Add ray stage with 4 h100 gpus to CI and fix sampled logprobs in TRTLLM sampler

Signed-off-by: Shuyi Xiong <219646547+shuyixiong@users.noreply.github.com>

Author: shuyixiong

jenkins/L0_MergeRequest.groovy

@@ -734,6 +734,7 @@ def getMultiGpuFileChanged(pipeline, testFilter, globalVars)
         "tests/unittest/_torch/auto_deploy/unit/multigpu",
         "tests/unittest/_torch/multi_gpu/",
         "tests/unittest/_torch/multi_gpu_modeling/",
+        "tests/unittest/_torch/ray_orchestrator/multi_gpu/",
         "tests/unittest/disaggregated/",
         "tests/unittest/llmapi/test_llm_multi_gpu.py",
         "tests/unittest/llmapi/test_llm_multi_gpu_pytorch.py",

jenkins/L0_Test.groovy

@@ -2938,6 +2938,7 @@ def launchTestJobs(pipeline, testFilter)
         "DGX_H100-4_GPUs-PyTorch-DeepSeek-1": ["dgx-h100-x4-oci", "l0_dgx_h100", 1, 1, 4],
         "DGX_H100-4_GPUs-PyTorch-GptOss-1": ["dgx-h100-x4-oci", "l0_dgx_h100", 1, 1, 4],
         "DGX_H100-4_GPUs-PyTorch-Others-1": ["dgx-h100-x4-oci", "l0_dgx_h100", 1, 1, 4],
+        "DGX_H100-4_GPUs-PyTorch-Ray-1": ["dgx-h100-x4-oci", "l0_dgx_h100", 1, 1, 4],
         "B300-PyTorch-1": ["b300-single", "l0_b300", 1, 1],
         "DGX_B200-4_GPUs-PyTorch-1": ["b200-x4", "l0_dgx_b200", 1, 2, 4],
         "DGX_B200-4_GPUs-PyTorch-2": ["b200-x4", "l0_dgx_b200", 2, 2, 4],

tensorrt_llm/_torch/pyexecutor/sampler.py

@@ -3049,7 +3049,7 @@ def update_requests_single_beam_single_step(self, state: SampleStateTRTLLM):
 
             log_probs_host = log_probs_host_tensor[seq_slots, 0, seq_lens].tolist()
             idx = 0
-            for request in reqs_with_new_tokens:
+            for seq_slot, request in zip(seq_slots, reqs_with_new_tokens):
                 if request.py_return_log_probs:
                     log_probs = [
                         {