chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

enable ut test for xpu devices (#11712)

Browse Source 
Co-authored-by: jundu <jun.du@intel.com>
Co-authored-by: Gao, Pengfei <pengfei.gao@intel.com>
This commit is contained in:
DiweiSun
2026-02-03 11:15:14 -08:00
committed by GitHub
co-authored by jundu Gao, Pengfei <pengfei.gao@intel.com>
parent 0a6925639b
commit 495290aefd
20 changed files with 237 additions and 151 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/radix_cache/test_mamba_unittest.py
+4 -3
View File
@@ -17,6 +17,7 @@ from sglang.srt.mem_cache.memory_pool import HybridLinearKVPool, HybridReqToToke
from sglang.srt.mem_cache.radix_cache import RadixKey
from sglang.srt.sampling.sampling_params import SamplingParams
from sglang.srt.server_args import ServerArgs, set_global_server_args_for_scheduler
from sglang.srt.utils import get_device
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")
@@ -39,7 +40,7 @@ class TestMamba(unittest.TestCase):
num_layers = 48
global_interval = 4
dtype = torch.bfloat16
device = "cuda"
device = get_device()
full_attention_layer_ids = [
i for i in range(global_interval - 1, num_layers, global_interval)
]
@@ -67,7 +68,7 @@ class TestMamba(unittest.TestCase):
max_num_reqs = 10
mamba_cache_size = 20
max_context_len = 128
device = "cuda"
device = get_device()
global_interval = 4
num_layers = 48
full_attention_layer_ids = [
@@ -152,7 +153,7 @@ class TestMamba(unittest.TestCase):
max_num_reqs = 10
mamba_cache_size = 20
max_context_len = 128
device = "cuda"
device = get_device()
full_attention_layer_ids = [
i for i in range(global_interval - 1, num_layers, global_interval)
]
test/registered/radix_cache/test_swa_unittest.py
+4 -3
View File
@@ -13,6 +13,7 @@ from sglang.srt.mem_cache.memory_pool import ReqToTokenPool
from sglang.srt.mem_cache.radix_cache import RadixKey
from sglang.srt.mem_cache.swa_memory_pool import SWAKVPool, SWATokenToKVPoolAllocator
from sglang.srt.mem_cache.swa_radix_cache import SWARadixCache
from sglang.srt.utils import get_device
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
register_cuda_ci(est_time=8, suite="stage-b-test-large-1-gpu")
@@ -37,7 +38,7 @@ class TestSWA(unittest.TestCase):
num_layers = 48
global_interval = 4
dtype = torch.bfloat16
device = "cuda"
device = get_device()
full_attention_layer_ids = [i for i in range(0, num_layers, global_interval)]
full_attention_layer_ids_set = set(full_attention_layer_ids)
swa_attention_layer_ids = [
@@ -89,7 +90,7 @@ class TestSWA(unittest.TestCase):
num_layers = 48
global_interval = 4
dtype = torch.bfloat16
device = "cuda"
device = get_device()
full_attention_layer_ids = [i for i in range(0, num_layers, global_interval)]
full_attention_layer_ids_set = set(full_attention_layer_ids)
swa_attention_layer_ids = [
@@ -243,7 +244,7 @@ class TestSWA(unittest.TestCase):
num_layers = 48
global_interval = 4
dtype = torch.bfloat16
device = "cuda"
device = get_device()
full_attention_layer_ids = [i for i in range(0, num_layers, global_interval)]
full_attention_layer_ids_set = set(full_attention_layer_ids)
swa_attention_layer_ids = [