Glucas Source

nuggetprime · 2011-01-08, 20:45

Where can I get the newest 2.9.2 glucas source?
The oxixares server reports some error if I want to download it.

BTW,is Glucas actively developed at the moment?

nuggetprime · 2011-01-11, 16:23

rogue · 2011-01-11, 16:46

As far as I know, glucas is no longer actively developed. I've had no recent contact with Guillermo.

If you PM me your e-mail I can zip up the source and send it to you.

tichy · 2011-01-11, 18:53

Maybe we could put it on launchpad or sourceforge ?

rogue · 2011-01-12, 21:51

You can d/l glucas 2.9.2 source from here.

nuggetprime · 2011-01-13, 14:22

doesn't compile:

Code:

glucas.h:344: error: alignment of array elements is greater than element size

Maybe upload another 2.9.2 subversion?

rogue · 2011-01-13, 15:42

Quote:

Originally Posted by nuggetprime

doesn't compile:

Code:

glucas.h:344: error: alignment of array elements is greater than element size

Maybe upload another 2.9.2 subversion?

I have no clue what that error means. It is possible that gcc doesn't like some of the code in glucas.

nuggetprime · 2011-01-13, 16:29

I believe it has to do with the glucas version.
Could you up an earlier 2.9.2 shapshot?
BTW,does it compile fine on your PowerPC machine?

rogue · 2011-01-13, 17:20

Quote:

Originally Posted by nuggetprime

I believe it has to do with the glucas version.
Could you up an earlier 2.9.2 shapshot?
BTW,does it compile fine on your PowerPC machine?

I have nothing earlier than 2.9.2.

It did before, I haven't had to rebuild the library for a long time. I can try again later tonight.

nuggetprime · 2011-01-14, 13:56

don't mind,Glucas 2.9.0 can be downloaded from sourceforge and it's the same speed as the 2.9.2 version.
but something else:
Why is LLR about 5 times faster than Phrot for Phrot tests?
On my P4 3.2 GHZ:
LLR 3.8.4:

Code:

3*2^140000+1 is not prime.  Proth RES64: 2820B15AB5652D8A  Time : 29.538 sec.

Phrot 0.7.2:

Code:

3*2^140000+1 is composite LLR64=2820b15ab5652d8a. (e=0.03042 (0.0602137~3.30672e-16@0.000) t=156.42s)

Why is that?

rogue · 2011-01-14, 16:08

Quote:

Originally Posted by nuggetprime

don't mind,Glucas 2.9.0 can be downloaded from sourceforge and it's the same speed as the 2.9.2 version.
but something else:
Why is LLR about 5 times faster than Phrot for Phrot tests?
On my P4 3.2 GHZ:
LLR 3.8.4:

Code:

3*2^140000+1 is not prime.  Proth RES64: 2820B15AB5652D8A  Time : 29.538 sec.

Phrot 0.7.2:

Code:

3*2^140000+1 is composite LLR64=2820b15ab5652d8a. (e=0.03042 (0.0602137~3.30672e-16@0.000) t=156.42s)

Why is that?

Note that glucas does not have hand-tuned assembly. gwnum does, and it does for many different flavors of x86 CPUs. Based upon what I've seen, glucas is better for RISC architectures. phrot on PPC is similar in speed to pfgw on x86 for most numbers, but phrot on x86 is much slower. I don't recommend phrot on x86 since pfgw can be built on any x86 platform (with a little help).

2011-01-08, 20:45	#1
nuggetprime Mar 2007 Austria 302₁₀ Posts	Glucas Source Where can I get the newest 2.9.2 glucas source? The oxixares server reports some error if I want to download it. BTW,is Glucas actively developed at the moment?

2011-01-13, 14:22	#6
nuggetprime Mar 2007 Austria 302₁₀ Posts	doesn't compile: Code: glucas.h:344: error: alignment of array elements is greater than element size Maybe upload another 2.9.2 subversion? Last fiddled with by nuggetprime on 2011-01-13 at 14:23

2011-01-13, 16:29	#8
nuggetprime Mar 2007 Austria 2·151 Posts	I believe it has to do with the glucas version. Could you up an earlier 2.9.2 shapshot? BTW,does it compile fine on your PowerPC machine? Last fiddled with by nuggetprime on 2011-01-13 at 16:29

2011-01-14, 13:56	#10
nuggetprime Mar 2007 Austria 2×151 Posts	don't mind,Glucas 2.9.0 can be downloaded from sourceforge and it's the same speed as the 2.9.2 version. but something else: Why is LLR about 5 times faster than Phrot for Phrot tests? On my P4 3.2 GHZ: LLR 3.8.4: Code: 32^140000+1 is not prime. Proth RES64: 2820B15AB5652D8A Time : 29.538 sec. Phrot 0.7.2: Code: 32^140000+1 is composite LLR64=2820b15ab5652d8a. (e=0.03042 (0.0602137~3.30672e-16@0.000) t=156.42s) Why is that?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Mlucas/Glucas for LLR tests?	mdettweiler	Mlucas	4	2008-03-12 22:25
OS X Glucas build	rtharper	Software	3	2007-06-13 23:28
Glucas and GIMPS	optim	Software	6	2004-04-05 21:32
GLucas....	bayanne	Software	5	2003-08-15 16:14
Factoring with GLucas	bayanne	Software	10	2003-02-07 18:25

2011-01-11, 16:23	#2
nuggetprime Mar 2007 Austria 12E₁₆ Posts

2011-01-11, 16:46	#3
rogue "Mark" Apr 2003 Between here and the 6347₁₀ Posts	As far as I know, glucas is no longer actively developed. I've had no recent contact with Guillermo. If you PM me your e-mail I can zip up the source and send it to you.

2011-01-11, 18:53	#4
tichy Nov 2010 2²×19 Posts	Maybe we could put it on launchpad or sourceforge ?

2011-01-12, 21:51	#5
rogue "Mark" Apr 2003 Between here and the 18CB₁₆ Posts	You can d/l glucas 2.9.2 source from here.