llama.cpp
f46bc30c
- HIP : optimize IQ2/IQ3 (`__vsub4` `__vcmpne4`) using SWAR (#27962)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
3 days ago
HIP : optimize IQ2/IQ3 (`__vsub4` `__vcmpne4`) using SWAR (#27962) * HIP : use bit manipulation for __vcmpne4 * HIP : use bit manipulation for __vsub4
References
#27962 - HIP : optimize IQ2/IQ3 (`__vsub4` `__vcmpne4`) using SWAR
Author
yanjs
Parents
709fe755
Loading