Fix bug in symm mem pre-allocation default (#19082)
This commit is contained in:
@@ -1068,16 +1068,13 @@ class ServerArgs:
|
||||
if model_config.is_multimodal and not self.language_only:
|
||||
self.adjust_mem_fraction_for_vlm(model_config)
|
||||
|
||||
# If symm mem is enabled and prealloc size is not set, set it to 4GB
|
||||
if (
|
||||
self.enable_symm_mem
|
||||
and not envs.SGLANG_SYMM_MEM_PREALLOC_GB_SIZE.is_set()
|
||||
):
|
||||
envs.SGLANG_SYMM_MEM_PREALLOC_GB_SIZE.set(4)
|
||||
logger.warning(
|
||||
"Symmetric memory is enabled, setting symmetric memory prealloc size to 4GB as default."
|
||||
"Use environment variable SGLANG_SYMM_MEM_PREALLOC_GB_SIZE to change the prealloc size."
|
||||
)
|
||||
# If symm mem is enabled and prealloc size is not set, set it to 4GB
|
||||
if self.enable_symm_mem and not envs.SGLANG_SYMM_MEM_PREALLOC_GB_SIZE.is_set():
|
||||
envs.SGLANG_SYMM_MEM_PREALLOC_GB_SIZE.set(4)
|
||||
logger.warning(
|
||||
"Symmetric memory is enabled, setting symmetric memory prealloc size to 4GB as default."
|
||||
"Use environment variable SGLANG_SYMM_MEM_PREALLOC_GB_SIZE to change the prealloc size."
|
||||
)
|
||||
|
||||
def _generate_cuda_graph_batch_sizes(self):
|
||||
"""
|
||||
|
||||
Reference in New Issue
Block a user