From 7924b313fe45512810fa33537c318c30af1dc6a6 Mon Sep 17 00:00:00 2001
From: jthomson04 <jothomson@nvidia.com>
Date: Wed, 4 Feb 2026 19:28:12 -0800
Subject: [PATCH 1/3] fix gb200 model prefix

Signed-off-by: jthomson04 <jothomson@nvidia.com>
---
 .github/configs/nvidia-master.yaml | 2 +-
 runners/launch_gb200-nv.sh         | 8 +++++---
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/.github/configs/nvidia-master.yaml b/.github/configs/nvidia-master.yaml
index a5cad5206..fe174a123 100644
--- a/.github/configs/nvidia-master.yaml
+++ b/.github/configs/nvidia-master.yaml
@@ -2286,7 +2286,7 @@ dsr1-fp4-gb200-dynamo-trt:
 dsr1-fp8-gb200-dynamo-trt:
   image: nvcr.io#nvidia/ai-dynamo/tensorrtllm-runtime:0.8.1.post2
   model: deepseek-ai/DeepSeek-R1-0528
-  model-prefix: dsr1-fp8
+  model-prefix: dsr1
   runner: gb200
   precision: fp8
   framework: dynamo-trt
diff --git a/runners/launch_gb200-nv.sh b/runners/launch_gb200-nv.sh
index 83e94005e..4b770194c 100755
--- a/runners/launch_gb200-nv.sh
+++ b/runners/launch_gb200-nv.sh
@@ -18,12 +18,14 @@ elif [[ $FRAMEWORK == "dynamo-trt" ]]; then
     if [[ $MODEL_PREFIX == "gptoss" ]]; then
         export MODEL_PATH="/mnt/lustre01/models/gpt-oss-120b"
         export SERVED_MODEL_NAME="gpt-oss-120b"
-    elif [[ $MODEL_PREFIX == "dsr1" ]]; then
+    elif [[ $MODEL_PREFIX == "dsr1" && $PRECISION == "fp4" ]]; then
         export MODEL_PATH="/mnt/lustre01/models/deepseek-r1-0528-fp4-v2/"
         export SERVED_MODEL_NAME="deepseek-r1-fp4"
-    elif [[ $MODEL_PREFIX == "dsr1-fp8" ]]; then
+        export SRT_SLURM_MODEL_PREFIX="dsr1"
+    elif [[ $MODEL_PREFIX == "dsr1" && $PRECISION == "fp8" ]]; then
         export MODEL_PATH="/mnt/numa1/groups/sa-shared/models/deepseek-r1-0528/"
         export SERVED_MODEL_NAME="deepseek-r1-fp8"
+        export SRT_SLURM_MODEL_PREFIX="dsr1-fp8"
     else
         echo "Unsupported model prefix: $MODEL_PREFIX. Supported prefixes are: gptoss or dsr1"
         exit 1
@@ -138,7 +140,7 @@ srtctl_root: "${SRTCTL_ROOT}"
 
 # Model path aliases
 model_paths:
-  "${MODEL_PREFIX}": "${MODEL_PATH}"
+  "${SRT_SLURM_MODEL_PREFIX}": "${MODEL_PATH}"
 containers:
   dynamo-trtllm: ${SQUASH_FILE}
   dynamo-sglang: ${SQUASH_FILE}

From 98e5ac1e0a47fcf43d3f83bacd198682b5723138 Mon Sep 17 00:00:00 2001
From: jthomson04 <jothomson@nvidia.com>
Date: Wed, 4 Feb 2026 19:33:19 -0800
Subject: [PATCH 2/3] update perf-changelog

Signed-off-by: jthomson04 <jothomson@nvidia.com>
---
 perf-changelog.yaml | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index aa8ad57f9..2329c58d4 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -361,3 +361,9 @@
     - "8k1k: 14 scenarios (7 MTP, 7 STP) for long context workloads"
     - "Prefill workers: 1-5P, Decode workers: 1-4D, TP/EP: 8/16/32"
   pr-link: https://github.com/InferenceMAX/InferenceMAX/pull/617
+
+- config-keys:
+    - dsr1-fp8-gb200-dynamo-trt
+  description:
+    - "Fix model_prefix argument in yaml configs"
+  pr-link: https://github.com/InferenceMAX/InferenceMAX/pull/646
\ No newline at end of file

From 2dc7f4b4787787cea6b52f668a7bd490ad2d508b Mon Sep 17 00:00:00 2001
From: jthomson04 <jothomson@nvidia.com>
Date: Wed, 4 Feb 2026 19:34:42 -0800
Subject: [PATCH 3/3] newline

Signed-off-by: jthomson04 <jothomson@nvidia.com>
---
 perf-changelog.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/perf-changelog.yaml b/perf-changelog.yaml
index 2329c58d4..813c173e6 100644
--- a/perf-changelog.yaml
+++ b/perf-changelog.yaml
@@ -366,4 +366,4 @@
     - dsr1-fp8-gb200-dynamo-trt
   description:
     - "Fix model_prefix argument in yaml configs"
-  pr-link: https://github.com/InferenceMAX/InferenceMAX/pull/646
\ No newline at end of file
+  pr-link: https://github.com/InferenceMAX/InferenceMAX/pull/646