chore: [Breaking Change] Rename cuda_graph_config padding_enabled field to enable_padding.

Signed-off-by: nv-guomingz <[email protected]>

Author: nv-guomingz

docs/source/blogs/Best_perf_practice_on_DeepSeek-R1_in_TensorRT-LLM.md

@@ -196,7 +196,7 @@ We are seeing meaningful speedup using FP8 KV cache, thus refreshing the numbers
 ```bash
 cat >./extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
   - 896
   - 512
@@ -263,7 +263,7 @@ YOUR_DATA_PATH=./dataset.txt
 
 cat >./extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
   - 1
   - 2

docs/source/blogs/tech_blog/blog3_Optimizing_DeepSeek_R1_Throughput_on_NVIDIA_Blackwell_GPUs.md

@@ -157,7 +157,7 @@ These optimizations target the overall execution flow, scheduling, and resource
 
     There is a feature called CUDA Graph padding in TensorRT-LLM, which is a good trade-off between the number of CUDA Graphs and the CUDA Graph hit ratio; it tries to pad a batch to the nearest one with a captured CUDA Graph. Normally you should enable the CUDA Graph padding feature to increase the CUDA Graph hit rate, but the padding itself has some overhead due to wasted tokens computation.
 
-    Users can opt-out the CUDA Graph padding feature to see the perf benefits, by setting the `cuda_graph_config:\n  padding_enabled: False`, see API here [Pytorch backend config](https://github.com/NVIDIA/TensorRT-LLM/blob/main/tensorrt_llm/_torch/pyexecutor/config.py#L41)
+    Users can opt-out the CUDA Graph padding feature to see the perf benefits, by setting the `cuda_graph_config:\n  enable_padding: False`, see API here [Pytorch backend config](https://github.com/NVIDIA/TensorRT-LLM/blob/main/tensorrt_llm/_torch/pyexecutor/config.py#L41)
 
 * Overlap Scheduler:
 

docs/source/performance/perf-overview.md

@@ -201,7 +201,7 @@ trtllm-bench --model $model_name throughput --dataset $dataset_file --backend py
 `llm_options.yml`
 ```yaml
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
     - 1
     - 2

docs/source/scripts/disaggregated/gen_yaml.py

@@ -190,7 +190,7 @@ def gen_config_file(config_path: str,
             'max_seq_len': 8576,
             'free_gpu_memory_fraction': gen_gpu_memory_fraction,
             'cuda_graph_config': {
-                'padding_enabled': True,
+                'enable_padding': True,
                 'batch_sizes': gen_cuda_graph_batch_sizes,
             },
             'print_iter_log': True,

examples/llm-api/quickstart_advanced.py

@@ -188,7 +188,7 @@ def setup_llm(args):
 
     cuda_graph_config = CudaGraphConfig(
         batch_sizes=args.cuda_graph_batch_sizes,
-        padding_enabled=args.cuda_graph_padding_enabled,
+        enable_padding=args.cuda_graph_padding_enabled,
     ) if args.use_cuda_graph else None
     llm = LLM(
         model=args.model_dir,

examples/models/core/deepseek_v3/README.md

@@ -142,7 +142,7 @@ python /app/tensorrt_llm/benchmarks/cpp/prepare_dataset.py \
 
 cat <<EOF > /tmp/extra-llm-api-config.yml
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes: [1, 4, 8, 12]
 EOF
 
@@ -169,7 +169,7 @@ python /app/tensorrt_llm/benchmarks/cpp/prepare_dataset.py \
 
 cat <<EOF > /tmp/extra-llm-api-config.yml
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes: [1, 2]
 moe_max_num_tokens: 16384
 EOF
@@ -237,7 +237,7 @@ To serve the model using `trtllm-serve`:
 ```bash
 cat >./extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
     - 1
     - 2
@@ -316,7 +316,7 @@ export TRTLLM_USE_UCX_KVCACHE=1
 
 cat >./gen-extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
     - 1
     - 2
@@ -538,7 +538,7 @@ python3 /path/to/TensorRT-LLM/benchmarks/cpp/prepare_dataset.py \
 
 cat >/path/to/TensorRT-LLM/extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
     - 1
     - 2

examples/models/core/qwen/README.md

@@ -745,7 +745,7 @@ To serve the model using `trtllm-serve`:
 ```bash
 cat >./extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
   - 1
   - 2
@@ -821,7 +821,7 @@ export TRTLLM_USE_UCX_KVCACHE=1
 
 cat >./gen-extra-llm-api-config.yml <<EOF
 cuda_graph_config:
-  padding_enabled: true
+  enable_padding: true
   batch_sizes:
     - 1
     - 2

examples/wide_ep/slurm_scripts/gen_yaml.py

@@ -196,7 +196,7 @@ def gen_config_file(config_path: str,
             'max_seq_len': 2176,
             'free_gpu_memory_fraction': gen_gpu_memory_fraction,
             'cuda_graph_config': {
-                'padding_enabled': True,
+                'enable_padding': True,
                 'batch_sizes': gen_cuda_graph_batch_sizes,
             },
             'print_iter_log': True,

tensorrt_llm/_torch/pyexecutor/model_engine.py

@@ -309,7 +309,7 @@ def get_rank_model_storage(model):
 def _filter_cuda_graph_batch_sizes(cuda_graph_batch_sizes: list[int],
                                    max_batch_size: int, max_num_tokens: int,
                                    max_draft_len: int,
-                                   padding_enabled: bool) -> list[int]:
+                                   enable_padding: bool) -> list[int]:
     # This is the largest possible batch size for a pure decoding batch.
     max_cuda_graph_bs = min(max_batch_size,
                             int(max_num_tokens / (1 + max_draft_len)))
@@ -326,8 +326,8 @@ def _filter_cuda_graph_batch_sizes(cuda_graph_batch_sizes: list[int],
             # is that if the user is OK padding to a batch size B, they should also
             # be OK with padding to some size B' < B since the performance will generally
             # just be better in the smaller case.
-            if padding_enabled and (i == 0
-                                    or result[i - 1] != max_cuda_graph_bs):
+            if enable_padding and (i == 0
+                                   or result[i - 1] != max_cuda_graph_bs):
                 logger.warning(
                     "CUDA graph padding is enabled, but one of the given CUDA graph "
                     f"batch sizes ({bs}) is larger than the executor's max batch size "

tensorrt_llm/bench/benchmark/utils/general.py

@@ -152,7 +152,7 @@ def get_settings(params: dict, dataset_metadata: DatasetMetadata, model: str,
             pass
 
     cuda_graph_config = {
-        "padding_enabled": True,
+        "enable_padding": True,
         "max_batch_size": max_batch_size
     }