Forum: Math
2021-01-29, 22:29
|
|
Replies: 6
Views: 1,467
The key term here is supersingular elliptic...
The key term here is supersingular elliptic curves. For p = 2 mod 3, any curve of the form y^2 = x^3 + B is supersingular, p > 3.
Another common case is p = 3 mod 4, in which case y^2 = x^3 + x...
|
|
Forum: Factoring
2016-10-11, 15:11
|
|
Replies: 32
Views: 5,938
|
|
Forum: Factoring
2016-10-11, 04:43
|
|
Replies: 32
Views: 5,938
Instead of using the where command on gdb, can...
Instead of using the where command on gdb, can you try the following instead: x/32i $rip? It should tell us precisely what the illegal instruction is.
By the way, the assembler is entirely in its...
|
|
Forum: Lounge
2013-07-23, 02:23
|
|
Replies: 60
Views: 6,421
|
|
Forum: Soap Box
2013-06-18, 23:01
|
|
Replies: 760
Views: 71,607
|
|
Forum: Programming
2013-05-16, 01:48
|
|
Replies: 13
Views: 3,168
|
|
Forum: Programming
2013-05-15, 21:58
|
|
Replies: 13
Views: 3,168
|
|
Forum: Programming
2013-04-13, 01:37
|
|
Replies: 4
Views: 8,471
The reason your the original code fails is deep...
The reason your the original code fails is deep in the C++ standard's language. In short, when you want the compiler to deduce argument types, those types must not be dependent types. In the above...
|
|
Forum: Programming
2013-03-30, 02:36
|
|
Replies: 77
Views: 16,736
It's actually that way in GMP:...
It's actually that way in GMP: http://gmplib.org/devel/asm.html
Nehalem is slightly slower than Core 2, probably due to some loop control change in Nehalem, as the performance of ADD/ADC is...
|
|
Forum: Factoring
2013-03-16, 21:31
|
|
Replies: 36
Views: 8,418
|
|
Forum: Factoring
2013-03-15, 22:45
|
|
Replies: 14
Views: 1,648
|
|
Forum: Factoring
2013-03-15, 22:00
|
|
Replies: 14
Views: 1,648
Can this not be somewhat fixed by lowering the...
Can this not be somewhat fixed by lowering the factor bases, and putting more pressure on the splitting of large primes (effectively trading space for time with rho, ECM, etc)? Or does doing this...
|
|
Forum: Computer Science & Computational Number Theory
2012-08-09, 00:44
|
|
Replies: 103
Views: 14,017
|
|
Forum: Computer Science & Computational Number Theory
2012-08-01, 00:44
|
|
Replies: 103
Views: 14,017
Just for the record, there is a document with...
Just for the record, there is a document with Granlund's timings for all recent architectures, including the Sandy Bridge:
http://gmplib.org/~tege/x86-timing.pdf
Good luck with the submission!
|
|
Forum: Computer Science & Computational Number Theory
2012-08-01, 00:09
|
|
Replies: 103
Views: 14,017
I'd like to point out that estimating the...
I'd like to point out that estimating the K-folded loops (end of page 13) as simply cycles/F is not extremely accurate, as it assumes infinite execution units. On the Core 2, a MUL can finish at most...
|
|
Forum: Programming
2012-07-26, 03:39
|
|
Replies: 36
Views: 5,490
|
|
Forum: Programming
2012-07-24, 23:48
|
|
Replies: 36
Views: 5,490
|
|
Forum: Programming
2012-07-23, 23:58
|
|
Replies: 36
Views: 5,490
|
|
Forum: Programming
2012-07-22, 03:26
|
|
Replies: 36
Views: 5,490
Not sure if I'm adding anything new here, but...
Not sure if I'm adding anything new here, but here's a new version of the loop I wrote, with a slight register change (thanks George for the SBB hint, how could I miss that?):
; r10 = x[]
...
|
|
Forum: Programming
2012-07-21, 00:54
|
|
Replies: 36
Views: 5,490
|
|
Forum: Programming
2012-07-20, 20:09
|
|
Replies: 36
Views: 5,490
Do you have a free register? You can avoid the...
Do you have a free register? You can avoid the comparison by performing a SETC REG right after the SUBQ, provided you have an available register to put that into.
Not sure if this helps the loop...
|
|
Forum: Science & Technology
2012-06-19, 12:10
|
|
Replies: 8
Views: 2,821
|
|
Forum: GMP-ECM
2012-02-24, 11:39
|
|
Replies: 13
Views: 4,911
|
|
Forum: Programming
2012-01-18, 05:18
|
|
Replies: 357
Views: 28,661
|
|
Forum: Programming
2011-11-28, 17:50
|
|
Replies: 60
Views: 15,786
|
