On 17 Nov, 01:22, [EMAIL PROTECTED] wrote:
> On Sunday 16 November 2008 00:05:55 Bill Hart wrote:> The timings indeed look
> good. Thanks.
>
> > You might be interested to know that I've heard rumours (fully
> > substantiated) that someone has code for mpn_addmul_1 that runs at 2.5
> > c/l (I'm
On Sun, Nov 16, 2008 at 4:07 PM, Bill Hart <[EMAIL PROTECTED]> wrote:
>
> Yes you are right. Of course for a project like MPIR it would have to
> be something like B. We'd have to have a well defined "MPIR Group"
> which would be allowed to make decisions about commercial use. This
> group could c
On Sunday 16 November 2008 00:05:55 Bill Hart wrote:
> The timings indeed look good. Thanks.
>
> You might be interested to know that I've heard rumours (fully
> substantiated) that someone has code for mpn_addmul_1 that runs at 2.5
> c/l (I'm relying on my memory here, but I think those are the
>
Wow, thanks for the really thorough answers.
I'm personally happy to wait until the Barrett code settles down to
how you want it.
I'll do as you say and start by looking at making the asm code
"ready". The biggest change required (as it is x86_64) for eMPIRe is
to convert the code to intel forma
On Saturday 15 November 2008 22:20:01 Bill Hart wrote:
> Hi Jason,
>
> I'm back home so I can take a better look.
>
> I see you have stuff for repeated division by a constant using
> Barret's algorithm (and also tdiv_qr_2exp?? - how does that differ?).
>
if you mean the mpz_tdiv_ then its for
Yes you are right. Of course for a project like MPIR it would have to
be something like B. We'd have to have a well defined "MPIR Group"
which would be allowed to make decisions about commercial use. This
group could consist of the current active developers (and all
interested past contributors) w
On Sun, Nov 16, 2008 at 9:23 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
> I realise there are only four things that really matter to me:
>
> 1) That my copyright notice be maintained.
> 2) That any offer to redistribute in binary form is accompanied by an
> equal offer to redistribute in source form
Yeah I certainly don't disagree with that. Of course all contributions
to date are LGPL anyway, so no problem there.
Bill.
2008/11/16 William Stein <[EMAIL PROTECTED]>:
>
> On Sun, Nov 16, 2008 at 6:23 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>>
>> I've been doing some thinking about licensing a
On Sun, Nov 16, 2008 at 6:23 AM, Bill Hart <[EMAIL PROTECTED]> wrote:
>
> I've been doing some thinking about licensing and what I personally
> really care about.
>
> I realise there are only four things that really matter to me:
>
> 1) That my copyright notice be maintained.
> 2) That any offer t
I've been doing some thinking about licensing and what I personally
really care about.
I realise there are only four things that really matter to me:
1) That my copyright notice be maintained.
2) That any offer to redistribute in binary form is accompanied by an
equal offer to redistribute in so
10 matches
Mail list logo