[Bugfix] Add constraints for sequence parallelism (#4014)
### What this PR does / why we need it? Add Add constraints for sequence parallelism for unsupported scenarios: 1. tp_size > 1 2. enable_expert_parallel must be True for MoE model ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? - vLLM version: v0.11.0 - vLLM main: https://github.com/vllm-project/vllm/commit/83f478bb19489b41e9d208b47b4bb5a95ac171ac --------- Signed-off-by: realliujiaxu <realliujiaxu@163.com>
This commit is contained in:
1 parent
259eb25f88
commit
22005c64c1
2 files changed
+12
-3
No files matched your search
@@ -115,12 +115,10 @@ def set_ascend_forward_context(
|
||||
# the performance may degrade due to the switching of communication methods.
|
||||
mmrs_fusion = True
|
||||
if is_moe_model(vllm_config):
|
||||
sp_enabled = enable_sp(vllm_config) and \
|
||||
tp_world_size > 1 and num_tokens is not None
|
||||
sp_enabled = enable_sp(vllm_config) and num_tokens is not None
|
||||
mmrs_fusion = False
|
||||
else:
|
||||
sp_enabled = enable_sp(vllm_config) and \
|
||||
tp_world_size > 1 and \
|
||||
num_tokens is not None and num_tokens > 1000
|
||||
forward_context.mmrs_fusion = mmrs_fusion
|
||||
|
||||
|
||||
@@ -659,6 +659,17 @@ def enable_sp(vllm_config=None) -> bool:
|
||||
# We retain the env VLLM_ASCEND_ENABLE_FLASHCOMM here for backward compatibility.
|
||||
or bool(int(os.getenv("VLLM_ASCEND_ENABLE_FLASHCOMM", '0'))))
|
||||
|
||||
if not _ENABLE_SP:
|
||||
return _ENABLE_SP
|
||||
|
||||
assert vllm_config.parallel_config.tensor_parallel_size > 1, \
|
||||
"Flash Comm v1 (Sequence Parallelism) is only supported when tp_size > 1."
|
||||
|
||||
assert (
|
||||
not is_moe_model(vllm_config)
|
||||
or vllm_config.parallel_config.enable_expert_parallel
|
||||
), "Flash Comm v1 (Sequence Parallelism) requires enable_expert_parallel=True for MoE models."
|
||||
|
||||
return _ENABLE_SP
|
||||
|
||||
|
||||
|
||||
Reference in new issue
Block a user