minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 13 Packages Projects Releases Wiki Activity

[CPU] Raise mem-fraction-static 0.1->0.2 in intel_amx backend a/b (#36260)

Browse Source 
Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
MingxuZh
2026-08-25 14:36:20 +08:00
committed by GitHub
co-authored by Claude
parent f2ef826f0c
commit b7f9fca26e
2 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/cpu/test_intel_amx_attention_backend_a.py
+2 -2
View File
@@ -34,7 +34,7 @@ class TestIntelAMXAttnBackend(CustomTestCase):
return DEFAULT_MLA_MODEL_NAME_FOR_TEST return DEFAULT_MLA_MODEL_NAME_FOR_TEST
@intel_amx_benchmark( @intel_amx_benchmark(
extra_args=["--batch-size", "4", "--mem-fraction-static", "0.1"], extra_args=["--batch-size", "4", "--mem-fraction-static", "0.2"],
min_throughput=40, min_throughput=40,
) )
def test_latency_default_model(self): def test_latency_default_model(self):
@@ -84,7 +84,7 @@ class TestDPAttention(CustomTestCase):
"--attention-backend", "--attention-backend",
"intel_amx", "intel_amx",
"--mem-fraction-static", "--mem-fraction-static",
"0.1", "0.2",
"--disable-overlap-schedule", "--disable-overlap-schedule",
"--tp", "--tp",
"2", "2",
test/registered/cpu/test_intel_amx_attention_backend_b.py
+2 -2
View File
@@ -20,14 +20,14 @@ register_cpu_ci(est_time=47, suite="base-b-test-cpu")
class TestIntelAMXAttnBackendQuant(CustomTestCase): class TestIntelAMXAttnBackendQuant(CustomTestCase):
@intel_amx_benchmark( @intel_amx_benchmark(
extra_args=["--batch-size", "4", "--mem-fraction-static", "0.1"], extra_args=["--batch-size", "4", "--mem-fraction-static", "0.2"],
min_throughput=150, min_throughput=150,
) )
def test_latency_fp8_qwen(self): def test_latency_fp8_qwen(self):
return DEFAULT_MODEL_NAME_FOR_TEST_QWEN_FP8 return DEFAULT_MODEL_NAME_FOR_TEST_QWEN_FP8
@intel_amx_benchmark( @intel_amx_benchmark(
extra_args=["--batch-size", "4", "--mem-fraction-static", "0.1"], extra_args=["--batch-size", "4", "--mem-fraction-static", "0.2"],
min_throughput=50, min_throughput=50,
) )
def test_latency_fp8_moe_model(self): def test_latency_fp8_moe_model(self):