Add vLLM v0.18.1 source tree with KV transfer abort fix
third_party/vllm/ now tracked in git for direct patch management.
Based on vLLM v0.18.1 release with one patch applied:
vllm/v1/core/sched/scheduler.py:
Replace fatal assert with graceful skip when KV transfer callback
arrives for an already-aborted request during PD disaggregated serving.
Future vLLM modifications should be made directly in third_party/vllm/
and committed normally. The patches/ directory is kept as documentation
of what changed from upstream.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
28
third_party/vllm/csrc/core/math.hpp
vendored
Normal file
28
third_party/vllm/csrc/core/math.hpp
vendored
Normal file
@@ -0,0 +1,28 @@
|
||||
#pragma once
|
||||
|
||||
#include <climits>
|
||||
#include <iostream>
|
||||
|
||||
inline constexpr uint32_t next_pow_2(uint32_t const num) {
|
||||
if (num <= 1) return num;
|
||||
return 1 << (CHAR_BIT * sizeof(num) - __builtin_clz(num - 1));
|
||||
}
|
||||
|
||||
template <typename A, typename B>
|
||||
static inline constexpr auto div_ceil(A a, B b) {
|
||||
return (a + b - 1) / b;
|
||||
}
|
||||
|
||||
// Round a down to the next multiple of b. The caller is responsible for making
|
||||
// sure that b is non-zero
|
||||
template <typename T>
|
||||
inline constexpr T round_to_previous_multiple_of(T a, T b) {
|
||||
return a % b == 0 ? a : (a / b) * b;
|
||||
}
|
||||
|
||||
// Round a up to the next multiple of b. The caller is responsible for making
|
||||
// sure that b is non-zero
|
||||
template <typename T>
|
||||
inline constexpr T round_to_next_multiple_of(T a, T b) {
|
||||
return a % b == 0 ? a : ((a / b) + 1) * b;
|
||||
}
|
||||
Reference in New Issue
Block a user