On 11/2/24 2:15 AM, Waldek Hebisch wrote:
I saw "Modular SIMD arithmetic in Mathemagix" by J van der Hoeven.
There's algorithm doing modular multiplication entirely in 32-bit.
Our current method stores tmp value in 64bit.
No, AFAICS Joris is extending 32-bit numbers to 64-bit.
Yes, you are right. I take a deeper look, for multiplication,
it unpacks each vector into 2 vectors (high bits and low bits)
and for u32, it is doing multiplication in u64.
The main contribution is that it is doing remainder in SIMD.
- Qian
--
You received this message because you are subscribed to the Google Groups "FriCAS -
computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/fricas-devel/bb886e52-2dd5-4c06-8095-b32c31e4890d%40gmail.com.
