![]() |
|
|
#1 |
|
Oct 2017
++41
1758 Posts |
When I set my Atom D2700 to "what make sense" I get a D-LL, which doesn't make sense because a job would take about a year. So i let it do PM1-L (It even found a factor! https://www.mersenne.org/report_expo...xp_lo=84476921 ) which takes about 9 days per exponent.
So, what would really make sense? I've never tried PRP-CF or ECM. The computer is always on and since it doesnt support speed step it almost doesn't make a difference on energy consumption whether it is running at 100% load or idle. Could even a raspberry pi be good for anything? |
|
|
|
|
|
#2 | |
|
Banned
"Luigi"
Aug 2002
Team Italia
130116 Posts |
Quote:
I am using mlucas with my Raspberry PI 3 (and Gentoo Linux 64 bit). A single DC exponent is done in about 1.5 months. But I am addicted to GIMPS since 1987... Last fiddled with by ET_ on 2017-12-17 at 18:00 |
|
|
|
|
|
|
#3 |
|
Bamboozled!
"๐บ๐๐ท๐ท๐ญ"
May 2003
Down not across
2E1416 Posts |
|
|
|
|
|
|
#4 | |
|
Sep 2002
Database er0rr
468510 Posts |
Quote:
Last fiddled with by paulunderwood on 2017-12-17 at 17:58 |
|
|
|
|
|
|
#5 | |
|
∂2ω=0
Sep 2002
Repรบblica de California
22·2,939 Posts |
Quote:
Mlucas has always been buildable in generic-C-code mode on Pi and such - the main thing that has changed in the past few months is that I released a version with inline-asm targeting the ARM v8 128-bit SIMD instruction set. See the related thread in the Software -> Mlucas subforum. (Page 11 of the thread is where the release is announced and timings form users on a variety of ARMv8 arches follow.) The gain from the SIMD code on ARMv8 is modest compared to (say) te SSE2 code running on my Core2 system, which I believe to be due to the x86 family having dedicated hardware-arithmetic functional units for SIMD versus scalar (64-bit-double for the former, 80-bit for the latter), compared to the low-power-optimized ARM having both scalar and SIMD instructions share a common set of functional units. (Someone please correct me if my surmise here is incorrect.) |
|
|
|
|
|
|
#6 |
|
Banned
"Luigi"
Aug 2002
Team Italia
5×7×139 Posts |
|
|
|
|
|
|
#7 |
|
"Composite as Heck"
Oct 2017
2·52·19 Posts |
Raspberry pi 3 (and anything else with an A53 chip) seems a good "minimum recommended hardware" watermark for mlucas, boards with older chipsets are probably best suited to something like factoring. I tried mlucas on an A7 but it didn't seem worth it.
If you have a raspberry pi 3 and want to use mlucas, me and ET have successfully tried it with this distro. |
|
|
|
|
|
#8 | |
|
"Thiago Arreguy"
Jan 2018
Brazil
1 Posts |
I try run George Woltman's Prime95 program (a.k.a. mprime for the linux version) in a Raspberry pi 3 with gentoo-on-rpi3-64bit and not works: "mprime cannot execute binary file: Exec format error".
Its only possible run mlucas on Raspberry pi 3 ? Quote:
|
|
|
|
|
|
|
#9 |
|
Sep 2002
Database er0rr
124D16 Posts |
|
|
|
|
|
|
#10 |
|
Sep 2003
2×5×7×37 Posts |
If a very slow machine used very little electricity and cost very little, then potentially every job could be for it. As long as it's not so utterly slow that you can't get enough useful work out of it before newer, more efficient technology gets invented.
I don't know if any of the current crop of ARM machines or Atom machines fit the bill. I suspect not, at least not yet. |
|
|
|
|
|
#11 |
|
Oct 2017
++41
53 Posts |
|
|
|
|
![]() |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Slow Down? | R.D. Silverman | GMP-ECM | 55 | 2011-10-16 17:28 |
| GMP ECM slow on my laptop? | R.D. Silverman | GMP-ECM | 9 | 2011-02-17 14:57 |
| How hot is too hot? Slow is too slow? | petrw1 | Hardware | 13 | 2008-11-10 23:25 |
| Really slow machines? | Doorbasher | Hardware | 5 | 2004-08-23 22:18 |
| Factoring so slow?? | xtreme2k | Software | 7 | 2003-03-13 04:12 |