guanguan0308 44ef9a36ac [fix]: fix precision issue in dispatch_ffn_combine_bf16 and remove redundant sync (#7198) ...

### What this PR does / why we need it?
Fix the precision issue in dispatch_ffn_combine_bf16 operator.
Remove redundant synchronization operations in dispatch_ffn_combine
operator.

- vLLM version: v0.16.0
- vLLM main:
4034c3d32e

---------

Signed-off-by: guanguan0308 <1546542263@qq.com>

2026-03-23 10:14:03 +08:00

..

moe_init_routing_quant_v2

[Build] Add support for Ascend950 chip (#7151)

2026-03-12 10:25:51 +08:00

unpermute

[Kernel]: Optimize DispatchFFNCombine performance (#6468)

2026-02-09 16:30:34 +08:00

utils

[Build] Add support for Ascend950 chip (#7151)

2026-03-12 10:25:51 +08:00

dispatch_ffn_combine_kernel.hpp

[fix]: fix precision issue in dispatch_ffn_combine_bf16 and remove redundant sync (#7198)

2026-03-23 10:14:03 +08:00

dispatch_ffn_combine_tiling.h

[CustomOp] support TensorList for dispatchFFNCombine (#5665)

2026-01-09 15:56:29 +08:00

dispatch_ffn_combine.cpp

[Kernel]: Optimize DispatchFFNCombine performance (#6468)

2026-02-09 16:30:34 +08:00

dispatch_ffn_combine.h

[Kernel]: Optimize DispatchFFNCombine performance (#6468)

2026-02-09 16:30:34 +08:00