Add nightly test multi gpu configs (#12721)

.github/workflows/nightly-test.yml

@@ -18,6 +18,7 @@ jobs:
   nightly-test-eval-text-models:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 2-gpu-runner
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -35,6 +36,7 @@ jobs:
   nightly-test-perf-text-models:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 2-gpu-runner
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -63,6 +65,7 @@ jobs:
   nightly-test-eval-vlms:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 2-gpu-runner
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -80,6 +83,7 @@ jobs:
   nightly-test-perf-vlms:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 2-gpu-runner
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -108,6 +112,7 @@ jobs:
   nightly-test-1-gpu:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 1-gpu-runner
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -125,6 +130,7 @@ jobs:
   nightly-test-4-gpu:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 4-gpu-h100
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -142,6 +148,7 @@ jobs:
   nightly-test-8-gpu-h200:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 8-gpu-h200
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -159,6 +166,7 @@ jobs:
   nightly-test-8-gpu-h20:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 8-gpu-h20
+    continue-on-error: true
     env:
       SGLANG_CI_RDMA_ALL_DEVICES: "mlx5_1,mlx5_2,mlx5_3,mlx5_4"
     steps:
@@ -178,6 +186,7 @@ jobs:
   nightly-test-4-gpu-b200:
     if: github.repository == 'sgl-project/sglang'
     runs-on: 4-gpu-b200
+    continue-on-error: true
     steps:
       - name: Checkout code
         uses: actions/checkout@v4

test/srt/run_suite.py

@@ -218,7 +218,10 @@ suites = {
     "nightly-4-gpu-b200": [
         TestFile("test_fp4_moe.py", 300),
     ],
+    "nightly-4-gpu": [],
     "nightly-8-gpu": [],
+    "nightly-8-gpu-h200": [],
+    "nightly-8-gpu-h20": [],
     "__not_in_ci__": [
         TestFile("ascend/test_ascend_w8a8_quantization.py"),
         TestFile("cpu/test_comm.py"),