add: gpu_regression test suite with per-push CI

- Move test_dflash.py to tests/gpu_regression/torch/speculative/
- Add cuda13-gpu-regression tox env
- Add gpu-regression matrix entry to gpu_tests.yml (15min timeout)
- Runs in parallel with existing gpu/gpu-megatron/gpu-trtllm jobs
- Triggers on changes to modelopt/**, tests/gpu_regression/**,
  examples/speculative_decoding/**, tools/launcher/**

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Chenhan Yu <chenhany@nvidia.com>

Author: ChenhanYu

.github/workflows/gpu_tests.yml

@@ -42,6 +42,11 @@ jobs:
             .github/workflows/gpu_tests.yml
             modelopt/**
             tests/gpu/**
+            tests/gpu_regression/**
+            examples/speculative_decoding/**
+            examples/dataset/**
+            modelopt_recipes/general/speculative_decoding/**
+            tools/launcher/**
             pyproject.toml
             tox.ini
           fail_on_initial_diff_error: true
@@ -66,6 +71,9 @@ jobs:
             timeout: 45
             container_image: pytorch:26.01-py3
             # tests/gpu/_extensions/test_onnx_extensions.py fails for newer containers until https://github.com/tbenthompson/cppimport/pull/98
+          - example: gpu-regression
+            timeout: 15
+            container_image: pytorch:26.01-py3
           - example: gpu-megatron
             timeout: 45
             container_image: pytorch:26.01-py3

…ples/speculative_decoding/test_dflash.py …ression/torch/speculative/test_dflash.pytests/examples/speculative_decoding/test_dflash.py renamed to tests/gpu_regression/torch/speculative/test_dflash.py tests/examples/speculative_decoding/test_dflash.py renamed to tests/gpu_regression/torch/speculative/test_dflash.py

@@ -3,6 +3,7 @@ envlist=
     pre-commit-all
     py312-torch210-tf_latest-unit
     cuda13-gpu
+    cuda13-gpu-regression
     cuda13-gpu-megatron
 skipsdist = True
 toxworkdir = /tmp/{env:USER}-modelopt-tox
@@ -68,6 +69,12 @@ commands_pre =
 commands =
     python -m pytest tests/gpu {env:COV_ARGS:}
 
+[testenv:cuda13-gpu-regression]
+commands_pre =
+    pip install -e .[hf,dev-test]
+commands =
+    python -m pytest tests/gpu_regression {env:COV_ARGS:}
+
 [testenv:cuda13-gpu-megatron]
 commands_pre =
     # Install deps here so that it gets installed even in --current-env