more a100 (#67) (#70)

Signed-off-by: Harrison King Saturley-Hall <hsaturleyhal@nvidia.com>
Co-authored-by: Jason Zhou <jasonzho@nvidia.com>

Author: saturley-hall

tests/cli/e2e_validation/README.md

@@ -40,7 +40,7 @@ python3 tests/cli/e2e_validation/e2e_runner.py --mode <smoke|selective|full> [op
 **Usage**: `python3 e2e_runner.py --mode full --parallel 4 --continue-on-error`
 
 - ALL model combinations (~15 models)
-- ALL system configurations (h100_sxm, h200_sxm)
+- ALL system configurations (h100_sxm, h200_sxm, b200_sxm, gb200_sxm, a100_sxm)
 - ALL GPU configurations (8, 512 GPUs)  
 - ALL workload patterns (3 ISL/OSL combinations)
 - ALL performance targets (2 TPOT values)
@@ -117,7 +117,7 @@ pipeline {
 The testing framework systematically validates:
 
 - **15+ Models**: QWEN, LLAMA, DEEPSEEK, MOE variants
-- **2 Systems**: h100_sxm, h200_sxm, b200_sxm, gb200_sxm
+- **2 Systems**: h100_sxm, h200_sxm, b200_sxm, gb200_sxm, a100_sxm
 - **2 GPU Configs**: 8 GPUs, 512 GPUs
 - **3 Workload Patterns**: (4000,1000), (1000,2), (32,1000) ISL/OSL
 - **2 Performance Targets**: 10ms, 100ms TPOT

tools/simple_sdk_demo/cli/main.py

@@ -43,7 +43,7 @@ def parse(args):
     parser.add_argument("--system",
                         type=str,
                         default='h200_sxm',
-                        choices=['h100_sxm', 'h200_sxm'],
+                        choices=['h100_sxm', 'h200_sxm', 'b200_sxm', 'gb200_sxm', 'a100_sxm'],
                         help='GPU+system Type')
     parser.add_argument("--model",
                         type=str,

tools/simple_sdk_demo/sla_service/sla_service.py

@@ -36,7 +36,7 @@ def get_supported_models():
 
 @app.post("/sla")
 def post_sla(
-    system: str = Body("h200_sxm", description="hardware name, h200_sxm, h100_sxm"),       
+    system: str = Body("h200_sxm", description="hardware name, h200_sxm, h100_sxm, b200_sxm, gb200_sxm, a100_sxm"),       
     backend: str = Body("trtllm", description="backend name, trtllm, sglang, vllm"),
     version: str = Body("0.20.0", description="trtllm version, 0.20.0"), 
     model_name: str = Body("QWEN3_32B", description="model name"),