vllm/csrc/cuda_utils.h

#pragma once

int64_t get_device_attribute(int64_t attribute, int64_t device_id);

int64_t get_max_shared_memory_per_block_device_attribute(int64_t device_id);
Avoid multiple redefinition (#1817) 2023-12-15 01:35:58 +08:00			`#pragma once`

[Kernel][Misc] Use TORCH_LIBRARY instead of PYBIND11_MODULE for custom ops (#5047) 2024-06-10 04:23:30 +08:00			`int64_t get_device_attribute(int64_t attribute, int64_t device_id);`
[Build] Avoid building too many extensions (#1624) 2023-11-24 08:31:19 +08:00
[Kernel][Misc] Use TORCH_LIBRARY instead of PYBIND11_MODULE for custom ops (#5047) 2024-06-10 04:23:30 +08:00			`int64_t get_max_shared_memory_per_block_device_attribute(int64_t device_id);`