Marco Bodrato bodrato at
Wed Mar 15 20:50:30 UTC 2017


Il Mer, 15 Marzo 2017 8:56 pm, Adrien Prost-Boucle ha scritto:
> Hi,
>> I miss a case: 32 bits; to fully evaluate the impact of the patch+FP on
>> one-limb operands in the range 1..62.
> Isn't 64-bit and 32-bit data identical, for one mpn_sqrtrem1 call on
> x86-64?
> I don't get why we would see a difference.

You measured:
128-bit, 2-limbs normalised;
96-bit, 2-limbs non-normalised;
64-bit, 1-limb normalised.

I'd like to see also:
32-bit, 1-limb non-normalised.

Your patch changes the way normalised and non-normalised numbers are
handled, for both 1 and 2 limbs. That's why I expect a difference.

> Or, do you mean we should add another test (along with tests for 1 and 2
> limbs), to check whether a 1-limb data fits in 32-bits?

No, I don't.

>> Did you try also with ABI=32 (16,32,48, and 64 bits)?

> No, not yet.

> I didn't know 16-bit and 48-bit ABI even existed... at least on a x86-64
> machine?

You are right. I mean: I'd like to see results with ABI=32 for operands of
size 16 (1-not-norm), 32 (1-norm), 48 (2-not-norm), and 64 (2-norm).



More information about the gmp-devel mailing list