PFGW 4.0.4 (with gwnum v30.10) Released

[bbb120]

I failed download source code from https://sourceforge.net/p/openpfgw/code/HEAD/tree/ by tortoiseSVN!
why it can not download the zip file of source code directly just like GitHub ?

[VBCurtis]

Why don't you ask sourceforge, instead of this forum?

[bbb120]

Originally Posted by VBCurtis

Why don't you ask sourceforge, instead of this forum?

maybe someone download source code successful on this forum !

[rebirther]

Originally Posted by bbb120

maybe someone download source code successful on this forum !

You can download the source in a zip file on the upper right site (download snapshot)

[bbb120]

Originally Posted by rebirther

You can download the source in a zip file on the upper right site (download snapshot)

ok thank you very much!

[kijinSeija]

Hi I have a question :

How to get the decimal expression on the console, but only PRP ? I tested with ./pfgw -od input.txt but I have all the numbers but I just want the PRP. This is possible ?

Thanks :)

[rogue]

Originally Posted by kijinSeija

How to get the decimal expression on the console, but only PRP ? I tested with ./pfgw -od input.txt but I have all the numbers but I just want the PRP. This is possible ?

Using -od only outputs the decimal expansion to the console, but you should be able to use > on the command line to redirect the output to a file, e.g. pfgw64 -od input.txt > decimal.txt. decimal.txt will only have PRPs in it.

[SethTro]

Originally Posted by paulunderwood

In my experience with multithreaded gwnum, there is little to be gained for numbers less than ~800k digits. YMMV.

I'm curious if there's any follow up on this. I'm seeing multiple threads run but I get the same or worse timing when I use -T2 and -T4
With PFGW Version 4.0.1.64BIT.20191203.x86_Dev [GWNUM 29.8]

I ran the same 185,000 digit number with -T1, -T2, and -T4 and it took the same wall time but twice and four times as much usertime time (and I saw more cores at 100% utilization)

$ time pfgw64 -T1 -Cverbose -f0 timing.abc 
PFGW Version 4.0.1.64BIT.20191203.x86_Dev [GWNUM 29.8]

(9^25966-7)/2 is composite: RES64: [10E24C23A947A45F] (0.9759s+0.0001s)                                    
(9^51862-7)/2 is composite: RES64: [F2C9E5065BF0C831] (4.4636s+0.0002s)                                    
(9^60760-7)/2 is composite: RES64: [9F8CAE8CB2B113E5] (6.5107s+0.0003s)                                    
(9^69322-7)/2 is composite: RES64: [B984A003CE68ADF5] (9.2732s+0.0003s)                                    
(9^78004-7)/2 is composite: RES64: [F493F4ACB7AE8514] (10.4510s+0.0004s)                                    
(9^86270-7)/2 is composite: RES64: [7FE53104F6AE236F] (13.3946s+0.0004s)                                    
(9^103415-7)/2 is composite: RES64: [DD34C02FDF4073D0] (16.1420s+0.0005s)                                    
(9^111653-7)/2 is composite: RES64: [C527F2956453C4C2] (20.7393s+0.0005s)                                    
(9^120498-7)/2 is composite: RES64: [C8587138F5CC8CD5] (26.0815s+0.0006s)                                    
(9^129382-7)/2 is composite: RES64: [17B4C355EC1755C0] (28.6893s+0.0006s)                                    
(9^138052-7)/2 is composite: RES64: [6D8974E2615AF80D] (36.5331s+0.0007s)                                    
(9^146706-7)/2 is composite: RES64: [F6E98B47CEE0442A] (37.2141s+0.0008s)                                    
(9^155454-7)/2 is composite: RES64: [33AC0CC2482A9E55] (39.7199s+0.0009s)                                    
(9^164094-7)/2 is composite: RES64: [19E66BBE89C9D21F] (47.3960s+0.0009s)                                    
(9^172705-7)/2 is composite: RES64: [1BDFC63E511C7E0D] (57.2845s+0.0011s)                                    
(9^181162-7)/2 is composite: RES64: [1C1F11811928AAEE] (55.6043s+0.0010s)                                    
(9^189780-7)/2 is composite: RES64: [C5E799687A4AA318] (60.1954s+0.0011s)                                    
real	8m54.152s
user	8m54.168s


# With -T4 It's slower and takes multiple cores
$ time pfgw64 -T4 -Cverbose -f0 timing.abc 
PFGW Version 4.0.1.64BIT.20191203.x86_Dev [GWNUM 29.8]

(9^25966-7)/2 is composite: RES64: [10E24C23A947A45F] (0.9269s+0.0001s)                                    
(9^51862-7)/2 is composite: RES64: [F2C9E5065BF0C831] (8.4412s+0.0002s)                                    
(9^60760-7)/2 is composite: RES64: [9F8CAE8CB2B113E5] (12.0932s+0.0003s)                                    
(9^69322-7)/2 is composite: RES64: [B984A003CE68ADF5] (17.1319s+0.0003s)                                    
(9^78004-7)/2 is composite: RES64: [F493F4ACB7AE8514] (19.1786s+0.0003s)                                    
(9^86270-7)/2 is composite: RES64: [7FE53104F6AE236F] (23.3000s+0.0004s)
(9^103415-7)/2 is composite: RES64: [DD34C02FDF4073D0] (16.4805s+0.0005s)                                    
(9^111653-7)/2 is composite: RES64: [C527F2956453C4C2] (34.9898s+0.0005s)                                    
(9^120498-7)/2 is composite: RES64: [C8587138F5CC8CD5] (37.4123s+0.0009s)                                    
(9^129382-7)/2 is composite: RES64: [17B4C355EC1755C0] (34.9322s+0.0008s)                                    
(9^138052-7)/2 is composite: RES64: [6D8974E2615AF80D] (47.8086s+0.0008s)                                    
(9^146706-7)/2 is composite: RES64: [F6E98B47CEE0442A] (46.4387s+0.0008s)                                    
(9^155454-7)/2 is composite: RES64: [33AC0CC2482A9E55] (52.2789s+0.0009s)                                    
(9^164094-7)/2 is composite: RES64: [19E66BBE89C9D21F] (63.7558s+0.0009s)                                    
(9^172705-7)/2 is composite: RES64: [1BDFC63E511C7E0D] (65.7027s+0.0010s)                                    
(9^181162-7)/2 is composite: RES64: [1C1F11811928AAEE] (64.5466s+0.0010s)                                    
(9^189780-7)/2 is composite: RES64: [C5E799687A4AA318] (84.8338s+0.0011s)

real	12m26.290s
user	28m56.258s

[rogue]

Originally Posted by SethTro

I'm curious if there's any follow up on this. I'm seeing multiple threads run but I get the same or worse timing when I use -T2 and -T4
With PFGW Version 4.0.1.64BIT.20191203.x86_Dev [GWNUM 29.8]

I ran the same 185,000 digit number with -T1, -T2, and -T4 and it took the same wall time but twice and four times as much usertime time (and I saw more cores at 100% utilization)

I think this is a question for George. The gwnum library is starting and managing the additional threads.

[Prime95]

Depending on the CPU, multithreading isn't even possible until around 250000 bits and isn't really helpful until around 1.3 million bits or more.

For these small numbers you are better off running one pfgw instance for each core.

[numberMX]

When I go to the Sourceforge page and download `pfgw_mac_4.0.1.7z`, then extract and run it...

> ./pfgw64 -v
PFGW Version 3.8.1.64BIT.20160603.Mac_Dev [GWNUM 28.6]

...it downloads version 3.8.1, not 4.0.1. It seems like the 4.0.1 download for Mac does not actually have version 4.0.1 in it.