Project Links
 

[B]Dat File[/B]
[URL="http://mersenneforum.org/base5/"]sieve files[/URL]

[B]Sieve Program[/B]
[URL="http://sites.google.com/site/geoffreywalterreynolds/programs/sr5sieve"]sr5sieve[/URL]

[B]Factor Submission[/B]
[URL="http://www.sr5.psp-project.de/sieveimport.php"]factor submission[/URL]

[B]Testing Programs[/B]
[URL="http://jpenne.free.fr/Development/"]LLR[/URL]
[URL="http://sourceforge.net/projects/openpfgw/"]PFGW[/URL]
[URL="http://home.roadrunner.com/~mrodenkirch/home/Phrot.html"]phrot[/URL]
[URL="http://tools.1up.no/psp/prpclient/"]PRPnet 2.4.7 [/URL] Replace programs with the following:
[URL="http://home.roadrunner.com/~mrodenkirch/prpnet_4.2.0.zipPRPnet 4.2.0 [/URL]


[B]Stats[/B]
[URL="http://www.sr5.psp-project.de/"]http://sr5.psp-project.de/[/URL]

Links here should also be updated.

[QUOTE=em99010pepe;216114]Links here should also be updated.[/QUOTE]

You are right, but I am not sure what direction this project is headed. The last few days it seems things are progressing again but in several different directions.

[quote=Beyond;216118]You are right, but I am not sure what direction this project is headed. The last few days it seems things are progressing again but in several different directions.[/quote]

Simple. For manual LLR reservations people should use LLR381 client from Pénne. For automated Network LLR client people should run LLRnetV08. For sieve people should run sr5sieve from Geoff. Someone should host the sr5data.txt file, if you want I can host it on google site.

[QUOTE=em99010pepe;216122]Simple. For manual LLR reservations people should use LLR381 client from Pénne. For automated Network LLR client people should run LLRnetV08. For sieve people should run sr5sieve from Geoff. Someone should host the sr5data.txt file, if you want I can host it on google site.[/QUOTE]

Joe_O already has the updated sieve file hosted, I think it will be updated more often now. There was/is no confusion in my mind about what programs to use, more on what was happening with the actual direction of the project (hint see reservation thread).

[quote=Beyond;216127]Joe_O already has the updated sieve file hosted, I think it will be updated more often now. There was/is no confusion in my mind about what programs to use, more on what was happening with the actual direction of the project (hint see reservation thread).[/quote]

The project has a dedicated server but for Lennart I think it is easier to point his machines to the prpnet server hosted by him. I think you made the right call by hidden his server address but I understand his concern on sharing the address.
Anyway, you can ask Lars to host a prpnet server too where he can easily set up one and have coordinate stats page.

I think the project should have three directions: manual sieve, manual LLR and automated LLR with llrnet and prpnet servers.

[QUOTE=em99010pepe;216134]...
I think the project should have three directions: manual sieve, manual LLR and automated LLR with llrnet and prpnet servers.[/QUOTE]

What about a PRPnet like sieve system that hands out 10G assignments? We might be able to get more people sieving. Nothing elaborate, just as simple as possible.

BOINC

The link for phrot in the first post is wrong. It should be this: [url]http://home.roadrunner.com/~mrodenkirch/home/Phrot.html[/url]

I think maybe it should be completely removed along with others since there is no more individual k reservations.

[QUOTE=Beyond;216193]I think maybe it should be completely removed along with others since there is no more individual k reservations.[/QUOTE]

I guess that depends upon the programs that are supported by the project. If it is LLRNET only, then I would agree.

Note that the following is my opinion. I haven't fully followed all of the discussions and I would guess that there were discussions outside of this forum regarding the choice of LLRNET. It is my opinion that PRPNet is a much better choice than LLRNet for this project. Why? It opens the door to more people participating. It has native support for Sierpinski/Riesel type projects. It is not married to a specific version of LLR.

[QUOTE=rogue;216190]The link for phrot in the first post is wrong. It should be this: [url]http://home.roadrunner.com/~mrodenkirch/home/Phrot.html[/url][/QUOTE]

[QUOTE=Beyond;216193]I think maybe it should be completely removed along with others since there is no more individual k reservations.[/QUOTE]

I think that the links to PRP and Phrot should be removed, leaving only LLR in all it's different manifiestations(AFAIK PRPnet uses LLR for both Riesel and Sierpinski. If not, someone please correct me.) so that the residues are all compatible.

As far as the LLRnet vs PRPnet debate, it may become a moot point. Since LLRnet server is no longer supported, it may break when LTD updates to a newer Linux. Then we will either have to scramble for a new host, or change over to PRPnet ASAP. Wouldn't it be better to be prepared?

[QUOTE=Joe O;216212]I think that the links to PRP and Phrot should be removed, leaving only LLR in all it's different manifiestations(AFAIK PRPnet uses LLR for both Riesel and Sierpinski. If not, someone please correct me.) so that the residues are all compatible.

As far as the LLRnet vs PRPnet debate, it may become a moot point. Since LLRnet server is no longer supported, it may break when LTD updates to a newer Linux. Then we will either have to scramble for a new host, or change over to PRPnet ASAP. Wouldn't it be better to be prepared?[/QUOTE]

Absolutely agreed on PRP. PRPNet uses LLR only for base that are powers of 2. It uses PFGW for other bases. phrot targets non-x86 users.

When using LLR 3.8.x, it does have the advantage of doing a first-pass primality test, which PRPNet does not take advantage of. I could modify PRPNet to support primality testing with LLR, but that would force EVERYONE to use LLR 3.8.x instead of phrot or PFGW because the residues would not match. Right now, the residues for PRP tests using PFGW, LLR, and phrot residues match for all bases (except base 2). Note that base 2 requires either phrot or LLR. PFGW is only used by base 2 for GFN divisibility testing. What I'm basically saying is that by limiting PRPNet to using LLR (or phrot on non-x86) for base 2 only provides the maximum amount of forward and backward compatibility. The only limit that PRPNet has is with base 2 where c=-1. phrot cannot do a primality test on numbers of this form, which is what LLR does.

[QUOTE=Joe O;216212]I think that the links to PRP and Phrot should be removed, ....[/QUOTE]

[QUOTE=rogue;216230]Absolutely agreed on PRP. ....[/QUOTE]

I stand corrected. I would now ask for only the PRP link to be removed. I did not know that Phrot was needed for non x-86 users to participate.

Why is it that this link: [URL="http://sr5.psp-project.de/r5stats.html"]http://sr5.psp-project.de/r5stats.html[/URL] shows 154 k's remaining when it should rightfully show 153 k's remaining?

AFAICS, it may be because k=193786 is still remaining.

On an additional question, did some of the faster sieve machines reach p=1M/sec, now that the sievefile went down to 208 k's?

Take care

KEP!

[QUOTE=KEP;217278]Why is it that this link: [URL="http://sr5.psp-project.de/r5stats.html"]http://sr5.psp-project.de/r5stats.html[/URL] shows 154 k's remaining when it should rightfully show 153 k's remaining?

AFAICS, it may be because k=193786 is still remaining.

On an additional question, did some of the faster sieve machines reach p=1M/sec, now that the sievefile went down to 208 k's?

Take care

KEP![/QUOTE]

It will reflect the 2 primes found within 6 hours of each other once the 2nd prime is verified on Top 5000 page

I would consider getting back on this project, but since it only supports LLR residues...

[QUOTE=rogue;217290]I would consider getting back on this project, but since it only supports LLR residues...[/QUOTE]

LLR residue for base-5 = PRP residue.

The stats page is refreshing according to the timestamp, but the returned values have stopped accruing about 24 hours ago
[url]http://www.sr5.psp-project.de/s5stats.html[/url]

Can ltd give it a kick?

Still stuck.

Still stuck.

Drat. I was enjoying this.
I just killed 17 clients as I have no visual confirmation that I am not sending my results to /dev/null :down:
I'll try to bring them back sometime in the future.

Thanks to Lennart for the sr5 prpnet status link:
[URL]http://www.psp-project.de:8105/all.html[/URL]
whilst the stats page is down, a participant can check that the results are still being collected there.
I will note this for next time.

Sorry for not answering before.
I had not noticed this thread till 20 minutes ago. My fault.
So far I can see that the results are transfered into the main DB but then
they are not used in the stats calculation.
All these tests have still the status transfered and waiting for stats calculation.

So the only thing I have to find out is why are they not seen by the stats calculation mechanism.

Keep you informed.

Think I found it.
We will see in 45 minutes.

Thing is that there is a state in the main db which shows some info about the distribution of a test.
When I made the import for the new tests on the server I did set this status on a wrong value. Normally this would make no difference but as there is an old optimization in one of the queries which needed some exact values for this status field instead of a range all the new tests were hidden from the stats engine.

You fixed it. Thanks. :smile:

So is [URL="http://www.sr5.psp-project.de/r5stats.html"]this page[/URL] ever going to be updated to show recent primes?