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

[CI] fix ci (#2464)

Browse Source 
### What this PR does / why we need it?
1. use action/checkout@v5 instead of v4
2. remove dbo test case because there is issue with it and will be
refactored later
3. make vllm-ascend compatible with vllm v0.10.1.1 and add CI for it
4. fix sampler api changes introduced by
https://github.com/vllm-project/vllm/pull/22387
6. fix qwen3 moe config changes intruoduced by
https://github.com/vllm-project/vllm/pull/20562
7. fix kvcache block changes introduced by
https://github.com/vllm-project/vllm/pull/23262

### Does this PR introduce _any_ user-facing change?
N/A

### How was this patch tested?
CI passed with existing test.


- vLLM version: v0.10.0
- vLLM main:
0c6e40bbaa

---------

Signed-off-by: MengqingCao <cmq0113@163.com>
This commit is contained in:
Mengqing Cao
2025-08-22 07:30:48 +08:00
committed by GitHub
parent 0ca3f48c90
commit b0403f8d8a
27 changed files with 389 additions and 199 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
vllm_ascend/core/scheduler.py
View File

@@ -31,6 +31,13 @@ from vllm.v1.outputs import ModelRunnerOutput
from vllm.v1.request import Request, RequestStatus
from vllm.v1.structured_output import StructuredOutputManager
from vllm_ascend.utils import vllm_version_is
if vllm_version_is("0.10.1.1"):
from vllm.v1.core.kv_cache_manager import KVCacheBlocks
else:
KVCacheBlocks = None
class AscendScheduler(Scheduler):
"""This Scheduler extends vllm's original v1 scheduler
@@ -59,7 +66,10 @@ class AscendScheduler(Scheduler):
scheduled_running_reqs: list[Request] = []
preempted_reqs: list[Request] = []
req_to_new_block_ids: dict[str, list[int]] = {}
if vllm_version_is("0.10.1.1"):
req_to_new_block_ids: dict[str, list[int]] = {}
else:
req_to_new_blocks: dict[str, KVCacheBlocks] = {}
num_scheduled_tokens: dict[str, int] = {}
token_budget = self.max_num_scheduled_tokens
# Spec decode-related.
@@ -217,8 +227,11 @@ class AscendScheduler(Scheduler):
if self.lora_config and request.lora_request:
scheduled_loras.add(request.lora_request.lora_int_id)
req_to_new_block_ids[request.request_id] = (
self.kv_cache_manager.get_block_ids(request.request_id))
if vllm_version_is("0.10.1.1"):
req_to_new_block_ids[request.request_id] = (
self.kv_cache_manager.get_block_ids(request.request_id))
else:
req_to_new_blocks[request.request_id] = new_blocks
# Update request info.
num_scheduled_tokens[request.request_id] = num_new_tokens
token_budget -= num_new_tokens
@@ -307,8 +320,11 @@ class AscendScheduler(Scheduler):
# Schedule the request.
scheduled_running_reqs.append(request)
self.scheduled_req_ids.add(request.request_id)
req_to_new_block_ids[request.request_id] = (
new_blocks.get_block_ids())
if vllm_version_is("0.10.1.1"):
req_to_new_block_ids[request.request_id] = (
new_blocks.get_block_ids())
else:
req_to_new_blocks[request.request_id] = new_blocks
num_scheduled_tokens[request.request_id] = num_new_tokens
token_budget -= num_new_tokens
req_index += 1
@@ -346,16 +362,27 @@ class AscendScheduler(Scheduler):
any_request, len(self.running)))
# Construct the scheduler output.
new_reqs_data = [
NewRequestData.from_request(req,
req_to_new_block_ids[req.request_id])
for req in scheduled_new_reqs
]
if vllm_version_is("0.10.1.1"):
new_reqs_data = [
NewRequestData.from_request(
req, req_to_new_block_ids[req.request_id])
for req in scheduled_new_reqs
]
cached_reqs_data = self._make_cached_request_data(
scheduled_running_reqs, scheduled_resumed_reqs,
num_scheduled_tokens, scheduled_spec_decode_tokens,
req_to_new_block_ids)
else:
new_reqs_data = [
NewRequestData.from_request(
req, req_to_new_blocks[req.request_id].get_block_ids())
for req in scheduled_new_reqs
]
cached_reqs_data = self._make_cached_request_data(
scheduled_running_reqs, scheduled_resumed_reqs,
num_scheduled_tokens, scheduled_spec_decode_tokens,
req_to_new_block_ids)
cached_reqs_data = self._make_cached_request_data(
scheduled_running_reqs, scheduled_resumed_reqs,
num_scheduled_tokens, scheduled_spec_decode_tokens,
req_to_new_blocks)
scheduled_cached_reqs = cached_reqs_data
scheduler_output = SchedulerOutput(