I really mean both. It's possible that new MPN_COPY code which we'll
have relatively soon, will remove some of the performance deficit, but
there are still many cases where one can avoid a copy altogether, use
mpn's directly instead of mpz's or otherwise avoid using
mpz_import/export with its asso
On May 22, 9:49 pm, Bill Hart wrote:
> The main reason I recommended not to use that function in time
> critical stuff is it is very, very slow (for reasons I don't
> understand).
When the endianness and everything is right, mpz_import and mpz_export
use MPN_COPY which calls copyi (on platforms
On Fri, May 22, 2009 at 11:29 PM, Craig Citro wrote:
>
>> That link no longer works but when I skimmed it early it recommended
>> to never use mpz_import and mpz_export. Is there a particular reason?
>> (I use both functions in gmpy to convert between the internal
>> representation of a Python lo
The main reason I recommended not to use that function in time
critical stuff is it is very, very slow (for reasons I don't
understand).
The talk was about writing highly time critical code. The best way
around that is to roll your own, which can often be made much faster.
Bill.
2009/5/23 Willi
On Fri, May 22, 2009 at 7:13 PM, Case Vanhorsen wrote:
>
> On Wed, May 20, 2009 at 8:29 PM, Bill Hart
> wrote:
>>
>> Today at Sage Days 15 in Seattle, I gave a talk on MPIR.
>>
>> I've been experimenting with the Sage notebook, which is very capable,
>> but I am no, hence the (somewhat broken)
> That link no longer works but when I skimmed it early it recommended
> to never use mpz_import and mpz_export. Is there a particular reason?
> (I use both functions in gmpy to convert between the internal
> representation of a Python longint and an mpz.)
>
We also do the same thing in Sage to c
On Wed, May 20, 2009 at 8:29 PM, Bill Hart wrote:
>
> Today at Sage Days 15 in Seattle, I gave a talk on MPIR.
>
> I've been experimenting with the Sage notebook, which is very capable,
> but I am no, hence the (somewhat broken) notes from my talk are
> available here:
>
> http://www.sagenb.org/h