[TRTC-43] [feat] Add config db and docs

.gitignore

@@ -55,6 +55,8 @@ tensorrt_llm/scripts
 *docs/source/_cpp_gen*
 docs/source/**/*.rst
 !docs/source/examples/index.rst
+!docs/source/deployment-guide/config_table.rst
+!docs/source/deployment-guide/note_sections.rst
 *.swp
 
 # Testing

docs/source/deployment-guide/deployment-guide-for-deepseek-r1-on-trtllm.md

@@ -66,15 +66,15 @@ We maintain YAML configuration files with recommended performance settings in th
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/deepseek-r1-throughput.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-throughput.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/deepseek-r1-throughput.yaml
+```{literalinclude} ../../../examples/configs/curated/deepseek-r1-throughput.yaml
 ---
 language: shell
 prepend: |
@@ -90,15 +90,15 @@ To use the `DeepGEMM` MOE backend on B200/GB200, use this config instead:
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/deepseek-r1-deepgemm.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-deepgemm.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/deepseek-r1-deepgemm.yaml
+```{literalinclude} ../../../examples/configs/curated/deepseek-r1-deepgemm.yaml
 ---
 language: shell
 prepend: |
@@ -154,7 +154,7 @@ These options provide control over TensorRT LLM's behavior and are set within th
 
 #### `trust_remote_code`
 
- **Description:** Allows TensorRT LLM to download models and tokenizers from Hugging Face. This flag is passed directly to the Hugging Face API.
+* **Description:** Allows TensorRT LLM to download models and tokenizers from Hugging Face. This flag is passed directly to the Hugging Face API.
 
 #### `kv_cache_config`
 
@@ -429,3 +429,23 @@ $$
 $$
 \text{TPS} = \frac{\text{Num Output Tokens}}{T_{last} - T_{first}}
 $$
+
+## Preconfigured Recipes
+
+The following tables list recommended configurations from the comprehensive database for different performance profiles.
+
+```{eval-rst}
+.. include:: note_sections.rst
+   :start-after: .. start-note-traffic-patterns
+   :end-before: .. end-note-traffic-patterns
+
+.. include:: config_table.rst
+   :start-after: .. start-deepseek-ai/DeepSeek-R1-0528
+   :end-before: .. end-deepseek-ai/DeepSeek-R1-0528
+```
+
+```{eval-rst}
+.. include:: config_table.rst
+   :start-after: .. start-nvidia/DeepSeek-R1-0528-FP4-v2
+   :end-before: .. end-nvidia/DeepSeek-R1-0528-FP4-v2
+```

docs/source/deployment-guide/deployment-guide-for-gpt-oss-on-trtllm.md

@@ -64,15 +64,15 @@ For low-latency use cases:
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/gpt-oss-120b-latency.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/gpt-oss-120b-latency.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/gpt-oss-120b-latency.yaml
+```{literalinclude} ../../../examples/configs/curated/gpt-oss-120b-latency.yaml
 ---
 language: shell
 prepend: |
@@ -88,15 +88,15 @@ For max-throughput use cases:
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/gpt-oss-120b-throughput.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/gpt-oss-120b-throughput.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/gpt-oss-120b-throughput.yaml
+```{literalinclude} ../../../examples/configs/curated/gpt-oss-120b-throughput.yaml
 ---
 language: shell
 prepend: |
@@ -377,3 +377,17 @@ $$
 $$
 \text{TPS} = \frac{\text{Num Output Tokens}}{T_{last} - T_{first}}
 $$
+
+## Preconfigured Recipes
+
+The following table lists recommended configurations from the comprehensive database for different performance profiles.
+
+```{eval-rst}
+.. include:: note_sections.rst
+   :start-after: .. start-note-traffic-patterns
+   :end-before: .. end-note-traffic-patterns
+
+.. include:: config_table.rst
+   :start-after: .. start-openai/gpt-oss-120b
+   :end-before: .. end-openai/gpt-oss-120b
+```

docs/source/deployment-guide/deployment-guide-for-llama3.3-70b-on-trtllm.md

@@ -58,15 +58,15 @@ We maintain YAML configuration files with recommended performance settings in th
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/llama-3.3-70b.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/llama-3.3-70b.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/llama-3.3-70b.yaml
+```{literalinclude} ../../../examples/configs/curated/llama-3.3-70b.yaml
 ---
 language: shell
 prepend: |

docs/source/deployment-guide/deployment-guide-for-llama4-scout-on-trtllm.md

@@ -57,15 +57,15 @@ We maintain YAML configuration files with recommended performance settings in th
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/llama-4-scout.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/llama-4-scout.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/llama-4-scout.yaml
+```{literalinclude} ../../../examples/configs/curated/llama-4-scout.yaml
 ---
 language: shell
 prepend: |

docs/source/deployment-guide/deployment-guide-for-qwen3-next-on-trtllm.md

@@ -35,15 +35,15 @@ We maintain YAML configuration files with recommended performance settings in th
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/qwen3-next.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/qwen3-next.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/qwen3-next.yaml
+```{literalinclude} ../../../examples/configs/curated/qwen3-next.yaml
 ---
 language: shell
 prepend: |

docs/source/deployment-guide/deployment-guide-for-qwen3-on-trtllm.md

@@ -40,15 +40,15 @@ We maintain YAML configuration files with recommended performance settings in th
 
 ```shell
 TRTLLM_DIR=/app/tensorrt_llm # change as needed to match your environment
-EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/qwen3.yaml
+EXTRA_LLM_API_FILE=${TRTLLM_DIR}/examples/configs/curated/qwen3.yaml
 ```
 
 Note: if you don't have access to the source code locally, you can manually create the YAML config file using the code in the dropdown below.
 
 ````{admonition} Show code
 :class: dropdown
 
-```{literalinclude} ../../../examples/configs/qwen3.yaml
+```{literalinclude} ../../../examples/configs/curated/qwen3.yaml
 ---
 language: shell
 prepend: |

docs/source/deployment-guide/index.rst

@@ -6,15 +6,20 @@ Quick Start for Popular Models
 
 The table below contains ``trtllm-serve`` commands that can be used to easily deploy popular models including DeepSeek-R1, gpt-oss, Llama 4, Qwen3, and more.
 
-We maintain LLM API configuration files for these models containing recommended performance settings in the `examples/configs <https://github.com/NVIDIA/TensorRT-LLM/tree/main/examples/configs>`_ directory. The TensorRT LLM Docker container makes the config files available at ``/app/tensorrt_llm/examples/configs``, but you can customize this as needed:
+We maintain LLM API configuration files for these models containing recommended performance settings in two locations:
+
+* **Curated Examples**: `examples/configs/curated <https://github.com/NVIDIA/TensorRT-LLM/tree/main/examples/configs/curated>`_ - Hand-picked configurations for common scenarios.
+* **Comprehensive Database**: `examples/configs/database <https://github.com/NVIDIA/TensorRT-LLM/tree/main/examples/configs/database>`_ - A more comprehensive set of known-good configurations for various GPUs and traffic patterns.
+
+The TensorRT LLM Docker container makes these config files available at ``/app/tensorrt_llm/examples/configs/curated`` and ``/app/tensorrt_llm/examples/configs/database`` respectively. You can reference them as needed:
 
 .. code-block:: bash
 
    export TRTLLM_DIR="/app/tensorrt_llm" # path to the TensorRT LLM repo in your local environment
 
-.. note::
-
-   The configs here are specifically optimized for a target ISL/OSL (Input/Output Sequence Length) of 1024/1024. If your traffic pattern is different, you may benefit from additional tuning. In the future, we plan to provide more configs for a wider range of traffic patterns.
+.. include:: note_sections.rst
+   :start-after: .. start-note-quick-start-isl-osl
+   :end-before: .. end-note-quick-start-isl-osl
 
 This table is designed to provide a straightforward starting point; for detailed model-specific deployment guides, check out the guides below.
 
@@ -30,53 +35,53 @@ This table is designed to provide a straightforward starting point; for detailed
    * - `DeepSeek-R1 <https://huggingface.co/deepseek-ai/DeepSeek-R1-0528>`_
      - H100, H200
      - Max Throughput
-     - `deepseek-r1-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/deepseek-r1-throughput.yaml>`_
-     - ``trtllm-serve deepseek-ai/DeepSeek-R1-0528 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/deepseek-r1-throughput.yaml``
+     - `deepseek-r1-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/deepseek-r1-throughput.yaml>`_
+     - ``trtllm-serve deepseek-ai/DeepSeek-R1-0528 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-throughput.yaml``
    * - `DeepSeek-R1 <https://huggingface.co/deepseek-ai/DeepSeek-R1-0528>`_
      - B200, GB200
      - Max Throughput
-     - `deepseek-r1-deepgemm.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/deepseek-r1-deepgemm.yaml>`_
-     - ``trtllm-serve deepseek-ai/DeepSeek-R1-0528 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/deepseek-r1-deepgemm.yaml``
+     - `deepseek-r1-deepgemm.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/deepseek-r1-deepgemm.yaml>`_
+     - ``trtllm-serve deepseek-ai/DeepSeek-R1-0528 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-deepgemm.yaml``
    * - `DeepSeek-R1 (NVFP4) <https://huggingface.co/nvidia/DeepSeek-R1-FP4>`_
      - B200, GB200
      - Max Throughput
-     - `deepseek-r1-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/deepseek-r1-throughput.yaml>`_
-     - ``trtllm-serve nvidia/DeepSeek-R1-FP4 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/deepseek-r1-throughput.yaml``
+     - `deepseek-r1-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/deepseek-r1-throughput.yaml>`_
+     - ``trtllm-serve nvidia/DeepSeek-R1-FP4 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-throughput.yaml``
    * - `DeepSeek-R1 (NVFP4) <https://huggingface.co/nvidia/DeepSeek-R1-FP4-v2>`_
      - B200, GB200
      - Min Latency
-     - `deepseek-r1-latency.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/deepseek-r1-latency.yaml>`_
-     - ``trtllm-serve nvidia/DeepSeek-R1-FP4-v2 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/deepseek-r1-latency.yaml``
+     - `deepseek-r1-latency.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/deepseek-r1-latency.yaml>`_
+     - ``trtllm-serve nvidia/DeepSeek-R1-FP4-v2 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/deepseek-r1-latency.yaml``
    * - `gpt-oss-120b <https://huggingface.co/openai/gpt-oss-120b>`_
      - Any
      - Max Throughput
-     - `gpt-oss-120b-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/gpt-oss-120b-throughput.yaml>`_
-     - ``trtllm-serve openai/gpt-oss-120b --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/gpt-oss-120b-throughput.yaml``
+     - `gpt-oss-120b-throughput.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/gpt-oss-120b-throughput.yaml>`_
+     - ``trtllm-serve openai/gpt-oss-120b --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/gpt-oss-120b-throughput.yaml``
    * - `gpt-oss-120b <https://huggingface.co/openai/gpt-oss-120b>`_
      - Any
      - Min Latency
-     - `gpt-oss-120b-latency.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/gpt-oss-120b-latency.yaml>`_
-     - ``trtllm-serve openai/gpt-oss-120b --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/gpt-oss-120b-latency.yaml``
+     - `gpt-oss-120b-latency.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/gpt-oss-120b-latency.yaml>`_
+     - ``trtllm-serve openai/gpt-oss-120b --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/gpt-oss-120b-latency.yaml``
    * - `Qwen3-Next-80B-A3B-Thinking <https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Thinking>`_
      - Any
      - Max Throughput
-     - `qwen3-next.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/qwen3-next.yaml>`_
-     - ``trtllm-serve Qwen/Qwen3-Next-80B-A3B-Thinking --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/qwen3-next.yaml``
+     - `qwen3-next.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/qwen3-next.yaml>`_
+     - ``trtllm-serve Qwen/Qwen3-Next-80B-A3B-Thinking --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/qwen3-next.yaml``
    * - Qwen3 family (e.g. `Qwen3-30B-A3B <https://huggingface.co/Qwen/Qwen3-30B-A3B>`_)
      - Any
      - Max Throughput
-     - `qwen3.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/qwen3.yaml>`_
-     - ``trtllm-serve Qwen/Qwen3-30B-A3B --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/qwen3.yaml`` (swap to another Qwen3 model name as needed)
+     - `qwen3.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/qwen3.yaml>`_
+     - ``trtllm-serve Qwen/Qwen3-30B-A3B --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/qwen3.yaml`` (swap to another Qwen3 model name as needed)
    * - `Llama-3.3-70B (FP8) <https://huggingface.co/nvidia/Llama-3.3-70B-Instruct-FP8>`_
      - Any
      - Max Throughput
-     - `llama-3.3-70b.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/llama-3.3-70b.yaml>`_
-     - ``trtllm-serve nvidia/Llama-3.3-70B-Instruct-FP8 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/llama-3.3-70b.yaml``
+     - `llama-3.3-70b.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/llama-3.3-70b.yaml>`_
+     - ``trtllm-serve nvidia/Llama-3.3-70B-Instruct-FP8 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/llama-3.3-70b.yaml``
    * - `Llama 4 Scout (FP8) <https://huggingface.co/nvidia/Llama-4-Scout-17B-16E-Instruct-FP8>`_
      - Any
      - Max Throughput
-     - `llama-4-scout.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/llama-4-scout.yaml>`_
-     - ``trtllm-serve nvidia/Llama-4-Scout-17B-16E-Instruct-FP8 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/llama-4-scout.yaml``
+     - `llama-4-scout.yaml <https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/configs/curated/llama-4-scout.yaml>`_
+     - ``trtllm-serve nvidia/Llama-4-Scout-17B-16E-Instruct-FP8 --extra_llm_api_options ${TRTLLM_DIR}/examples/configs/curated/llama-4-scout.yaml``
 
 Model-Specific Deployment Guides
 ---------------------------------
@@ -94,3 +99,10 @@ The deployment guides below provide more detailed instructions for serving speci
    deployment-guide-for-qwen3-on-trtllm.md
    deployment-guide-for-qwen3-next-on-trtllm.md
    deployment-guide-for-kimi-k2-thinking-on-trtllm.md
+
+Comprehensive Configuration Database
+------------------------------------
+
+The table below lists all available pre-configured model scenarios in the TensorRT LLM configuration database. Each row represents a specific model, GPU, and performance profile combination with recommended request settings.
+
+.. include:: config_table.rst

docs/source/deployment-guide/note_sections.rst

@@ -0,0 +1,36 @@
+..
+   Reusable note sections for deployment guides.
+   Include specific notes using:
+
+   .. include:: note_sections.rst
+      :start-after: .. start-note-<name>
+      :end-before: .. end-note-<name>
+
+.. start-note-traffic-patterns
+
+.. note::
+
+   **Traffic Patterns**: The ISL (Input Sequence Length) and OSL (Output Sequence Length)
+   values in each configuration represent the **maximum supported values** for that config.
+   Requests exceeding these limits may result in errors.
+
+   To handle requests with input sequences **longer than the configured ISL**, add the following
+   to your config file:
+
+   .. code-block:: yaml
+
+      enable_chunked_prefill: true
+
+   This enables chunked prefill, which processes long input sequences in chunks rather than
+   requiring them to fit within a single prefill operation. Note that enabling chunked prefill
+   does **not** guarantee optimal performance—these configs are tuned for the specified ISL/OSL.
+
+.. end-note-traffic-patterns
+
+.. start-note-quick-start-isl-osl
+
+.. note::
+
+   The configs here are specifically optimized for a target ISL/OSL (Input/Output Sequence Length) of 1024/1024. If your traffic pattern is different, refer to the :ref:`Comprehensive Configuration Database` section below which covers a larger set of traffic patterns and performance profiles.
+
+.. end-note-quick-start-isl-osl