model-dir --live-output --timeout 28800
+
+ROCm MAD launches a Docker container with the name
+``container_ci-pyt_vllm_llama-3.1-8b``. The latency and throughput reports of the
+model are collected in the following path: ``~/MAD/reports_float16/``.
+
+Although the following models are preconfigured to collect latency and
+throughput performance data, you can also change the benchmarking parameters.
+Refer to the :ref:`Standalone benchmarking <vllm-benchmark-standalone>` section.
+
+Available models
+----------------
+
+.. hlist::
+   :columns: 3
+
+   * ``pyt_vllm_llama-3.1-8b``
+
+   * ``pyt_vllm_llama-3.1-70b``
+
+   * ``pyt_vllm_llama-3.1-405b``
+
+   * ``pyt_vllm_llama-2-7b``
+
+   * ``pyt_vllm_llama-2-70b``
+
+   * ``pyt_vllm_mixtral-8x7b``
+
+   * ``pyt_vllm_mixtral-8x22b``
+
+   * ``pyt_vllm_mistral-7b``
+
+   * ``pyt_vllm_qwen2-7b``
+
+   * ``pyt_vllm_qwen2-72b``
+
+   * ``pyt_vllm_jais-13b``
+
+   * ``pyt_vllm_jais-30b``
+
+   * ``pyt_vllm_llama-3.1-8b_fp8``
+
+   * ``pyt_vllm_llama-3.1-70b_fp8``
+
+   * ``pyt_vllm_llama-3.1-405b_fp8``
+
+   * ``pyt_vllm_mixtral-8x7b_fp8``
+
+   * ``pyt_vllm_mixtral-8x22b_fp8``
+
+.. _vllm-benchmark-standalone:
+
+Standalone benchmarking
+=======================
+
+You can run the vLLM benchmark tool independently by starting the
+:ref:`Docker container <vllm-benchmark-get-started>` as shown in the following
+snippet.
+
+.. code-block::
+
+   docker pull rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
+   docker run -it --device=/dev/kfd --device=/dev/dri --group-add video --shm-size 128G --security-opt seccomp=unconfined --security-opt apparmor=unconfined --cap-add=SYS_PTRACE -v $(pwd):/workspace --env HUGGINGFACE_HUB_CACHE=/workspace --name vllm_v0.6.4 rocm/vllm:rocm6.2_mi300_ubuntu20.04_py3.9_vllm_0.6.4
+
+In the Docker container, clone the ROCm MAD repository and navigate to the
+benchmark scripts directory at ``~/MAD/scripts/vllm``.
+
+.. code-block::
+
+   git clone https://github.com/ROCm/MAD
+   cd MAD/scripts/vllm
+
+Command
+-------
+
+To start the benchmark, use the following command with the appropriate options.
+See :ref:`Options <vllm-benchmark-standalone-options>` for the list of
+options and their descriptions.
+
+.. code-block:: shell
+
+   ./vllm_benchmark_report.sh -s $test_option -m $model_repo -g $num_gpu -d $datatype
+
+See the :ref:`examples <vllm-benchmark-run-benchmark>` for more information.
+
+.. note::
+
+   The input sequence length, output sequence length, and tensor parallel (TP) are
+   already configured. You don't need to specify them with this script.
+
+.. note::
+
+   If you encounter the following error, pass your access-authorized Hugging
+   Face token to the gated models.
+
+   .. code-block:: shell
+
+      OSError: You are trying to access a gated repo.
+
+      # pass your HF_TOKEN
+      export HF_TOKEN=$your_personal_hf_token
+
+.. _vllm-benchmark-standalone-options:
+
+Options
+-------
+
+.. list-table::
+   :header-rows: 1
+   :align: center
+
+   * - Name
+     - Options
+     - Description
+
+   * - ``$test_option``
+     - latency
+     - Measure decoding token latency
+
+   * -
+     - throughput
+     - Measure token generation throughput
+
+   * -
+     - all
+     - Measure both throughput and latency
+
+   * - ``$model_repo``
+     - ``meta-llama/Meta-Llama-3.1-8B-Instruct``
+     - Llama 3.1 8B
+
+   * - (``float16``)
+     - ``meta-llama/Meta-Llama-3.1-70B-Instruct``
+     - Llama 3.1 70B
+
+   * -
+     - ``meta-llama/Meta-Llama-3.1-405B-Instruct``
+     - Llama 3.1 405B
+
+   * -
+     - ``meta-llama/Llama-2-7b-chat-hf``
+     - Llama 2 7B
+
+   * -
+     - ``meta-llama/Llama-2-70b-chat-hf``
+     - Llama 2 70B
+
+   * -
+     - ``mistralai/Mixtral-8x7B-Instruct-v0.1``
+     - Mixtral 8x7B
+
+   * -
+     - ``mistralai/Mixtral-8x22B-Instruct-v0.1``
+     - Mixtral 8x22B
+
+   * -
+     - ``mistralai/Mistral-7B-Instruct-v0.3``
+     - Mixtral 7B
+
+   * -
+     - ``Qwen/Qwen2-7B-Instruct``
+     - Qwen2 7B
+
+   * -
+     - ``Qwen/Qwen2-72B-Instruct``
+     - Qwen2 72B
+
+   * -
+     - ``core42/jais-13b-chat``
+     - JAIS 13B
+
+   * -
+     - ``core42/jais-30b-chat-v3``
+     - JAIS 30B
+
+   * - ``$model_repo``
+     - ``amd/Meta-Llama-3.1-8B-Instruct-FP8-KV``
+     - Llama 3.1 8B
+
+   * - (``float8``)
+     - ``amd/Meta-Llama-3.1-70B-Instruct-FP8-KV``
+     - Llama 3.1 70B
+
+   * -
+     - ``amd/Meta-Llama-3.1-405B-Instruct-FP8-KV``
+     - Llama 3.1 405B
+
+   * -
+     - ``amd/Mixtral-8x7B-Instruct-v0.1-FP8-KV``
+     - Mixtral 8x7B
+
+   * -
+     - ``amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV``
+     - Mixtral 8x22B
+
+   * - ``$num_gpu``
+     - 1 or 8
+     - Number of GPUs
+
+   * - ``$datatype``
+     - ``float16`` or ``float8``
+     - Data type
+
+.. _vllm-benchmark-run-benchmark:
+
+Running the benchmark on the MI300X accelerator
+-----------------------------------------------
+
+Here are some examples of running the benchmark with various options.
+See :ref:`Options <vllm-benchmark-standalone-options>` for the list of
+options and their descriptions.
+
+Example 1: latency benchmark
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+ 
+Use this command to benchmark the latency of the Llama 3.1 8B model on one GPU with the ``float16`` and ``float8`` data types.
+
+.. code-block::
+
+   ./vllm_benchmark_report.sh -s latency -m meta-llama/Meta-Llama-3.1-8B-Instruct -g 1 -d float16
+   ./vllm_benchmark_report.sh -s latency -m amd/Meta-Llama-3.1-8B-Instruct-FP8-KV -g 1 -d float8
+
+Find the latency reports at:
+
+- ``./reports_float16/summary/Meta-Llama-3.1-8B-Instruct_latency_report.csv``
+
+- ``./reports_float8/summary/Meta-Llama-3.1-8B-Instruct-FP8-KV_latency_report.csv``
+
+Example 2: throughput benchmark
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Use this command to benchmark the throughput of the Llama 3.1 8B model on one GPU with the ``float16`` and ``float8`` data types.
+
+.. code-block:: shell
+
+   ./vllm_benchmark_report.sh -s throughput -m meta-llama/Meta-Llama-3.1-8B-Instruct -g 1 -d float16
+   ./vllm_benchmark_report.sh -s throughput -m amd/Meta-Llama-3.1-8B-Instruct-FP8-KV -g 1 -d float8
+
+Find the throughput reports at:
+
+- ``./reports_float16/summary/Meta-Llama-3.1-8B-Instruct_throughput_report.csv``
+
+- ``./reports_float8/summary/Meta-Llama-3.1-8B-Instruct-FP8-KV_throughput_report.csv``
+
+.. raw:: html
+
+   <style>
+   mjx-container[jax="CHTML"][display="true"] {
+       text-align: left;
+       margin: 0;
+   }
+   </style>
+
+.. note::
+
+   Throughput is calculated as:
+
+   - .. math:: throughput\_tot = requests \times (\mathsf{\text{input lengths}} + \mathsf{\text{output lengths}}) / elapsed\_time
+
+   - .. math:: throughput\_gen = requests \times \mathsf{\text{output lengths}} / elapsed\_time
+
+Further reading
+===============
+
+- For application performance optimization strategies for HPC and AI workloads,
+  including inference with vLLM, see :doc:`/how-to/tuning-guides/mi300x/workload`.
+
+- To learn more about the options for latency and throughput benchmark scripts,
+  see `<https://github.com/ROCm/vllm/tree/main/benchmarks>`_.
+
+- To learn more about system settings and management practices to configure your system for
+  MI300X accelerators, see :doc:`/how-to/system-optimization/mi300x`.
+
+- To learn how to run LLM models from Hugging Face or your own model, see
+  :doc:`Using ROCm for AI </how-to/rocm-for-ai/index>`.
+
+- To learn how to optimize inference on LLMs, see
+  :doc:`Fine-tuning LLMs and inference optimization </how-to/llm-fine-tuning-optimization/index>`.
+
+- For a list of other ready-made Docker images for ROCm, see the
+  :doc:`Docker image support matrix <rocm-install-on-linux:reference/docker-image-support-matrix>`.
+
+- To compare with the previous version of the ROCm vLLM Docker image for performance validation, refer to
+  `LLM inference performance validation on AMD Instinct MI300X (ROCm 6.2.0) <https://rocm.docs.amd.com/en/docs-6.2.0/how-to/performance-validation/mi300x/vllm-benchmark.html>`_.
+

docs/how-to/tuning-guides/mi300x/index.rst

@@ -8,6 +8,8 @@ accelerators. They include detailed instructions on system settings and
 application tuning suggestions to help you fully leverage the capabilities of
 these accelerators, thereby achieving optimal performance.
 
+* :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`
+
 * :doc:`/how-to/tuning-guides/mi300x/system`
 
 * :doc:`/how-to/tuning-guides/mi300x/workload`

docs/how-to/tuning-guides/mi300x/workload.rst

@@ -152,9 +152,7 @@ address any new bottlenecks that may emerge.
 
 ROCm provides a prebuilt optimized Docker image that has everything required to implement
 the tips in this section. It includes ROCm, vLLM, PyTorch, and tuning files in the CSV 
-format. For more information, see the guide to 
+format. For more information, see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
-`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_ 
-on the ROCm GitHub repository.
 
 .. _mi300x-profiling-tools:
 
@@ -378,11 +376,10 @@ Refer to `vLLM documentation <https://docs.vllm.ai/en/latest/models/performance.
 for additional performance tips. :ref:`fine-tuning-llms-vllm` describes vLLM
 usage with ROCm.
 
-ROCm provides a prebuilt optimized Docker image for validating the performance of LLM inference with vLLM 
+ROCm provides a prebuilt optimized Docker image for validating the performance
-on the MI300X accelerator. The Docker image includes ROCm, vLLM, PyTorch, and tuning files in the CSV 
+of LLM inference with vLLM on the MI300X accelerator. The Docker image includes
-format. For more information, see the guide to 
+ROCm, vLLM, PyTorch, and tuning files in the CSV format. For more information,
-`LLM inference performance validation with vLLM on the AMD Instinct™ MI300X accelerator <https://github.com/ROCm/MAD/blob/develop/benchmark/vllm/README.md>`_ 
+see :doc:`/how-to/performance-validation/mi300x/vllm-benchmark`.
-on the ROCm GitHub repository.
 
 Maximize throughput
 -------------------

docs/index.md

@@ -45,7 +45,7 @@ ROCm documentation is organized into the following categories:
 * [Using ROCm for HPC](./how-to/rocm-for-hpc/index.rst)
 * [Fine-tuning LLMs and inference optimization](./how-to/llm-fine-tuning-optimization/index.rst)
 * [System optimization](./how-to/system-optimization/index.rst)
-* [AMD Instinct MI300X tuning guides](./how-to/tuning-guides/mi300x/index.rst)
+* [AMD Instinct MI300X performance validation and tuning](./how-to/tuning-guides/mi300x/index.rst)
 * [GPU cluster networking](https://rocm.docs.amd.com/projects/gpu-cluster-networking/en/latest/index.html)
 * [System debugging](./how-to/system-debugging.md)
 * [Using MPI](./how-to/gpu-enabled-mpi.rst)

docs/sphinx/_toc.yml.in

@@ -70,9 +70,11 @@ subtrees:
       - file: how-to/system-optimization/w6000-v620.md
         title: AMD RDNA 2
   - file: how-to/tuning-guides/mi300x/index.rst
-    title: AMD MI300X tuning guides
+    title: AMD MI300X performance validation and tuning
     subtrees:
     - entries:
+      - file: how-to/performance-validation/mi300x/vllm-benchmark.rst
+        title: Performance validation
       - file: how-to/tuning-guides/mi300x/system.rst
         title: System tuning
       - file: how-to/tuning-guides/mi300x/workload.rst