[CPU][CI]: rename Xeon CPU CI suites to stage-*-intel (#37395)

Co-authored-by: Ma Mingfei <mingfei.ma@intel.com>
2026-09-04 10:08:41 +08:00
parent c1b4d535d7
commit 8770c1db1f
109 changed files with 115 additions and 115 deletions
@@ -90,7 +90,7 @@ jobs:
          # full machine, no partitioning. 4 files: 3 TP + autoround.
          - runner: xeon-gnr
            role: gnr
-            suite: "base-b-tp-test-cpu"
+            suite: "stage-a-tp-test-cpu-intel"
            name: gnr

          # Three SPR boxes each run ONE full-machine container taking one
@@ -102,19 +102,19 @@ jobs:
          # per box keeps each server within the box's memory budget.
          - runner: xeon-spr
            role: spr
-            suite: "base-b-test-cpu,base-c-test-cpu"
+            suite: "stage-a-test-cpu-intel,stage-b-test-cpu-intel"
            name: spr1
            part_id: 0
            part_size: 3
          - runner: xeon-spr
            role: spr
-            suite: "base-b-test-cpu,base-c-test-cpu"
+            suite: "stage-a-test-cpu-intel,stage-b-test-cpu-intel"
            name: spr2
            part_id: 1
            part_size: 3
          - runner: xeon-spr
            role: spr
-            suite: "base-b-test-cpu,base-c-test-cpu"
+            suite: "stage-a-test-cpu-intel,stage-b-test-cpu-intel"
            name: spr3
            part_id: 2
            part_size: 3
@@ -61,7 +61,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=30, suite="base-a-test-cpu")
-register_cpu_ci(est_time=46, suite="base-c-test-cpu")
+register_cpu_ci(est_time=46, suite="stage-b-test-cpu-intel")


 _BENCH_SERVING_CLI_CASES = {
@@ -24,7 +24,7 @@ from sglang.test.test_utils import (
    popen_launch_server,
 )

-register_cpu_ci(est_time=183, suite="base-b-tp-test-cpu")
+register_cpu_ci(est_time=183, suite="stage-a-tp-test-cpu-intel")


 class TestAutoRoundCPUConfig(CustomTestCase):
@@ -8,7 +8,7 @@ from sglang.srt.server_args import ServerArgs, set_global_server_args_for_schedu
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import GeluAndMul, SiluAndMul, precision

-register_cpu_ci(est_time=9, suite="base-b-test-cpu")
+register_cpu_ci(est_time=9, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")

 torch.manual_seed(1234)
@@ -11,7 +11,7 @@ from sglang.srt.utils.numa_utils import init_threads_binding
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=6, suite="base-b-tp-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-a-tp-test-cpu-intel")


 class TestBinding(CustomTestCase):
@@ -11,7 +11,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=13, suite="base-b-test-cpu")
+register_cpu_ci(est_time=13, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -9,7 +9,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import parametrize, precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=7, suite="base-b-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-a-test-cpu-intel")

 causal_conv1d_weight_pack = torch.ops.sgl_kernel.causal_conv1d_weight_pack
 causal_conv1d_fwd = torch.ops.sgl_kernel.causal_conv1d_fwd_cpu
@@ -13,7 +13,7 @@ import torch.multiprocessing as mp
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase, find_available_port

-register_cpu_ci(est_time=43, suite="base-b-test-cpu")
+register_cpu_ci(est_time=43, suite="stage-a-test-cpu-intel")


 def run_distributed_test(rank, world_size, master_port, output_writer, fn):
@@ -21,7 +21,7 @@ from sglang.test.test_utils import (
    popen_launch_server,
 )

-register_cpu_ci(est_time=315, suite="base-b-tp-test-cpu")
+register_cpu_ci(est_time=315, suite="stage-a-tp-test-cpu-intel")


 class TestCPUGraph(CustomTestCase):
@@ -7,7 +7,7 @@ from torch.nn.functional import scaled_dot_product_attention
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=10, suite="base-b-test-cpu")
+register_cpu_ci(est_time=10, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")

 torch.manual_seed(1234)
@@ -7,7 +7,7 @@ from torch.nn.functional import scaled_dot_product_attention
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=7, suite="base-b-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -8,7 +8,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import parametrize, precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=54, suite="base-b-test-cpu")
+register_cpu_ci(est_time=54, suite="stage-a-test-cpu-intel")

 flash_attn_varlen_func = torch.ops.sgl_kernel.flash_attn_varlen_func

@@ -18,7 +18,7 @@ from sglang.test.cpu_test_utils import (
 )
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=9, suite="base-b-test-cpu")
+register_cpu_ci(est_time=9, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -21,7 +21,7 @@ from sglang.test.test_utils import (
    popen_launch_server,
 )

-register_cpu_ci(est_time=685, suite="base-b-tp-test-cpu")
+register_cpu_ci(est_time=685, suite="stage-a-tp-test-cpu-intel")


 class TestIntelAMXAttnBackend(CustomTestCase):
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
    intel_amx_benchmark,
 )

-register_cpu_ci(est_time=47, suite="base-b-test-cpu")
+register_cpu_ci(est_time=47, suite="stage-a-test-cpu-intel")


 class TestIntelAMXAttnBackendQuant(CustomTestCase):
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
    intel_amx_benchmark,
 )

-register_cpu_ci(est_time=477, suite="base-b-tp-test-cpu")
+register_cpu_ci(est_time=477, suite="stage-a-tp-test-cpu-intel")


 class TestIntelAMXAttnBackendQuant(CustomTestCase):
@@ -9,7 +9,7 @@ from torch.nn.functional import softplus
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision

-register_cpu_ci(est_time=7, suite="base-b-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -8,7 +8,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=8, suite="base-b-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -34,7 +34,7 @@ from sglang.test.cpu_test_utils import (
    unpack_and_dequant_awq,
 )

-register_cpu_ci(est_time=7, suite="base-b-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-a-test-cpu-intel")


 def run_fused_experts(
@@ -8,7 +8,7 @@ import torch
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import make_non_contiguous, precision

-register_cpu_ci(est_time=5, suite="base-b-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")

 torch.manual_seed(1234)
@@ -14,7 +14,7 @@ from sglang.test.cpu_test_utils import (
 )
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=7, suite="base-b-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-a-test-cpu-intel")

 convert_weight_packed = torch.ops.sgl_kernel.convert_weight_packed
 qkv_proj_with_rope = torch.ops.sgl_kernel.qkv_proj_with_rope
@@ -8,7 +8,7 @@ from sglang.srt.utils import is_host_cpu_arm64
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision

-register_cpu_ci(est_time=5, suite="base-b-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")

 torch.manual_seed(1234)
@@ -25,7 +25,7 @@ from sglang.srt.utils.rank_consensus_checker import (
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase, find_available_port

-register_cpu_ci(est_time=193, suite="base-b-test-cpu")
+register_cpu_ci(est_time=193, suite="stage-a-test-cpu-intel")


 def run_distributed_test(
@@ -8,7 +8,7 @@ from sglang.srt.entrypoints.http_request_decompression import (
 )
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=5, suite="base-b-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-a-test-cpu-intel")

 PAYLOAD = b'{"text":"hello world","n":7}'
 COMPRESSED = zstandard.ZstdCompressor().compress(PAYLOAD)
@@ -7,7 +7,7 @@ from starlette.datastructures import Headers
 from sglang.srt.entrypoints.request_headers import apply_header_overrides
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=4, suite="base-b-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-a-test-cpu-intel")


 def _obj():
@@ -17,7 +17,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=33, suite="base-b-test-cpu")
+register_cpu_ci(est_time=33, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")

 torch.manual_seed(1234)
@@ -10,7 +10,7 @@ from sglang.srt.arg_groups.validation_hook import validate_ib_devices
 from sglang.srt.server_args import ServerArgs
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=6, suite="base-b-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")


@@ -20,7 +20,7 @@ from sglang.test.cpu_test_utils import (
 )
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=8, suite="base-b-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-a-test-cpu-intel")

 torch.manual_seed(1234)

@@ -20,7 +20,7 @@ from sglang.test.server_fixtures.spec_eagle_fixture import EagleLlama2Base
 # Measured 780s all-green on a 40-core GNR socket (1 launch + 18 methods).
 register_cpu_ci(
    est_time=800,
-    suite="base-b-test-cpu",
+    suite="stage-a-test-cpu-intel",
    disabled="EagleLlama2Base needs gated meta-llama/Llama-2-7b-chat-hf",
 )

@@ -9,7 +9,7 @@ from sglang.test.server_fixtures.spec_eagle_fixture import Eagle3Base
 # (CUDA sibling: 360); tune from CI TIMINGS once it has run there.
 register_cpu_ci(
    est_time=480,
-    suite="base-b-test-cpu",
+    suite="stage-a-test-cpu-intel",
    disabled="EAGLE3 numerical parity mismatches on CPU intel_amx",
 )

@@ -21,7 +21,7 @@ from sglang.test.server_fixtures.spec_eagle_fixture import EagleLlama2Base
 # Measured 830s all-green on a 40-core GNR socket (1 launch + 14 methods).
 register_cpu_ci(
    est_time=850,
-    suite="base-b-test-cpu",
+    suite="stage-a-test-cpu-intel",
    disabled="EagleLlama2Base needs gated meta-llama/Llama-2-7b-chat-hf",
 )

@@ -9,7 +9,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.cpu_test_utils import precision
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=6, suite="base-b-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-a-test-cpu-intel")


 def _topk1_chain_inputs(bs, num_steps):
@@ -11,7 +11,7 @@ from sglang.srt.layers.quantization.fp4_kv_cache_quant_method import (
 from sglang.srt.mem_cache.memory_pool import MHATokenToKVPool
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=14, suite="base-b-test-cpu")
+register_cpu_ci(est_time=14, suite="stage-a-test-cpu-intel")

 torch.manual_seed(42)

@@ -43,7 +43,7 @@ from sglang.multimodal_gen.runtime.platforms.cuda import (
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=9, suite="base-b-test-cpu")
+register_cpu_ci(est_time=9, suite="stage-a-test-cpu-intel")


 class TestSubBlockSparseAttentionDispatch(CustomTestCase):
@@ -15,7 +15,7 @@ from sglang.srt.models.llama4 import Llama4MoE
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=30, suite="base-b-test-cpu")
+register_cpu_ci(est_time=30, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=10, suite="base-b-test-cpu-arm64")


@@ -23,7 +23,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=70, stage="nightly", runner_config="1-gpu-large")
 register_amd_ci(est_time=40, suite="nightly-amd-1-gpu", nightly=True)
-register_cpu_ci(est_time=48, suite="base-c-test-cpu")
+register_cpu_ci(est_time=48, suite="stage-b-test-cpu-intel")


 class TestCrashDump(CustomTestCase):
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=240, stage="nightly", runner_config="1-gpu-large")
 register_amd_ci(est_time=120, suite="nightly-amd-1-gpu", nightly=True)
-register_cpu_ci(est_time=237, suite="base-c-test-cpu")
+register_cpu_ci(est_time=237, suite="stage-b-test-cpu-intel")


 class BaseTestSoftWatchdog:
@@ -12,7 +12,7 @@ from sglang.srt.parser.reasoning_parser import KimiK2Detector as KimiK2Reasoning
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(5, "base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")


 def _make_tool(name, parameters=None):
@@ -22,7 +22,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=42, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=38, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=58, suite="base-c-test-cpu")
+register_cpu_ci(est_time=58, suite="stage-b-test-cpu-intel")


 class TestInputEmbeds(CustomTestCase):
@@ -36,7 +36,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=43, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=43, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=54, suite="base-c-test-cpu")
+register_cpu_ci(est_time=54, suite="stage-b-test-cpu-intel")

 CHUNKED_PREFILL_SIZE = 256

@@ -8,7 +8,7 @@ from sglang.test.kv_canary.e2e_base import (
 )
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestCanaryE2EBase(CustomTestCase):
@@ -8,7 +8,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.kv_canary.e2e_base import CanaryE2EBase
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 _GOOD_LINE: str = SwaDivergenceLog(
@@ -13,7 +13,7 @@ from sglang.srt.server_args import ServerArgs
 from sglang.test.ci.ci_register import register_amd_ci, register_cpu_ci

 register_amd_ci(est_time=30, suite="nightly-amd-1-gpu", nightly=True)
-register_cpu_ci(est_time=7, suite="base-c-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-b-test-cpu-intel")


 class MockTokenizerManager:
@@ -11,7 +11,7 @@ from sglang.test.test_utils import CustomTestCase

 register_cuda_ci(est_time=29, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=45, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=32, suite="base-c-test-cpu")
+register_cpu_ci(est_time=32, suite="stage-b-test-cpu-intel")


 class TestExternalModels(CustomTestCase):
@@ -13,7 +13,7 @@ from sglang.test.server_fixtures.default_fixture import DefaultServerBase

 register_cuda_ci(est_time=48, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=48, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=206, suite="base-c-test-cpu")
+register_cpu_ci(est_time=206, suite="stage-b-test-cpu-intel")


 class TestTransformersBackendEval(DefaultServerBase):
@@ -15,7 +15,7 @@ from sglang.srt.runtime_context import get_parallel
 from sglang.srt.server_args import ServerArgs, set_global_server_args_for_scheduler
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=6, suite="base-b-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-a-test-cpu-intel")


@pytest.fixture(autouse=True)
@@ -26,7 +26,7 @@ register_cuda_ci(
    runner_config="1-gpu-small",
 )
 register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=49, suite="base-c-test-cpu")
+register_cpu_ci(est_time=49, suite="stage-b-test-cpu-intel")

 _MODEL_NAME = "Qwen/Qwen3-0.6B"

@@ -31,7 +31,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=50, stage="base-b", runner_config="1-gpu-large")
 register_amd_ci(est_time=140, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=54, suite="base-c-test-cpu")
+register_cpu_ci(est_time=54, suite="stage-b-test-cpu-intel")

 # System message to guide Llama3.2 to produce proper tool call format
 SYSTEM_MESSAGE = (
@@ -28,7 +28,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=58, stage="base-b", runner_config="1-gpu-large")
 register_amd_ci(est_time=41, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=101, suite="base-c-test-cpu")
+register_cpu_ci(est_time=101, suite="stage-b-test-cpu-intel")


 class TestLargeMaxNewTokens(CustomTestCase):
@@ -16,7 +16,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=52, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=60, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=83, suite="base-c-test-cpu")
+register_cpu_ci(est_time=83, suite="stage-b-test-cpu-intel")


 class TestMatchedStop(CustomTestCase, MatchedStopMixin):
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=91, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=141, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=91, suite="base-c-test-cpu")
+register_cpu_ci(est_time=91, suite="stage-b-test-cpu-intel")


 class TestOpenAIEmbedding(CustomTestCase):
@@ -6,7 +6,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=15, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestQuantLogString(CustomTestCase):
@@ -20,7 +20,7 @@ from sglang.test.test_utils import (
 # RadixAttention server integration tests
 register_cuda_ci(est_time=100, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=100, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=405, suite="base-c-test-cpu")
+register_cpu_ci(est_time=405, suite="stage-b-test-cpu-intel")


 class TestRadixCacheFCFS(CustomTestCase):
@@ -17,7 +17,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=55, stage="base-b", runner_config="1-gpu-small")
 register_amd_ci(est_time=55, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=113, suite="base-c-test-cpu")
+register_cpu_ci(est_time=113, suite="stage-b-test-cpu-intel")

 MODEL = DEFAULT_SMALL_MODEL_NAME_FOR_TEST

@@ -7,7 +7,7 @@ from sglang.test.server_fixtures.default_fixture import DefaultServerBase
 MODEL = "openai/gpt-oss-20b"

 register_cuda_ci(est_time=151, stage="base-b", runner_config="1-gpu-large")
-register_cpu_ci(est_time=1602, suite="base-c-test-cpu")
+register_cpu_ci(est_time=1602, suite="stage-b-test-cpu-intel")


 class TestSWARadixCacheKL(KLDivergenceMixin, DefaultServerBase):
@@ -10,7 +10,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 # ---------------------------------------------------------------------------
@@ -19,7 +19,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")

 _HTTP_SCOPE = {
    "type": "http",
@@ -22,7 +22,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=50, stage="weekly", runner_config="1-gpu-large")
 register_amd_ci(est_time=120, suite="nightly-amd-1-gpu", nightly=True)
-register_cpu_ci(est_time=55, suite="base-c-test-cpu")
+register_cpu_ci(est_time=55, suite="stage-b-test-cpu-intel")


 class TestRoutingKeyScheduling(CustomTestCase):
@@ -14,7 +14,7 @@ from sglang.test.server_fixtures.streaming_session_fixture import (
 )

 register_cuda_ci(est_time=130, stage="extra-a", runner_config="1-gpu-large")
-register_cpu_ci(est_time=449, suite="base-c-test-cpu")
+register_cpu_ci(est_time=449, suite="stage-b-test-cpu-intel")


 class TestStreamingSessionSWARetractMixedChunk(
@@ -11,7 +11,7 @@ from sglang.test.simple_eval_mixed_prefix_gsm8k import (
 )
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 def _write_synthetic_dataset(path: str, n: int) -> None:
@@ -10,7 +10,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.run_eval import _run_sgl_eval, run_eval
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=6, suite="base-b-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-a-test-cpu-intel")


 def _write_fake_metrics(out_parent: Path, eval_name: str, payload: dict) -> None:
@@ -34,7 +34,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 register_cpu_ci(2.0, "base-a-test-cpu")


-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 def _make_scheduler(grammar_backend_name="none", skip_tokenizer=False):
@@ -18,7 +18,7 @@ from sglang.srt.parser.reasoning_parser import KimiK3Detector as KimiK3Reasoning
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(2.0, "base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class _DummyTokenizer:
@@ -15,7 +15,7 @@ from sglang.srt.constrained.utils import is_legacy_structural_tag
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(1.0, "base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class TestIsLegacyStructuralTag(unittest.TestCase):
@@ -59,7 +59,7 @@ from sglang.srt.managers.io_struct import EmbeddingReqInput
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=10, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 # Mock TokenizerManager for embedding tests
@@ -37,7 +37,7 @@ from sglang.srt.function_call.utils import get_schema_properties
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=20, suite="base-a-test-cpu")
-register_cpu_ci(est_time=70, suite="base-c-test-cpu")
+register_cpu_ci(est_time=70, suite="stage-b-test-cpu-intel")


@functools.lru_cache(maxsize=None)
@@ -19,7 +19,7 @@ from sglang.srt.function_call.utils import (
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(5, "base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 class TestJsonSchemaConstraint(unittest.TestCase):
@@ -24,7 +24,7 @@ from sglang.srt.function_call.json_array_parser import JsonArrayParser
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(5, "base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")


 class TestParallelToolCalls(unittest.TestCase):
@@ -10,7 +10,7 @@ from sglang.srt.function_call.core_types import StreamingParseResult
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(5, "base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class DummyDetector(BaseFormatDetector):
@@ -14,7 +14,7 @@ from sglang.srt.layers.moe.moe_runner.aiter import (
 from sglang.srt.layers.moe.moe_runner.base import MoeRunnerConfig
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=7, suite="base-c-test-cpu")
+register_cpu_ci(est_time=7, suite="stage-b-test-cpu-intel")


 def _runner_input():
@@ -11,7 +11,7 @@ from sglang.srt.layers.moe.utils import MoeA2ABackend, MoeRunnerBackend
 from sglang.srt.runtime_context import get_flags
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


@pytest.fixture
@@ -32,7 +32,7 @@ from sglang.srt.lora.layers import FusedMoEWithLoRA
 from sglang.srt.runtime_context import get_context, get_flags, get_parallel
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=15, suite="base-c-test-cpu")
+register_cpu_ci(est_time=15, suite="stage-b-test-cpu-intel")


 class _TestDispatchRunnerCore(DispatchMoeRunnerCore):
@@ -39,7 +39,7 @@ from sglang.test.test_utils import (

 register_cuda_ci(est_time=8, stage="base-b", runner_config="1-gpu-large")
 register_amd_ci(est_time=8, suite="stage-b-test-1-gpu-small-amd")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 class TestTokenizedReqInputMsgpack(unittest.TestCase):
@@ -4,7 +4,7 @@ import torch
 from sglang.srt.managers.mm_utils import _scatter_mm_embedding
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=5, suite="base-b-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-a-test-cpu-intel")

 NUM_TOKENS = 64

@@ -7,7 +7,7 @@ from sglang.srt.environ import envs
 from sglang.srt.managers import mm_schedule as mm_utils
 from sglang.test.ci.ci_register import register_cpu_ci

-register_cpu_ci(est_time=8, suite="base-b-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-a-test-cpu-intel")


@pytest.mark.parametrize(
@@ -42,7 +42,7 @@ from sglang.srt.utils.weight_versions import WeightVersionSpan
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 def _round_trip(obj):
@@ -13,7 +13,7 @@ from sglang.srt.managers.scheduler_components.flush_wrapper import (
 )

 register_cpu_ci(est_time=14, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestSchedulerFlushCache(unittest.TestCase):
@@ -24,7 +24,7 @@ from sglang.srt.runtime_context import publish, reset_context
 from sglang.srt.sampling.sampling_params import SamplingParams

 register_cpu_ci(est_time=15, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestSchedulerPauseGeneration(unittest.TestCase):
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")

 import unittest
 from unittest.mock import MagicMock
@@ -7,7 +7,7 @@ from unittest.mock import MagicMock, patch
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=60, suite="base-a-test-cpu", nightly=True)
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class TestCpuMonitor(unittest.TestCase):
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")

 import asyncio
 import unittest
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")

 import unittest

@@ -7,7 +7,7 @@ from sglang.srt.observability.utils import (
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class TestMetricsUtils(unittest.TestCase):
@@ -6,7 +6,7 @@ from typing import Any, Dict, List, Optional
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")

 import asyncio
 import json
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")

 import unittest
 from unittest.mock import MagicMock, patch
@@ -19,7 +19,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 class TestRegisterCompletionTemplate(CustomTestCase):
@@ -33,7 +33,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestConversationGetPrompt(CustomTestCase):
@@ -13,7 +13,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")


 class TestPrefixHold(CustomTestCase):
@@ -12,7 +12,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 class TestTemplateContentFormatDetection(CustomTestCase):
@@ -5,7 +5,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")

 # Simulated model output that contains think tags (e.g. from DeepSeek-R1)
 THINK_OUTPUT = (
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")

 import json
 import unittest
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=9, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")

 import unittest
 from unittest.mock import MagicMock
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=9, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")

 import unittest
 from types import SimpleNamespace
@@ -3,7 +3,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci, register_xpu_ci

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=8, suite="base-c-test-cpu")
+register_cpu_ci(est_time=8, suite="stage-b-test-cpu-intel")
 register_xpu_ci(est_time=10, suite="stage-a-test-1-gpu-xpu")

 import copy
@@ -94,7 +94,7 @@ from sglang.test.test_utils import (
 )

 register_cpu_ci(est_time=10, suite="base-a-test-cpu")
-register_cpu_ci(est_time=11, suite="base-c-test-cpu")
+register_cpu_ci(est_time=11, suite="stage-b-test-cpu-intel")

 # Mock get_device() so all tests run on CPU-only CI runners
 _mock_device = patch(
@@ -6,7 +6,7 @@ from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase

 register_cpu_ci(est_time=7, suite="base-a-test-cpu")
-register_cpu_ci(est_time=6, suite="base-c-test-cpu")
+register_cpu_ci(est_time=6, suite="stage-b-test-cpu-intel")


 class TestRunaiUtils(CustomTestCase):
@@ -14,7 +14,7 @@ from sglang.kernels.ops.diffusion.common.fallback_torch import (
 from sglang.test.ci.ci_register import register_cpu_ci, register_mlx_ci

 register_cpu_ci(est_time=1, suite="base-a-test-cpu")
-register_cpu_ci(est_time=1, suite="base-b-test-cpu")
+register_cpu_ci(est_time=1, suite="stage-a-test-cpu-intel")
 register_cpu_ci(est_time=1, suite="base-b-test-cpu-arm64")
 register_mlx_ci(est_time=1, suite="stage-a-unit-test-mlx")

@@ -1,7 +1,7 @@
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=4, suite="base-c-test-cpu")
+register_cpu_ci(est_time=4, suite="stage-b-test-cpu-intel")

 import unittest

@@ -11,7 +11,7 @@ from sglang.srt.utils.auth import AuthLevel, decide_request_auth
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class TestHttpServerAdminAuth(unittest.TestCase):
@@ -11,7 +11,7 @@ from sglang.srt.utils.json_response import (
 from sglang.test.ci.ci_register import register_cpu_ci

 register_cpu_ci(est_time=6, suite="base-a-test-cpu")
-register_cpu_ci(est_time=5, suite="base-c-test-cpu")
+register_cpu_ci(est_time=5, suite="stage-b-test-cpu-intel")


 class TestJSONResponseUtils(unittest.TestCase):
--- a/Show More
+++ b/Show More