[Klaud Cold] Update dsr1-fp4-b200-trt (+mtp) TRT-LLM image to v1.3.0rc14 (#1489)

* Update dsr1-fp4-b200-trt (+mtp) TRT-LLM image to v1.3.0rc14

Update TensorRT-LLM image (off: v1.2.0rc6.post2 104d / mtp: v1.2.0rc6.post3 101d) to v1.3.0rc14 (latest pre-release)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* chore: fill pr-link for #1489

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: functionstackx

.github/configs/nvidia-master.yaml

@@ -1873,7 +1873,7 @@ dsr1-fp4-b300-sglang:
       - { tp: 8, ep: 8, conc-start: 4, conc-end: 16 }
 
 dsr1-fp4-b200-trt:
-  image: nvcr.io#nvidia/tensorrt-llm/release:1.2.0rc6.post2
+  image: nvcr.io#nvidia/tensorrt-llm/release:1.3.0rc14
   model: nvidia/DeepSeek-R1-0528-FP4-V2
   model-prefix: dsr1
   runner: b200
@@ -1906,7 +1906,7 @@ dsr1-fp4-b200-trt:
       - { tp: 8, ep: 8, dp-attn: true, conc-start: 128, conc-end: 256 }
 
 dsr1-fp4-b200-trt-mtp:
-  image: nvcr.io#nvidia/tensorrt-llm/release:1.2.0rc6.post3
+  image: nvcr.io#nvidia/tensorrt-llm/release:1.3.0rc14
   model: nvidia/DeepSeek-R1-0528-FP4-V2
   model-prefix: dsr1
   runner: b200

perf-changelog.yaml

@@ -2869,3 +2869,10 @@
     - "1k1k and 8k1k STP low-latency and max-throughput srt-slurm recipes under benchmarks/multi_node/srt-slurm-recipes/sglang/glm5/b200-fp8/"
     - "Use the B200 local model alias glm5-fp8 mapped to /scratch/fsw/models/GLM-5-FP8"
   pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1372
+
+- config-keys:
+    - dsr1-fp4-b200-trt
+    - dsr1-fp4-b200-trt-mtp
+  description:
+    - "Update TensorRT-LLM image (off: v1.2.0rc6.post2 104d / mtp: v1.2.0rc6.post3 101d) to v1.3.0rc14 (latest pre-release)"
+  pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/1489