EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Revert "drop ascend scheduler" (#4580)

Browse Source 
Reverts vllm-project/vllm-ascend#4498
- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2
This commit is contained in:
Mengqing Cao
2025-11-29 22:20:48 +08:00
committed by GitHub
parent 4dbe4fd123
commit 517fd9272d
52 changed files with 2948 additions and 85 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/e2e/nightly/multi_node/config/models/DeepSeek-R1-W8A8-A2-torchair.yaml
View File

@@ -30,7 +30,7 @@ deployment:
--quantization ascend
--gpu-memory-utilization 0.9
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}'
--additional-config '{"torchair_graph_config":{"enabled":true,"enable_multistream_moe":true},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
--additional-config '{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_moe":true},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
-
server_cmd: >
@@ -51,7 +51,7 @@ deployment:
--quantization ascend
--gpu-memory-utilization 0.9
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}'
--additional-config '{"torchair_graph_config":{"enabled":true,"enable_multistream_moe":true},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
--additional-config '{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_moe":true},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
benchmarks:
acc:
case_type: accuracy

4
tests/e2e/nightly/multi_node/config/models/DeepSeek-R1-W8A8-A2.yaml
View File

@@ -31,7 +31,7 @@ deployment:
--gpu-memory-utilization 0.9
--enforce-eager
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}'
--additional-config '{"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
--additional-config '{"ascend_scheduler_config":{"enabled":false},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
-
server_cmd: >
@@ -53,5 +53,5 @@ deployment:
--gpu-memory-utilization 0.9
--enforce-eager
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}'
--additional-config '{"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
--additional-config '{"ascend_scheduler_config":{"enabled":false},"chunked_prefill_for_mla":true,"enable_weight_nz_layout":true}'
benchmarks:

8
tests/e2e/nightly/multi_node/config/models/DeepSeek-R1-W8A8-EPLB.yaml
View File

@@ -50,7 +50,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
-
server_cmd: >
@@ -80,7 +80,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
-
server_cmd: >
vllm serve vllm-ascend/DeepSeek-R1-0528-W8A8
@@ -111,7 +111,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
-
server_cmd: >
vllm serve vllm-ascend/DeepSeek-R1-0528-W8A8
@@ -141,7 +141,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true,"dynamic_eplb":true,"num_iterations_eplb_update":2048,"num_wait_worker_iterations":200}'
benchmarks:
perf:
case_type: performance

8
tests/e2e/nightly/multi_node/config/models/DeepSeek-R1-W8A8.yaml
View File

@@ -49,7 +49,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true}'
-
server_cmd: >
@@ -79,7 +79,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":false,"enable_multistream_shared_expert":false},"enable_prefill_optimizations":true,"enable_weight_nz_layout":true}'
-
server_cmd: >
vllm serve vllm-ascend/DeepSeek-R1-0528-W8A8
@@ -110,7 +110,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true}'
-
server_cmd: >
vllm serve vllm-ascend/DeepSeek-R1-0528-W8A8
@@ -140,7 +140,7 @@ deployment:
"kv_connector_module_path": "vllm_ascend.distributed.llmdatadist_c_mgr_connector"
}'
--additional-config
'{"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true}'
'{"ascend_scheduler_config":{"enabled":false},"torchair_graph_config":{"enabled":true,"enable_multistream_mla":true,"graph_batch_sizes":[28],"use_cached_graph":true,"enable_super_kernel":false},"multistream_overlap_shared_expert":true}'
benchmarks:
perf:
case_type: performance

4
tests/e2e/nightly/multi_node/config/models/DeepSeek-V3_2-Exp-bf16.yaml
View File

@@ -29,7 +29,7 @@ deployment:
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.9
--additional-config '{"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
--additional-config '{"ascend_scheduler_config":{"enabled":true},"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
-
server_cmd: >
@@ -49,5 +49,5 @@ deployment:
--trust-remote-code
--no-enable-prefix-caching
--gpu-memory-utilization 0.92
--additional-config '{"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
--additional-config '{"ascend_scheduler_config":{"enabled":true},"torchair_graph_config":{"enabled":true,"graph_batch_sizes":[16]}}'
benchmarks: