mersenneforum.org  

Go Back   mersenneforum.org > Search Forums

Showing results 1 to 25 of 1000
Search took 0.36 seconds.
Search: Posts Made By: bsquared
Forum: YAFU 2022-10-07, 16:30
Replies: 19
Views: 1,158
Posted By bsquared
These changes require the AVX512F, AVX512BW, and...

These changes require the AVX512F, AVX512BW, and AVX512VL flags
Forum: YAFU 2022-10-07, 02:03
Replies: 19
Views: 1,158
Posted By bsquared
No I haven't worked with it much. I'd be...

No I haven't worked with it much. I'd be interested in comparisons as well! When I'm done here I'll look into it more... maybe some things I'm doing now can be useful there as well.
Forum: YAFU 2022-10-06, 19:55
Replies: 19
Views: 1,158
Posted By bsquared
Fortunately with Zen4's AVX512 implementation...

Fortunately with Zen4's AVX512 implementation looking pretty good, I think AVX512 is here to stay, and will become a mainstream part of Intel's and AMD's server and eventually consumer products going...
Forum: YAFU 2022-10-06, 19:36
Replies: 19
Views: 1,158
Posted By bsquared
The medium prime bucket sieve can use the same...

The medium prime bucket sieve can use the same improvement as the large prime one...


Original GGNFS:
size time (s)
300 41
330 53
360 121
390 44
420 59
Forum: YAFU 2022-10-06, 16:13
Replies: 19
Views: 1,158
Posted By bsquared
Another small improvement. This one helps the...

Another small improvement. This one helps the sieving portion; more so as the siever size and factor base size go up.

Same test cases as before.


Original GGNFS:
size time (s)
300 ...
Forum: YAFU 2022-10-05, 16:57
Replies: 19
Views: 1,158
Posted By bsquared
New improvement

The large prime bucket sieve (lasched) and trial division (MMX-TD) benefitted from AVX512 as well. The changes were actually very similar to things I did in yafu's SIQS.

Here is a summary of a...
Forum: YAFU 2022-10-03, 14:52
Replies: 19
Views: 1,158
Posted By bsquared
You are right! The microecm improvement doesn't...

You are right! The microecm improvement doesn't require AVX512. I just checked in a version that should compile without -DAVX512_LASIEVE_SETUP in CFLAGS.

Probably -DAVX512_LASIEVE_SETUP will...
Forum: YAFU 2022-10-02, 19:31
Replies: 19
Views: 1,158
Posted By bsquared
I'll have to check, maybe I didn't protect all of...

I'll have to check, maybe I didn't protect all of the avx512 code. You need avx512 to get any speedup.
Forum: YAFU 2022-10-02, 15:25
Replies: 69
Views: 1,310
Posted By bsquared
Yes, probably. I suspect there is an exit()...

Yes, probably. I suspect there is an exit() buried in msieve for that condition... but I'm unable to check at the moment. There is a option -snfs_xover (I think?) which you should probably set...
Forum: YAFU 2022-10-02, 14:54
Replies: 5
Views: 136
Posted By bsquared
I think above 110 or 120 digits it will try to...

I think above 110 or 120 digits it will try to use the triple large prime variation. And it's possible there are intrinsics in use there that are not supported on some processors. It's not a path...
Forum: YAFU 2022-09-30, 13:55
Replies: 19
Views: 1,158
Posted By bsquared
I've been studying and working with the ggnfs...

I've been studying and working with the ggnfs lattice siever lately. The assembly code is pretty much unassailable, but there are C equivalents for most of the assembly code. I've made some...
Forum: Information & Answers 2022-09-29, 04:57
Replies: 36
Views: 1,126
Posted By bsquared
I haven't tried getting Cuda to work on my WSL...

I haven't tried getting Cuda to work on my WSL install, and now I'm thinking I don't want to try :max:

The stuff you are trying makes sense... wish I could help.
Forum: YAFU 2022-09-28, 21:43
Replies: 19
Views: 1,158
Posted By bsquared
I don't know, their ecm looks pretty Serious......

I don't know, their ecm looks pretty Serious... Edwards curves, bytecode driven...

But I've communicated with Paul Z. before, I can see if he's interested.
Forum: Number Theory Discussion Group 2022-09-28, 15:25
Replies: 53
Views: 12,780
Posted By bsquared
You should not be discouraged, your sieve does...

You should not be discouraged, your sieve does very well and demonstrates knowledge of several important optimizations. And the 2310 version is definitely the best way to go for a huge interval like...
Forum: YAFU 2022-09-28, 14:58
Replies: 19
Views: 1,158
Posted By bsquared
Just to note, building with CFLAGS=-DGGNFS_MPQS...

Just to note, building with CFLAGS=-DGGNFS_MPQS during the final make should enable the original mpqs code.
Forum: Number Theory Discussion Group 2022-09-28, 14:41
Replies: 53
Views: 12,780
Posted By bsquared
Yafu does this too. Multithreading in yafu works...

Yafu does this too. Multithreading in yafu works by taking the wheel concept a step further. We set up N sieve arrays, each array is a bitvector over one of the residue classes of the wheel...
Forum: Number Theory Discussion Group 2022-09-28, 13:10
Replies: 53
Views: 12,780
Posted By bsquared
The code at that link has different parameters...

The code at that link has different parameters for input to the segmented_bit_sieve_wheel function... it appears to want a start and stop interval, not stop and wheel_size parameters.

But anyway,...
Forum: YAFU 2022-09-28, 03:27
Replies: 19
Views: 1,158
Posted By bsquared
ggnfs improvements

As mentioned here (https://www.mersenneforum.org/showpost.php?p=614360&postcount=217), the microecm/tinyecm code in yafu appears to be helpful in the ggnfs lattice sievers as a replacement for mpqs...
Forum: Factoring 2022-09-28, 02:25
Replies: 216
Views: 33,599
Posted By bsquared
Agreed on the CADO note. However I think on most...

Agreed on the CADO note. However I think on most jobs, microecm is completely sufficient. Anything with mfba/r of 64 or less can use microecm as a complete replacement for mpqs. For really big...
Forum: Factoring 2022-09-27, 13:24
Replies: 7
Views: 574
Posted By bsquared
Thank you Jeff, I am happy to see this...

Thank you Jeff, I am happy to see this announcement!

I just wanted to add, Jeff did much more than just add significant optimizations to the low-level code in microecm. He also spotted an...
Forum: YAFU 2022-09-27, 01:10
Replies: 69
Views: 1,310
Posted By bsquared
No, not using the improved version that I know...

No, not using the improved version that I know of. It has been working for me just fine on several windows systems. Does it run single threaded?
Forum: Programming 2022-09-26, 17:28
Replies: 4
Views: 1,262
Posted By bsquared
Agreed. Yafu uses compressstoreu along with...

Agreed. Yafu uses compressstoreu along with popcount in its bucket sieve.
Forum: Hardware 2022-09-26, 14:05
Replies: 20
Views: 88,911
Posted By bsquared
Thank you for this! Lots of great info.

Thank you for this! Lots of great info.
Forum: YAFU 2022-09-26, 02:19
Replies: 69
Views: 1,310
Posted By bsquared
When there is no final newline then feof()...

When there is no final newline then feof() returns true regardless of if the previous fgets() read anything or not. So there has to be an extra check if the fgets() has meaningful data in it in...
Forum: YAFU 2022-09-23, 17:23
Replies: 69
Views: 1,310
Posted By bsquared
Here's ...

Here's (https://www.mersenneforum.org/showthread.php?t=16480)a thread for GPU-ECM.

And here's (https://www.mersenneforum.org/showthread.php?t=27103)a thread for Seth Trosi's improved CGBN...
Showing results 1 to 25 of 1000

 
All times are UTC. The time now is 22:52.


Fri Oct 7 22:52:56 UTC 2022 up 50 days, 20:21, 0 users, load averages: 0.97, 0.92, 0.85

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.

≠ ± ∓ ÷ × · − √ ‰ ⊗ ⊕ ⊖ ⊘ ⊙ ≤ ≥ ≦ ≧ ≨ ≩ ≺ ≻ ≼ ≽ ⊏ ⊐ ⊑ ⊒ ² ³ °
∠ ∟ ° ≅ ~ ‖ ⟂ ⫛
≡ ≜ ≈ ∝ ∞ ≪ ≫ ⌊⌋ ⌈⌉ ∘ ∏ ∐ ∑ ∧ ∨ ∩ ∪ ⨀ ⊕ ⊗ 𝖕 𝖖 𝖗 ⊲ ⊳
∅ ∖ ∁ ↦ ↣ ∩ ∪ ⊆ ⊂ ⊄ ⊊ ⊇ ⊃ ⊅ ⊋ ⊖ ∈ ∉ ∋ ∌ ℕ ℤ ℚ ℝ ℂ ℵ ℶ ℷ ℸ 𝓟
¬ ∨ ∧ ⊕ → ← ⇒ ⇐ ⇔ ∀ ∃ ∄ ∴ ∵ ⊤ ⊥ ⊢ ⊨ ⫤ ⊣ … ⋯ ⋮ ⋰ ⋱
∫ ∬ ∭ ∮ ∯ ∰ ∇ ∆ δ ∂ ℱ ℒ ℓ
𝛢𝛼 𝛣𝛽 𝛤𝛾 𝛥𝛿 𝛦𝜀𝜖 𝛧𝜁 𝛨𝜂 𝛩𝜃𝜗 𝛪𝜄 𝛫𝜅 𝛬𝜆 𝛭𝜇 𝛮𝜈 𝛯𝜉 𝛰𝜊 𝛱𝜋 𝛲𝜌 𝛴𝜎𝜍 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔