srsieve/sr2sieve enhancements

[ValerieVonck]

Originally Posted by rogue

I have posted srsieve (v 1.0.6), sr2sieve (v 1.9.2), and sr1sieve (v 1.4.6)

These should build cleanly on both 64-bit Windows and OS X.

Rogue, in your links you posted a reference to sr1sieve 1.4.5?
When I go to home.roadrunner.com/~mrodenkirch/sr1sieve_1.4.6.zip => file not found?

regards
c

[rogue]

Originally Posted by CedricVonck

Rogue, in your links you posted a reference to sr1sieve 1.4.5?
When I go to home.roadrunner.com/~mrodenkirch/sr1sieve_1.4.6.zip => file not found?

regards
c

Sorry that is 1.4.5. I just remembered one problem I ran into with the asm module lookup-x86-64.S. mingw64 allows for ".align 16" but gcc (llvm) does not. I haven't tried with clang. Changing to ".align 15" works on gcc (llvm), but not mingw64.

[G30ffr3y]

I have modified sr2sieve to read from pfgw format files (and general ABC format). To do so, simply add the -w flag when calling from the command line. You can also use the -i flag along with it.

It has worked correctly in all tests thus far, and I cannot envisage any possible problems. There is a slight slowdown when reading a large number of sequences and/or terms (seconds rather than minutes), but it is less than the time to run srfile for the same file.

I'm not really sure if it should be 1.9.3, but there you go.

[rogue]

Originally Posted by G30ffr3y

I have modified sr2sieve to read from pfgw format files (and general ABC format). To do so, simply add the -w flag when calling from the command line. You can also use the -i flag along with it.

It has worked correctly in all tests thus far, and I cannot envisage any possible problems. There is a slight slowdown when reading a large number of sequences and/or terms (seconds rather than minutes), but it is less than the time to run srfile for the same file.

I'm not really sure if it should be 1.9.3, but there you go.

Thanks. I'll take a look. Presuming there are no problems, I'll post on my website.

[rogue]

I have posted sr2sieve 1.9.3 on my website. I made one change. I removed the -w option that was added. My version of sr2sieve will read the first line of the input file and use that to determine if it is an abcd or pfgw file.

[Xentar]

sr1sieve 1.4.5, windows x64
I think there is a double line break after a found factor:

138069900833 | 244*17^1974974+1

p=138069991991, 38610079 p/sec, 3542 factors, 0.1% done, 16 sec/factor, ETA 05 M
138197831713 | 244*17^1567610+1

p=138197918107, 39337674 p/sec, 3543 factors, 0.1% done, 16 sec/factor, ETA 05 M

[f1pokerspeed]

I'm using sr1sieve v1.4.6 and I'm having trouble trying to multithread the sieve.
I've tried putting the command -t 4 and its various forms at the start and end of the arguments and nothing seems to work - the program exits in roughly 0.2s. A quick flash, at the most. Without the command, the sieve works fine - so I am not sure what the problem is. Any ideas?

By the way, I've also tried to do this with v1.4.5.

[rogue]

Originally Posted by f1pokerspeed

I'm using sr1sieve v1.4.6 and I'm having trouble trying to multithread the sieve.
I've tried putting the command -t 4 and its various forms at the start and end of the arguments and nothing seems to work - the program exits in roughly 0.2s. A quick flash, at the most. Without the command, the sieve works fine - so I am not sure what the problem is. Any ideas?

By the way, I've also tried to do this with v1.4.5.

Which OS?

[f1pokerspeed]

Windows 7 x64 Home Premium.
CPU: Intel i5-2430M @ 2.4 -> 3.2 GHz w/ TB. 2 core with HT.
GPU Intel iGPU 2000

[pepi37]

sr2sieve 1.93 ( Win7 x64)

D:\PGEN>sr2sieve -t2 -A0 -A2 -i sr_10.abcd -p59004308141 -P15000000000000
sr2sieve: unknown option -- t

[henryzz]

I don't think multithreading is possible on windows.
From memory it uses pthreads which I think is officially linux only. It does seem to be possible to build on win32 now and there seem to be claims of having it working for 64-bit. You would have to compile sr1sieve yourself to take advantage of this on windows.