Skip to content

[BugFix][KVCache] Add enc_dec_block_num to prefill_kvcache_block_num check #1

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
kevincheng2 wants to merge 2 commits into from
Closed

[BugFix][KVCache] Add enc_dec_block_num to prefill_kvcache_block_num check #1

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
36d78cc
[Cherry-Pick][BugFix][KVCache] Add enc_dec_block_num to prefill_kvcac…
kevincheng2 Mar 30, 2026
255f572
[Cherry-Pick][BugFix][KVCache] Add enc_dec_block_num to prefill_kvcac…
kevincheng2 Mar 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 10 additions & 6 deletions fastdeploy/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1409,9 +1409,11 @@ def postprocess(self, num_total_tokens, number_of_tasks):
self.prefill_kvcache_block_num = self.total_block_num
else:
self.prefill_kvcache_block_num = int(self.total_block_num * self.kv_cache_ratio)
assert (
self.prefill_kvcache_block_num >= self.max_block_num_per_seq
), f"current block number :{self.prefill_kvcache_block_num} should be greater than or equal to current model len needed minimum block number :{self.max_block_num_per_seq}"
assert self.prefill_kvcache_block_num >= self.max_block_num_per_seq + self.enc_dec_block_num, (
f"prefill_kvcache_block_num: {self.prefill_kvcache_block_num} should be larger "
f"than or equal to {self.max_block_num_per_seq + self.enc_dec_block_num}, please reduce "
"the max_model_len or increase num_gpu_blocks_override"
)
else:
length = num_total_tokens // number_of_tasks
block_num = (length + self.block_size - 1 + self.dec_token_num) // self.block_size
Expand All @@ -1432,9 +1434,11 @@ def reset(self, num_gpu_blocks):
f"Reset block num, the total_block_num:{self.total_block_num},"
f" prefill_kvcache_block_num:{self.prefill_kvcache_block_num}"
)
assert (
self.prefill_kvcache_block_num >= self.max_block_num_per_seq
), f"current block number :{self.prefill_kvcache_block_num} should be greater than or equal to current model len needed minimum block number :{self.max_block_num_per_seq}"
assert self.prefill_kvcache_block_num >= self.max_block_num_per_seq + self.enc_dec_block_num, (
f"current device block num: {self.prefill_kvcache_block_num} "
f"should be larger than or equal to {self.max_block_num_per_seq + self.enc_dec_block_num}, please reduce "
"the max_model_len or replace the machine with larger GPU cards"
)

def print(self):
"""
Expand Down
Loading