Fix DeepSeek DP Attention + torch compile (#5367)
Co-authored-by: ispobock <ispobaoke@163.com>
This commit is contained in:
@@ -192,8 +192,7 @@ def _dp_gather(
|
||||
|
||||
if local_tokens.shape[0] > 0 and (is_partial or get_attention_tp_rank() == 0):
|
||||
assert (
|
||||
global_tokens.untyped_storage().data_ptr()
|
||||
!= local_tokens.untyped_storage().data_ptr()
|
||||
local_tokens.untyped_storage() is not global_tokens.untyped_storage()
|
||||
), "aliasing between global_tokens and local_tokens not allowed"
|
||||
memcpy_triton(
|
||||
global_tokens, local_tokens, 0, local_start_pos, local_num_tokens, False
|
||||
@@ -243,8 +242,7 @@ def dp_scatter(
|
||||
assert global_tokens.is_contiguous()
|
||||
if local_tokens.shape[0] > 0:
|
||||
assert (
|
||||
local_tokens.untyped_storage().data_ptr()
|
||||
!= global_tokens.untyped_storage().data_ptr()
|
||||
local_tokens.untyped_storage() is not global_tokens.untyped_storage()
|
||||
), "aliasing between local_tokens and global_tokens not allowed"
|
||||
memcpy_triton(
|
||||
local_tokens, global_tokens, 0, local_start_pos, local_num_tokens, True
|
||||
|
||||
@@ -1,7 +1,8 @@
|
||||
import json
|
||||
import pandas as pd
|
||||
import argparse
|
||||
import json
|
||||
import os
|
||||
|
||||
import pandas as pd
|
||||
from tabulate import tabulate
|
||||
|
||||
# Parse command-line arguments
|
||||
|
||||
@@ -28,6 +28,9 @@ class TestDPAttentionDP2TP2(CustomTestCase):
|
||||
"--enable-dp-attention",
|
||||
"--dp",
|
||||
"2",
|
||||
"--enable-torch-compile",
|
||||
"--torch-compile-max-bs",
|
||||
"2",
|
||||
],
|
||||
)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user