The notices and links on their site now seem to be much clearer and
more correct. So it looks to me like they have made some effort to
attend to the requirements of the license.
Bill.
On 2 November 2012 20:35, Bill Hart wrote:
> Not only do they not distribute the source code with the binary, bu
I understand. But note that a good proportion of the code in MPIR is
copyright the Free Software Foundation. They will hold us to account
as much as they will hold you to account, which is why I am not happy
for the link to remain live unless you can adjust the zip file to
comply with the license.
Thanks for the correction, but this program is written for my BSc thesis
and for learning purposes. The DVD will contain the MPIR files as well as
every other source i have used.
I really doubt this program will leave my computer and my college though.
On Friday, November 9, 2012 12:32:17 AM UTC
Visual Studio supports OpenMP, it is very simple for enabling parallelism
in C++.
Also you can try Intel C++ compiler, it has better implementation of OpenMP.
On Fri, Nov 9, 2012 at 7:08 AM, Richard Marton wrote:
> Hi there!
>
> I have made an RSA coder in C++ which runs on a single thread. The
On Thursday, November 8, 2012 11:27:22 PM UTC+1, Bill Hart wrote:
>
> I take it that the primes are up to about 512 bits in the examples you
> gave?
>
> Lower limit is 8, upper is 3072 bits.
> On my core 2 duo CPU 256bits finish in a second, 512 finishes in 2-4s,
> 1024 in 10-15s, 3072 bits a
-Original Message-
From: mpir-devel@googlegroups.com [mailto:mpir-devel@googlegroups.com] On
Behalf Of Bill Hart
Sent: Thursday, November 08, 2012 4:33 PM
To: mpir-devel@googlegroups.com
Cc: sage-de...@googlegroups.com; flint-devel
Subject: Re: [mpir-devel] MPIR 2.6.0 released + appeal for
By the way, I happened to notice that it is looking for tune_fft.c for
some reason (I think it shouldn't be doing that). This file is in the
directory fft/tune, so you could copy it wherever it is expecting to
find it (tune or build.vc10/tune I think).
Maybe that will get the tune project to work
Yes, no work has been done to maintain mpir_bench.
Bill.
On 9 November 2012 00:20, Dann Corbit wrote:
> The mpir_bench program appears to have aged a bit. There are three
> unresolved externals:
>
> 1>bench_two.obj : error LNK2001: unresolved external symbol __gmp_version
> 1>fermat_prime_p.ob
I'm sorry, I got that wrong. Here is what Brian wrote last time:
"I should mention that the tuning is very unreliable on Windows".
I misremembered that and thought that it wasn't supported.
It should work, but as Brian said last time, you do have to interleave
builds with tests and the same with
The mpir_bench program appears to have aged a bit. There are three unresolved
externals:
1>bench_two.obj : error LNK2001: unresolved external symbol __gmp_version
1>fermat_prime_p.obj : error LNK2001: unresolved external symbol
__gmpn_mulmod_2expp1
1>trn.obj : error LNK2001: unresolved external
-Original Message-
From: mpir-devel@googlegroups.com [mailto:mpir-devel@googlegroups.com] On
Behalf Of Bill Hart
Sent: Thursday, November 08, 2012 4:06 PM
To: mpir-devel@googlegroups.com
Cc: sage-de...@googlegroups.com; flint-devel
Subject: Re: [mpir-devel] MPIR 2.6.0 released + appeal for
I get zero failures on the tests, and three expected warnings:
c:\math\mpir-2.6.0\build.vc10\x64\Release>misc.locale.exe
Test suppressed for windows DLL
c:\math\mpir-2.6.0\build.vc10\x64\Release>mpf.reuse.exe
Test suppressed for windows DLL
c:\math\mpir-2.6.0\build.vc10\x64\Release>mpz.reuse.exe
T
I believe tune is not supported on Windows.
I'll let Brian answer the question about the tests. Glad they built on
the second attempt.
Bill.
On 9 November 2012 00:01, Dann Corbit wrote:
> All of the projects build for me except for the tune project. The mpir-tests
> project had three failures
All of the projects build for me except for the tune project. The mpir-tests
project had three failures to build, but the three failures for mpir-tests
built OK on a second pass (I suspect that there is a missing dependency for
those three projects and the first pass built some object or librar
Hi Dann,
Yes, that is correct. The i7 core is the Nehalem core basically. Those
i3/i5/i7 processors all use the same core I think, just different
cache and so on. At least that is my understanding. Either way, we do
detect the correct thing.
Bill.
On 8 November 2012 23:43, Dann Corbit wrote:
>
I have a Core i7 intel chip.
I assume that Nehalem build is preferred over Core2, since i7 is a descendent
of Nehalem rather than the Xeon core. Is that right?
-Original Message-
From: mpir-devel@googlegroups.com [mailto:mpir-devel@googlegroups.com] On
Behalf Of Bill Hart
Sent: Thursday
Sure, you can link against it without becoming LGPL.
But when I run the program at that link on my computer, it runs
without me providing an MPIR library. Since the author is asking about
this on our list, I assumed his binary must contain a copy of MPIR.
That is called a "combined work" in the LG
If it is an LPGL application, and it does not modify the library, does the
application that uses it also become LGPL?
According to my understanding, even a commercial application can link against
LGPL libraries.
-Original Message-
From: mpir-devel@googlegroups.com [mailto:mpir-devel@goog
Hi all,
After a very long alpha/beta release cycle, MPIR 2.6.0 has finally
been released!
It is available from our website http://mpir.org/
This release is dedicated to the memory of Jason Moxham who passed
away recently.
The main new features in this release are:
* new FFT implementation for
Apologies if I am missing something here, but shouldn't there be a
license text somewhere? Or does this not actually depend on MPIR
(which is LGPL v3+)?
Bill.
On 8 November 2012 22:11, Richard Marton wrote:
> https://www.dropbox.com/s/ufokd88d5o47ata/FINAL.zip
>
> This is the executable under wi
That number is trivially factored by brute force in a fraction of a second:
c:\Program Files\IBM\Informix\11.70\bin>factor 642565121024
first trying brute force division by small primes
PRIME FACTOR 2
PRIME FACTOR 2
PRIME FACTOR 2
PRIME FACTOR 2
PRIME FACTOR 2
PRIME FACTOR 2
I take it that the primes are up to about 512 bits in the examples you gave?
If so, then perhaps a different method of primality testing would be
better. There are good libraries dedicated to this, like ECPP.
Primality testing is an entire industry.
I don't personally know of a really fast algori
From: mpir-devel@googlegroups.com [mailto:mpir-devel@googlegroups.com] On
Behalf Of Richard Marton
Sent: Thursday, November 08, 2012 2:09 PM
To: mpir-devel@googlegroups.com
Subject: [mpir-devel] how would you recommend me to multithread an MPIR C++
application? pthread?
Hi there!
I have made an
https://www.dropbox.com/s/ufokd88d5o47ata/FINAL.zip
This is the executable under windows 7/8
if a dll is missing, install the redist first. When it runs, test it on
number like 64,256,512,1024
On Thursday, November 8, 2012 11:08:51 PM UTC+1, Richard Marton wrote:
>
> Hi there!
>
> I have made an
Hi there!
I have made an RSA coder in C++ which runs on a single thread. The prime
generation takes up most of the CPU time, I'd prefer to speed it up using
multiple cores.
The random number is generated by MPIR and tested by the Miller-Rabin
primality test with 40 iterations. I am using Visual
25 matches
Mail list logo