[AMD] Fix CI - unit-test-backend-1-gpu-amd-mi35x and unit-test-backend-2-gpu-amd, stage-b-test-small-1-gpu-amd (#16675)
Co-authored-by: Bingxu Chen <bingxche@amd.com>
This commit is contained in:
YC Tseng
@@ -9,7 +9,7 @@ from sglang.test.test_utils import CustomTestCase
|
||||
|
||||
# Triton kernel unit test for KV indices creation
|
||||
register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestCreateKvIndices(CustomTestCase):
|
||||
|
||||
@@ -15,7 +15,7 @@ from sglang.test.test_utils import (
|
||||
|
||||
# RadixAttention server integration tests
|
||||
register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestRadixCacheFCFS(CustomTestCase):
|
||||
|
||||
@@ -10,7 +10,7 @@ from sglang.srt.mem_cache.swa_radix_cache import SWARadixCache
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestSWA(unittest.TestCase):
|
||||
|
||||
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (
|
||||
|
||||
# Torch native attention backend integration test with MMLU eval
|
||||
register_cuda_ci(est_time=150, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestTorchNativeAttnBackend(CustomTestCase):
|
||||
|
||||
@@ -21,7 +21,7 @@ from sglang.test.test_utils import (
|
||||
|
||||
# Triton attention backend integration test with latency benchmark and MMLU eval
|
||||
register_cuda_ci(est_time=200, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=1110, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=1110, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestTritonAttnBackend(CustomTestCase):
|
||||
|
||||
@@ -26,7 +26,7 @@ from sglang.test.test_utils import CustomTestCase
|
||||
register_cuda_ci(est_time=30, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=30,
|
||||
suite="stage-b-test-small-1-gpu",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="test was never enabled for AMD CI, needs validation",
|
||||
)
|
||||
|
||||
|
||||
@@ -16,7 +16,7 @@ from sglang.test.test_utils import (
|
||||
|
||||
# Sliding window attention with Triton backend (Gemma-3 model)
|
||||
register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=100, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestSlidingWindowAttentionTriton(CustomTestCase):
|
||||
|
||||
@@ -17,7 +17,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=190, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestTorchCompile(CustomTestCase):
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=111, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=179, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=179, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
|
||||
|
||||
@@ -4,7 +4,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
from sglang.test.gpt_oss_common import BaseTestGptOss
|
||||
|
||||
register_cuda_ci(est_time=402, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=750, suite="stage-b-test-small-1-gpu-amd")
|
||||
register_amd_ci(est_time=750, suite="stage-b-test-small-1-gpu-amd-mi35x")
|
||||
|
||||
|
||||
class TestGptOss1Gpu(BaseTestGptOss):
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=9, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci
|
||||
|
||||
register_amd_ci(est_time=520, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=520, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
"""
|
||||
Test LLaDA2 (Diffusion Language Model) on AMD GPUs.
|
||||
|
||||
@@ -25,7 +25,7 @@ from sglang.test.lora_utils import (
|
||||
from sglang.test.test_utils import CustomTestCase, is_in_ci
|
||||
|
||||
register_cuda_ci(est_time=82, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestLoRA(CustomTestCase):
|
||||
|
||||
@@ -32,7 +32,7 @@ from sglang.test.test_utils import CustomTestCase, is_in_ci
|
||||
register_cuda_ci(est_time=200, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=200,
|
||||
suite="stage-b-test-small-1-gpu",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/13107",
|
||||
)
|
||||
|
||||
|
||||
@@ -24,7 +24,7 @@ from sglang.test.runners import SRTRunner
|
||||
from sglang.test.test_utils import CustomTestCase
|
||||
|
||||
register_cuda_ci(est_time=224, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=224, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=224, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
PROMPTS = [
|
||||
"AI is a field of computer science focused on",
|
||||
|
||||
@@ -25,7 +25,7 @@ from sglang.test.lora_utils import (
|
||||
from sglang.test.test_utils import CustomTestCase, is_in_ci
|
||||
|
||||
register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
# All prompts are used at once in a batch.
|
||||
PROMPTS = [
|
||||
|
||||
@@ -6,7 +6,7 @@ import requests
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=32, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=32, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=32, suite="stage-b-test-small-1-gpu-amd")
|
||||
from prometheus_client.parser import text_string_to_metric_families
|
||||
from prometheus_client.samples import Sample
|
||||
|
||||
|
||||
@@ -16,7 +16,7 @@ from sglang.test.test_utils import (
|
||||
register_cuda_ci(est_time=194, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=242,
|
||||
suite="stage-a-test-1",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/13107",
|
||||
)
|
||||
|
||||
|
||||
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (
|
||||
register_cuda_ci(est_time=442, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=221,
|
||||
suite="stage-a-test-1",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/12574",
|
||||
)
|
||||
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Model tests for compressed tensors (FP8)
|
||||
register_cuda_ci(est_time=42, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=42, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=42, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
from types import SimpleNamespace
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Cross encoder model tests
|
||||
register_cuda_ci(est_time=100, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=150, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import multiprocessing as mp
|
||||
import random
|
||||
|
||||
@@ -4,7 +4,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
register_cuda_ci(est_time=73, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=73,
|
||||
suite="stage-b-test-small-1-gpu",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/11127",
|
||||
)
|
||||
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Qwen model tests
|
||||
register_cuda_ci(est_time=90, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=130, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=130, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
from types import SimpleNamespace
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Reward model tests
|
||||
register_cuda_ci(est_time=103, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=132, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=132, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
# Copyright 2023-2024 SGLang Team
|
||||
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Transformers fallback model tests
|
||||
register_cuda_ci(est_time=245, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=320, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=320, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import dataclasses
|
||||
import multiprocessing as mp
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# VLM (Vision Language Model) tests
|
||||
register_cuda_ci(est_time=270, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=420, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=420, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import argparse
|
||||
import random
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=80, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=30, suite="stage-a-test-1")
|
||||
register_amd_ci(est_time=30, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
|
||||
|
||||
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=70, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=141, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=141, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestOpenAIEmbedding(CustomTestCase):
|
||||
|
||||
@@ -29,7 +29,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=184, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=149, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=149, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestOpenAIServer(CustomTestCase):
|
||||
|
||||
@@ -31,7 +31,7 @@ from sglang.srt.entrypoints.openai.protocol import (
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=3, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestModelCard(unittest.TestCase):
|
||||
|
||||
@@ -24,7 +24,7 @@ from sglang.srt.utils import get_or_create_event_loop
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class _MockTokenizerManager:
|
||||
|
||||
@@ -14,7 +14,7 @@ from sglang.srt.managers.tokenizer_manager import TokenizerManager
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class _MockTemplateManager:
|
||||
|
||||
@@ -17,7 +17,7 @@ from sglang.srt.managers.io_struct import EmbeddingReqInput
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=10, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
# Mock TokenizerManager for embedding tests
|
||||
|
||||
@@ -22,7 +22,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=70, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=200, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=200, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestEnableThinking(CustomTestCase):
|
||||
|
||||
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=109, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=180, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=180, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestJSONModeMixin:
|
||||
|
||||
@@ -14,7 +14,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=7, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=20, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=20, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
# -------------------------------------------------------------------------
|
||||
|
||||
@@ -19,7 +19,7 @@ from sglang.test.test_utils import (
|
||||
register_cuda_ci(est_time=186, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(
|
||||
est_time=186,
|
||||
suite="stage-b-test-small-1-gpu",
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/11127",
|
||||
)
|
||||
|
||||
|
||||
@@ -24,7 +24,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=89, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=89, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=89, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestReasoningContentAPI(CustomTestCase):
|
||||
|
||||
@@ -15,7 +15,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=73, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=73, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestOpenAIServerFunctionCalling(CustomTestCase):
|
||||
|
||||
@@ -23,7 +23,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=120, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=258, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=258, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestToolChoiceLlama32(CustomTestCase):
|
||||
|
||||
@@ -23,7 +23,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=41, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=41, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestLargeMaxNewTokens(CustomTestCase):
|
||||
|
||||
@@ -12,7 +12,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=40, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestMatchedStop(CustomTestCase, MatchedStopMixin):
|
||||
|
||||
@@ -12,7 +12,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=6, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=47, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=47, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestOpenAIServerIgnoreEOS(CustomTestCase):
|
||||
|
||||
@@ -13,7 +13,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=38, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=31, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=31, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestRequestLengthValidation(CustomTestCase):
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Ops - Repeat Interleave tests (1-GPU)
|
||||
register_cuda_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=75, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=75, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import time
|
||||
from typing import Tuple
|
||||
|
||||
@@ -30,7 +30,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
OUTPUT_DIR = "./profiler_dir"
|
||||
|
||||
|
||||
@@ -11,7 +11,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
from sglang.test.test_utils import CustomTestCase
|
||||
|
||||
register_cuda_ci(est_time=44, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=22, suite="stage-a-test-1")
|
||||
register_amd_ci(est_time=22, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
# For test
|
||||
|
||||
@@ -15,7 +15,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=250, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=303, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=303, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestEvalFP8Accuracy(CustomTestCase):
|
||||
|
||||
@@ -9,7 +9,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
from sglang.test.test_utils import CustomTestCase
|
||||
|
||||
register_cuda_ci(est_time=8, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=12, suite="stage-a-test-1")
|
||||
register_amd_ci(est_time=12, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
def torch_scaled_mm(
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=9, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
import unittest
|
||||
from types import SimpleNamespace
|
||||
|
||||
@@ -1,7 +1,9 @@
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=210, suite="stage-b-test-small-1-gpu", disabled="see #14021")
|
||||
register_amd_ci(est_time=210, suite="stage-b-test-small-1-gpu", disabled="see #14021")
|
||||
register_amd_ci(
|
||||
est_time=210, suite="stage-b-test-small-1-gpu-amd", disabled="see #14021"
|
||||
)
|
||||
|
||||
import json
|
||||
import random
|
||||
|
||||
@@ -2,7 +2,7 @@ from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
# Rotary Embedding - MRoPE tests (1-GPU)
|
||||
register_cuda_ci(est_time=10, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=15, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
from typing import NamedTuple
|
||||
|
||||
|
||||
@@ -18,7 +18,7 @@ import os
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=41, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=60, suite="stage-b-test-small-1-gpu-amd")
|
||||
import random
|
||||
import unittest
|
||||
|
||||
|
||||
@@ -10,7 +10,7 @@ from sglang.srt.utils import kill_process_tree
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=82, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=82, suite="stage-b-test-small-1-gpu-amd")
|
||||
from sglang.test.test_utils import (
|
||||
DEFAULT_SMALL_MODEL_NAME_FOR_TEST,
|
||||
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
|
||||
|
||||
@@ -7,7 +7,7 @@ from sglang.srt.utils import kill_process_tree
|
||||
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
|
||||
|
||||
register_cuda_ci(est_time=66, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=66, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=66, suite="stage-b-test-small-1-gpu-amd")
|
||||
from sglang.test.run_eval import run_eval
|
||||
from sglang.test.test_utils import (
|
||||
DEFAULT_MODEL_NAME_FOR_TEST,
|
||||
|
||||
@@ -18,7 +18,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=230, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=345, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=345, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestMultiTokenizer(CustomTestCase):
|
||||
|
||||
@@ -24,7 +24,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
register_cuda_ci(est_time=77, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=117, suite="stage-b-test-small-1-gpu")
|
||||
register_amd_ci(est_time=117, suite="stage-b-test-small-1-gpu-amd")
|
||||
|
||||
|
||||
class TestSkipTokenizerInit(CustomTestCase):
|
||||
|
||||
+2
-1
@@ -20,7 +20,8 @@ PER_COMMIT_SUITES = {
|
||||
HWBackend.CPU: ["default", "stage-a-cpu-only"],
|
||||
HWBackend.AMD: [
|
||||
"stage-a-test-1",
|
||||
"stage-b-test-small-1-gpu",
|
||||
"stage-b-test-small-1-gpu-amd",
|
||||
"stage-b-test-small-1-gpu-amd-mi35x",
|
||||
"stage-b-test-large-2-gpu-amd",
|
||||
],
|
||||
HWBackend.CUDA: [
|
||||
|
||||
Reference in New Issue
Block a user