chore: update CI test est_time values (#38238)

Co-authored-by: sglang-bot <sglang-bot@users.noreply.github.com>

test/registered/unit/model_executor/model_runner_components/test_attention_backend_setup.py

@@ -13,7 +13,7 @@ from sglang.srt.model_executor.model_runner_components.attention_backend_setup i
 )
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 class _FakeBackend:

test/registered/unit/model_executor/model_runner_components/test_cuda_graph_setup.py

@@ -12,7 +12,7 @@ from sglang.srt.model_executor.model_runner_components.cuda_graph_setup import (
 )
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 def test_standard_gqa_gate_uses_pipeline_local_layer_range():

test/registered/unit/model_executor/model_runner_components/test_layer_setup.py

@@ -8,7 +8,7 @@ from sglang.srt.model_executor.model_runner_components.layer_setup import (
 )
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=6, suite="base-a-test-cpu")
 
 
 class TestComputeAttentionAndMoeLayers(unittest.TestCase):

test/registered/unit/model_executor/model_runner_components/test_ngram_embedding_manager.py

@@ -15,7 +15,7 @@ from sglang.srt.model_executor.model_runner_components.ngram_embedding_manager i
     update_ngram_token_table_after_sampling,
 )
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 def _make_ngram_info(batch_size: int, skip_token_table_update=None):

test/registered/unit/model_executor/model_runner_components/test_spec_aux_hidden_state.py

@@ -8,7 +8,7 @@ from sglang.srt.model_executor.model_runner_components.spec_aux_hidden_state imp
 )
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 @pytest.mark.parametrize(

test/registered/unit/model_executor/model_runner_components/test_startup_weight_load.py

@@ -31,7 +31,7 @@ from sglang.srt.model_loader.weight_utils import initialize_capture_safe_weights
 from sglang.srt.runtime_context import get_context, publish, reset_context
 from sglang.srt.server_args import ServerArgs
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 _STARTUP_MODULE = (

test/registered/unit/model_executor/runner/test_decode_cuda_graph_runner.py

@@ -36,7 +36,7 @@ from sglang.srt.utils import profile_utils as putils
 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-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 _CAPTURE_TRACE = "SGLANG_ENABLE_CUDA_GRAPH_CAPTURE_TRACE"
 _BATCH_CAPTURE = "SGLANG_GRAPH_BATCH_CAPTURE"

test/registered/unit/model_executor/runner/test_decode_cuda_graph_shared_read_fence.py

@@ -13,7 +13,7 @@ from sglang.srt.model_executor.runner.decode_cuda_graph_runner import (
 )
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 DECODE = ForwardMode.DECODE
 

test/registered/unit/model_executor/runner/test_flashinfer_autotune.py

@@ -8,7 +8,7 @@ from sglang.srt.model_executor.forward_batch_info import CaptureHiddenMode, Forw
 from sglang.srt.model_executor.runner import base_runner, flashinfer_autotune
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 @pytest.mark.parametrize(

test/registered/unit/model_executor/runner/test_flashinfer_autotune_sync.py

@@ -8,7 +8,7 @@ cover that gate and the digest it decides on.
 
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=50, suite="base-a-test-cpu")
+register_cpu_ci(est_time=52, suite="base-a-test-cpu")
 
 import json
 import multiprocessing

test/registered/unit/model_executor/runner/test_hidden_state_graph_recapture.py

@@ -19,7 +19,7 @@ from sglang.srt.runtime_context import get_context
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 class TestHiddenStateGraphRecapture(CustomTestCase):

test/registered/unit/model_executor/runner/test_prefill_cuda_graph_padding.py

@@ -13,7 +13,7 @@ from sglang.srt.model_executor.runner.prefill_cuda_graph_runner import (
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 class TestPrefillCudaGraphPadding(CustomTestCase):

test/registered/unit/model_executor/runner/test_prefill_shared_read_done.py

@@ -15,7 +15,7 @@ from sglang.srt.model_executor.runner_utils import (
 from sglang.srt.speculative.spec_info import SpeculativeAlgorithm
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 class _Event:

test/registered/unit/model_executor/runner_utils/test_graph_pool_borrow.py

@@ -17,7 +17,7 @@ from sglang.srt.speculative.dflash_worker_v2 import DFlashWorkerV2
 from sglang.test.ci.ci_register import register_cuda_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cuda_ci(est_time=60, stage="base-b", runner_config="1-gpu-small")
+register_cuda_ci(est_time=13, stage="base-b", runner_config="1-gpu-small")
 
 
 class TestGraphPoolBorrow(CustomTestCase):

test/registered/unit/model_executor/test_chunked_prefix_cache_gate.py

@@ -7,7 +7,7 @@ saw the flip, so an unsupported backend kept chunked prefix enabled.
 
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=4, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 import unittest
 

test/registered/unit/model_executor/test_cuda_graph_buffer_registry.py

@@ -29,7 +29,7 @@ from sglang.srt.model_executor.cuda_graph_buffer_registry import (
 from sglang.srt.model_executor.input_buffers import ForwardInputBuffers
 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=11, suite="base-a-test-cpu")
 
 
 @dataclasses.dataclass

test/registered/unit/model_executor/test_draft_runner_skips_lora.py

@@ -18,7 +18,7 @@ from sglang.srt.speculative.spec_info import SpeculativeAlgorithm
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=12, suite="base-a-test-cpu")
 
 
 class TestDraftRunnerSkipsLoRA(CustomTestCase):

test/registered/unit/model_executor/test_forward_metadata_plan_record.py

@@ -21,7 +21,7 @@ from sglang.srt.model_executor.forward_batch_info import ForwardBatch, ForwardMo
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=10, suite="base-a-test-cpu")
 
 
 def _make_batch(bs: int = 2, num_tokens: int = 2) -> ForwardBatch:

test/registered/unit/model_executor/test_hisparse_pool_configurator.py

@@ -9,7 +9,7 @@ from sglang.srt.runtime_context import get_context, get_parallel
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 class TestHiSparsePoolConfigurator(CustomTestCase):

test/registered/unit/model_executor/test_mlp_sync_pad_unpad.py

@@ -22,7 +22,7 @@ from sglang.srt.model_executor.runner.decode_cuda_graph_runner import (
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=10, suite="base-a-test-cpu")
 
 
 def _mock_model_runner(seq_len_fill_value: int = 1) -> MagicMock:

test/registered/unit/model_executor/test_model_runner_decode_rows.py

@@ -6,7 +6,7 @@ from unittest.mock import patch
 from sglang.srt.model_executor.model_runner import ModelRunner
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=1, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 class _FakeModelRunner:

test/registered/unit/model_executor/test_num_token_non_padded_localization.py

@@ -13,7 +13,7 @@ tensor and host-int twin agree, so a change to the sharding math fails loudly.
 
 from sglang.test.ci.ci_register import register_cpu_ci
 
-register_cpu_ci(est_time=2, suite="base-a-test-cpu")
+register_cpu_ci(est_time=7, suite="base-a-test-cpu")
 
 import unittest
 

test/registered/unit/model_executor/test_pool_configurator.py

@@ -21,7 +21,7 @@ from sglang.srt.runtime_context import (
 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-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 @contextlib.contextmanager

test/registered/unit/model_executor/test_prefill_cuda_graph_runner.py

@@ -27,7 +27,7 @@ from sglang.srt.runtime_context import get_context
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=2, suite="base-a-test-cpu")
+register_cpu_ci(est_time=11, suite="base-a-test-cpu")
 
 
 class _FakeAttentionBackend:

test/registered/unit/model_executor/test_prefill_cuda_graph_runner_helpers.py

@@ -21,7 +21,7 @@ from sglang.srt.model_loader.utils import resolve_language_model
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=2, suite="base-a-test-cpu")
+register_cpu_ci(est_time=10, suite="base-a-test-cpu")
 
 
 class _LayerModel:

test/registered/unit/model_executor/test_unified_out_cache_loc_rebind.py

@@ -36,7 +36,7 @@ from sglang.srt.model_executor.forward_batch_info import ForwardBatch, ForwardMo
 from sglang.test.ci.ci_register import register_cpu_ci
 from sglang.test.test_utils import CustomTestCase
 
-register_cpu_ci(est_time=5, suite="base-a-test-cpu")
+register_cpu_ci(est_time=10, suite="base-a-test-cpu")
 
 _DEV = "cpu"