llama.cpp
bd562fe4 - cuda : use fast copy when src and dst are of different type and contiguous (#16789)

Commit

93 days ago

cuda : use fast copy when src and dst are of different type and contiguous (#16789) * use fast copy when src and dst are contiguous and same shape * use int64_t ne and ignore shape

References

#16789 - cuda : use fast copy when src and dst are of different type and contiguous

Author

CISC

Parents

bbac6a26

llama.cpp bd562fe4 - cuda : use fast copy when src and dst are of different type and contiguous (#16789)

llama.cpp
bd562fe4 - cuda : use fast copy when src and dst are of different type and contiguous (#16789)