Gentoo Archives: gentoo-science

From: "François Bissey" <frp.bissey@×××××.com>
To: gentoo-science@l.g.o
Subject: Re: [gentoo-science] ntl bump
Date: Fri, 03 Apr 2015 08:13:47
Message-Id: 2CA54A2A-FAD3-437C-B01F-95C4E71CD63B@gmail.com
In Reply to: Re: [gentoo-science] ntl bump by Thomas Kahle
1 > On 3/04/2015, at 20:52, Thomas Kahle <tomka@g.o> wrote:
2 >
3 > Bump request was open for a while already:
4 > https://bugs.gentoo.org/show_bug.cgi?id=542682
5 > I bumped it yesterday, is the ntl-62 compat patch essential?
6
7 Well ntl 6.2.x only live in the sage-on-gentoo tree but I would imagine
8 it helps with later version. My pull request has actually been merged
9 in master before the 2.4.5 release but it seems to only live in master
10 for future flint 2.5:
11 https://github.com/wbhart/flint2/pull/95
12
13 > Patrick reported a test failure that I cannot reproduce yet:
14 > https://bugs.gentoo.org/show_bug.cgi?id=545378
15 > Do you know anything about it?
16 >
17
18 No. I am running tests again to see if it is brought by a recent change.
19 I’ll see if I can track to something.
20
21 François
22
23
24 > On 03/04/15 00:53, François Bissey wrote:
25 >> Speaking of flint would you mind importing 2.4.5 from the sage-on-gentoo
26 >> overlay to the main tree. I should have opened a bump request last month
27 >> when it was released.
28 >>
29 >> François
30 >>
31 >>> On 3/04/2015, at 01:33, Thomas Kahle <tomka@g.o> wrote:
32 >>>
33 >>> Bumped, but now flint became incompatible:
34 >>>
35 >>> https://github.com/wbhart/flint2/issues/131
36 >>>
37 >>> On 28/03/15 20:54, François Bissey wrote:
38 >>>> And Victor just announce ntl 9.0 on sage-devel:
39 >>>>
40 >>>> With much trepidation, I have introduced a (hopefully minor)
41 >>>> backward incompatibility into NTL.
42 >>>>
43 >>>> The interface to the single-precision modular arithmetic
44 >>>> routines has been modified slightly.
45 >>>> This interface change allows for more flexible and more
46 >>>> efficient implementation of these routines,
47 >>>> which play a crucial role at many levels in NTL.
48 >>>>
49 >>>> Basically, these changes to the interface abstract away
50 >>>> some implementation details that arguably should never been there
51 >>>> in the first place.
52 >>>> By coding to the new interface, NTL clients will be able to
53 >>>> benefit from the current and future improvements.
54 >>>>
55 >>>> In particular, on 64-bit x86/GCC platforms, single precision
56 >>>> moduli can now be up to 60 bits, rather than 50 bits.
57 >>>> While some operations may in fact be a little slower, the most important
58 >>>> ones (like MulModPrecon) should not be.
59 >>>> Using larger moduli speeds up a number of things, like ZZ_pX
60 >>>> arithmetic, as fewer primes need to be used in Chinese Remaindering steps.
61 >>>> Other applications benefit from larger moduli as well.
62 >>>>
63 >>>> It is expected that most NTL clients will not be affected at all.
64 >>>> Moreover, any code that needs to be updated will be detected
65 >>>> by the compiler, and the updates should be simple and mechanical.
66 >>>> There is also a configuration flag that will enable the legacy
67 >>>> interface (although this is not recommended practice).
68 >>>>
69 >>>> For more, go to http://www.shoup.net/ntl
70 >>>>
71 >>>>
72 >>>
73 >>> --
74 >>> Thomas Kahle
75 >>> http://dev.gentoo.org/~tomka/
76 >>>
77 >>
78 >>
79 >
80 > --
81 > Thomas Kahle
82 > http://dev.gentoo.org/~tomka/
83 >

Replies

Subject Author
Re: [gentoo-science] ntl bump Thomas Kahle <tomka@g.o>