750-800T complete. Factors available [URL="http://www.amphoria.co.uk/dl/mega"]here[/URL].

550-750 almost finished

reserving 850-1050

[URL="http://www.sendspace.com/file/di85rq"]550-750[/URL]

801-850T complete. Factors available [URL="http://www.amphoria.co.uk/dl/mega"]here[/URL]. Note I combined with 800-801T for processing so both ranges appear in the one file.

Taking 1050-1100T.

I have made a new sieve file available [URL="http://www.amphoria.co.uk/dl/mega"]here[/URL] sieved to p=850T. There are 65,704,418 k's remaining.

1050-1100T complete. Factors available [URL="http://www.amphoria.co.uk/dl/mega"]here[/URL].

Taking 1100-1150T.

Taking 1150-1250T.

1100-1150T complete. Factors available [URL="http://www/amphoria.co.uk/dl/mega"]here[/URL].

Taking 1250-1300T.

[URL="http://www.amphoria.co.uk/dl/mega"]1200-1250T[/URL] complete.

[URL="http://www.amphoria.co.uk/dl/mega"]1150-1200T[/URL] complete.

Taking 1300-1350T.

Taking 1350-1500T.

[URL="http://www.amphoria.co.uk/dl/mega"]1250-1350T[/URL] complete.

Taking 1500-1800T.

[URL="http://www.amphoria.co.uk/dl/mega"]1350-1500T[/URL] complete.
[URL="http://www.amphoria.co.uk/dl/mega"] [/URL]

Taking 1800-2100T.

[URL="http://www.amphoria.co.uk/dl/mega"]1500-1800T[/URL] complete.

You might want to change the 5 occurrences of "emirp" to "prime" in the first post of this thread.

[QUOTE=gd_barnes;261811]You might want to change the 5 occurrences of "emirp" to "prime" in the first post of this thread.[/QUOTE]
Done. For some reason, it wasn't changed back after this event:
[URL]http://www.mersenneforum.org/showthread.php?t=15478[/URL]

Dougal, could you give us an update of your progress? You haven't posted a status update since early March, so I'm wondering how that 850T-1050T range is coming along.

1800-2100T complete. I will upload the factors [URL="http://www.amphoria.co.uk/dl/mega"]here [/URL]this evening.

Taking 2100-2400T.[U][/U][URL="http://www.amphoria.co.uk/dl/mega"][/URL]

I have uploaded a new file sieved to p=2100T (except 850-1050T) to the usual [URL="http://www.amphoria.co.uk/dl/mega"]place[/URL].

It is currently taking about 2 weeks to sieve 100T on one 2.4 GHz Core 2 core and about 180,000 unique factors are removed.

Taking 2400-2700T.

[URL="http://www.amphoria.co.uk/megabit-twin/"]2100-2400T[/URL] complete.

Oddball - I have redesigned my website. The archive can now be accessed from [URL="http://www.amphoria.co.uk/megabit-twin/"]http://www.amphoria.co.uk/megabit-twin[/URL]/. Please can you change the link at the top of this thread.

Thanks.

[QUOTE=amphoria;263779][URL="http://www.amphoria.co.uk/megabit-twin/"]2100-2400T[/URL] complete.

Oddball - I have redesigned my website. The archive can now be accessed from [URL="http://www.amphoria.co.uk/megabit-twin/"]http://www.amphoria.co.uk/megabit-twin[/URL]/. Please can you change the link at the top of this thread.

Thanks.[/QUOTE]
Done. I'd also like to mention that if Dougal doesn't give us an update of his progress by the end of the month, his 850T-1050T range will be released.

[URL="http://www.amphoria.co.uk/megabit-twin/1050-5000t/"]2400-2700T[/URL] complete.

As it is almost the end of the month, I will take 850-1050T. If Dougal posts an update in the next couple of days, then I will stop.

Taking 2700-3000T.

[URL="http://www.amphoria.co.uk/megabit-twin/45-1050t/"]850-1050T[/URL] complete.

I have also uploaded a new sieve file complete to p=2700T [URL="http://www.amphoria.co.uk/megabit-twin/"]here[/URL]. There are 61,528,767 k's remaining.

Taking 3000-3300T.

[URL="http://www.amphoria.co.uk/megabit-twin/1050-5000t/"]2700-3000T[/URL] complete.

[URL="http://www.amphoria.co.uk/megabit-twin"]3000-3200T[/URL] complete.

Taking 3300-3500T.

[URL="http://www.amphoria.co.uk/megabit-twin"]3200-3300T[/URL] complete.

Taking 3500-3600T.

[URL="http://www.amphoria.co.uk/megabit-twin"]3300-3500T[/URL] complete.

Taking 3600-3800T.

[URL="http://www.amphoria.co.uk/megabit-twin"]3500-3600T[/URL] complete.

Taking 3800-3900T.

I uploaded a new sieve file sieved to p=3600T to the [URL="http://www.amphoria.co.uk/megabit-twin"]archive site[/URL]. There are 60,550,811 k's remaining.

Taking 3900-4000T.

Taking 4000T - 4800T

[QUOTE=Puzzle-Peter;269917]Taking 4000T - 4800T[/QUOTE]

Welcome to the project, Peter.

[QUOTE=amphoria;269933]Welcome to the project, Peter.[/QUOTE]

Thanks!

After doing some calculations of how long it would take me to find a world record twin, I decided to add my efforts to yours. Hopefully others will join in...

Taking 4800T - 5000T

[QUOTE=monst;270053]Taking 4800T - 5000T[/QUOTE]
Welcome to TPS!

I've added your reservation to the first post in this thread.

[URL="http://www.amphoria.co.uk/megabit-twin"]3600-3800T[/URL] complete.

Taking 5000-5200T.

Taking 5200-5300T.

[URL="http://www.amphoria.co.uk/megabit-twin"]3800-3900T[/URL] complete.

4200T-5000T is finished. The factor file is too big to upload. Could you privide an email address where I can send it?

Is that range correct? I am running 4800T to 5000T. Should I stop my run?

[QUOTE=monst;270645]Is that range correct? I am running 4800T to 5000T. Should I stop my run?[/QUOTE]

Sorry, I was off by 200T. I did 4000T to 4800T as reserved.

[QUOTE=Puzzle-Peter;270642]4200T-5000T is finished. The factor file is too big to upload. Could you privide an email address where I can send it?[/QUOTE]

Hopefully this is a typo as you reserved 4000T-4800T. You can upload the factor file to [url]www.sendspace.com[/url] and then post the link here.

OK, here we go 4000T - 4800T: [URL]http://www.sendspace.com/file/1k9q6t[/URL]

[URL="http://www.amphoria.co.uk/megabit-twin"]3900-4000T[/URL] complete

Taking 5300-5500T.

Peter - Are you planning on reserving another range? Someone calculated that the target was somewhere around 100000T. You will have noted that factors are still being found frequently.

[QUOTE=amphoria;270653]
Peter - Are you planning on reserving another range? Someone calculated that the target was somewhere around 100000T. You will have noted that factors are still being found frequently.[/QUOTE]


Not right now. I found about one factor in 10s. So optimum sieve depth is still far from where we are now. Without a good number of people helping this will take several years, as we'll need to test millions of candidates at 20(?) minutes per test (haven't done any tests yet, so I'm guessing here).

Plus NewPGen is not very comfortable for me (I'm using Linux command line versions). I'll try tpsieve to compare speed and user friendlyness.

[QUOTE=Puzzle-Peter;270654]Plus NewPGen is not very comfortable for me (I'm using Linux command line versions). I'll try tpsieve to compare speed and user friendlyness.[/QUOTE]

tpsieve complains about the range of k, otherwise I would be using tpsieve. :down:

[QUOTE=amphoria;270656]tpsieve complains about the range of k, otherwise I would be using tpsieve. :down:[/QUOTE]

I asked Ken about this and he told me this check was included to avoid using too much memory. You can simply kick out the lines checking for kmax-kmin<3*2^36 and compile it. Nowadays the machines should be able to cope with the large ranges.

I asked for the sources via the Prime Grid forum but haven't got an answer yet. When I try to compile what is included in the tpsieve package, I get an error because some included files are not found. I am not a programmer, so maybe this is easy to fix. I'll try this next week on a Linux box.

[QUOTE=Puzzle-Peter;270660]I asked Ken about this and he told me this check was included to avoid using too much memory. You can simply kick out the lines checking for kmax-kmin<3*2^36 and compile it. Nowadays the machines should be able to cope with the large ranges.

I asked for the sources via the Prime Grid forum but haven't got an answer yet. When I try to compile what is included in the tpsieve package, I get an error because some included files are not found. I am not a programmer, so maybe this is easy to fix. I'll try this next week on a Linux box.[/QUOTE]

I can get it to compile on Linux by changing to the pps dir and running make-bins.sh. This stops it from trying to build the BOINC version. I got some errors regarding gnu/stubs-32.h not being found because I am compiling on 64-bit Linux, but these are harmless. It just means that the 32-bit version is not built. This is on Ubuntu 11.04 BTW.

However, that's where the good news stops. After removing the check and compiling, I get a runtime error "free(): invalid pointer" just after it prints the version of gcc it was compiled with.

[QUOTE=amphoria;270665]However, that's where the good news stops. After removing the check and compiling, I get a runtime error "free(): invalid pointer" just after it prints the version of gcc it was compiled with.[/QUOTE]

Hmmm.
I'll give it a spin next week when I'm on Linux again. I also contacted Ken again - waiting for his answer...

I remember having issues with tpsieve before because my gcc version was too old. Could this be your problem?

I was actually wondering it it was too new - it's version 4.5. I intend to try v4.4 which is also available with Ubuntu 11.04. Ken's binaries are compiled with v4.3 but this has been removed from the standard Ubuntu 11.04 install.

I haven't had much joy getting a working Linux version, but I did manage to get it to compile on Windows using MinGW. It loaded the sieve file OK but crashed soon after starting the sieve. As this was the Win-32 version I am guessing that it ran out of memory.

I tried a cut down sieve file to prove that my binary worked and it revealed another limitation of tpsieve. The error message was "twin searching not implemented for such a small P". The error message is a bit misleading as it is nothing to with pmin or pmax. It is complaining that nstep is not greater than 2 - it is in fact equal to 2.

I don't know enough about the algorithm to sure about this, but as nstart=1000000, it might only work with multiple n's. I confirmed that the error message is genuine by duplicating it with Ken's Linux binary and the cut down sieve file.

[QUOTE=Puzzle-Peter;270654]Plus NewPGen is not very comfortable for me (I'm using Linux command line versions). I'll try tpsieve to compare speed and user friendlyness.[/QUOTE]

I never answered your concern about using NewPGen on Linux. It is possible to drive it from the command line which makes it no harder to use than tpsieve. I have attached a shell script which can be called as

./newpgen.sh sievefile.txt 5500

The last parameter is the stopping p in T. If you want the output file different from the input file change -osp. The only parameter you don't appear to be able to set from the command line is the flag to use 2 save files, but as this is stored in NewPGen.ini this only needs to be set once. -om sets the minutes between saving the file.

[QUOTE=amphoria;270731]I never answered your concern about using NewPGen on Linux. It is possible to drive it from the command line which makes it no harder to use than tpsieve. I have attached a shell script which can be called as

./newpgen.sh sievefile.txt 5500

The last parameter is the stopping p in T. If you want the output file different from the input file change -osp. The only parameter you don't appear to be able to set from the command line is the flag to use 2 save files, but as this is stored in NewPGen.ini this only needs to be set once. -om sets the minutes between saving the file.[/QUOTE]

Thanks a lot for this. One more question: How does stopping and restarting work? Do I need to modify the command or does it start from where it stopped automatically?

I just tried my old Xubuntu, I couldn't even get it to compile. Couldn't find main.c, clock.c and the others...

[QUOTE=Puzzle-Peter;270744]Thanks a lot for this. One more question: How does stopping and restarting work? Do I need to modify the command or does it start from where it stopped automatically?[/QUOTE]

If you run it with the input file and output file pointing to the same sieve file (the default in my script), then it will automatically start from where it stopped. It updates the value of p in the header of the sieve file when it stops. If you overwrite the sieve file like this, then you definitely want to set the option to use two save files. Note that you have to give both parameters again when you restart.

[URL="http://www.amphoria.co.uk/megabit-twin"]5000-5200T[/URL] complete.

Taking 5500-5700T.

[URL="http://www.amphoria.co.uk/megabit-twin"]5200-5300T[/URL] complete.

Taking 5700-5800T.

Taking 5800-6000T.

[URL="http://www.amphoria.co.uk/megabit-twin"]5300-5500T[/URL] complete.

Taking 6000-6400T.

amphoria, you might want to update the file onyour website, the situation have evolved since the last time you updated it.
and as i tried to run 6400T-6405T, newpgen complain that n is not increasing :
Opened the input file - counting the number of n's and finding the range of n...
The numbers in the file are not strictly increasing : I saw 1000000 followed by 1000000!

scratch my last remark, i had forgot to hit enter at the end of my header line.
also it seems that you can't use tpsieve yet,
[code]
tpsieve -p6400T -P6405T -i 0-120T.txt -m2048

tpsieve version cuda-0.2.3b (testing)
Found K's from 926835 to 119999996287275.
Found N's from 1000000 to 1000001.
kmax-kmin < 3*2^36 is required
[/code]

[QUOTE=firejuggler;272152]amphoria, you might want to update the file onyour website, the situation have evolved since the last time you updated it.
[/QUOTE]

I was waiting for monst to complete 4800-5000T before updating the file, but as this has taken longer than I expected I will update it tomorrow evening when I will have finished 5700-5800T.

Taking 6405-6600T.

[QUOTE=firejuggler;272159]scratch my last remark, i had forgot to hit enter at the end of my header line.
also it seems that you can't use tpsieve yet,
[code]
tpsieve -p6400T -P6405T -i 0-120T.txt -m2048

tpsieve version cuda-0.2.3b (testing)
Found K's from 926835 to 119999996287275.
Found N's from 1000000 to 1000001.
kmax-kmin < 3*2^36 is required
[/code][/QUOTE]

Look earlier in this thread for the results of our investigations with tpsieve. We got further than that but not a working version.

[QUOTE=amphoria;272173]I was waiting for monst to complete 4800-5000T before updating the file, but as this has taken longer than I expected I will update it tomorrow evening when I will have finished 5700-5800T.[/QUOTE]


I just checked my machine. It is at 4984T. I'll send in the results as soon as it is finished. You may want to wait just a bit longer to update the file.

-- Rich

thanks for the update monst.
and too bad for tpsieve.

@amphoria

I'm moving this sieving to a stronger machine and/or I'll break up the problem to get the results by tomorrow morning.

-- Rich

[QUOTE=monst;272187]@amphoria

I'm moving this sieving to a stronger machine and/or I'll break up the problem to get the results by tomorrow morning.

-- Rich[/QUOTE]

Excellent - thanks for the update.

4800 - 5000T complete

[url]http://www.sendspace.com/file/grif80[/url]

[URL="http://www.amphoria.co.uk/megabit-twin"]5700-5800T[/URL] complete.

The latest sieve file sieved to p=5800T is available [URL="http://www.amphoria.co.uk/megabit-twin"]here[/URL]. There are 58,982,026 k's remaining.

here it is (6400-6405T)

Time for a little break from CRUS, so:

I'm taking 6600T-7600T ETA ~26 days :smile:

Regards

Kenneth

[URL="http://www.amphoria.co.uk/megabit-twin"]5800-6000T[/URL] complete.

Taking 7600-8000T.

[URL="http://www.amphoria.co.uk/megabit-twin"]6000-6400T[/URL] complete.

Taking 8000-8400T.

[URL="http://www.amphoria.co.uk/megabit-twin"]6405-6600T[/URL] complete.

Taking 8400-8600T

I have uploaded a new sieve file sieved to p=6600T [URL="http://www.amphoria.co.uk/megabit-twin"]here[/URL]. There are 58,567,989 k's remaining.

Taking 8600-9000T.

[URL="http://www.amphoria.co.uk/megabit-twin"]7600-8000T[/URL] complete.

6600T-7600T range is complete and factors can be found here:

[URL="http://www.sendspace.com/file/dd73st"]http://www.sendspace.com/file/dd73st[/URL]

Take care

Kenneth

[URL="http://www.amphoria.co.uk/megabit-twin"]8400-9000T[/URL] complete.

Taking 9400-10100T.

I have uploaded a new sieve file sieved to p=9000T [URL="http://www.amphoria.co.uk/megabit-twin"]here[/URL]. There are 57,591,335 k's remaining.

Taking 10100-10300T.

Lennart

[URL="http://www.amphoria.co.uk/megabit-twin"]9000-9400T[/URL] complete.

Taking 10300-10800T.

[URL="http://www.amphoria.co.uk/megabit-twin"]9400-9800T[/URL] complete.

Taking 10800-11200T.

[URL="http://www.amphoria.co.uk/megabit-twin"]10300-10800T[/URL] complete.

Taking 11200-11600T.

[QUOTE=Lennart;277741]Taking 10100-10300T.

Lennart[/QUOTE]


Sorry but I have to unreserve this range becuase of a HD crach.


I will take some range later when I have fixed it.


Lennart

[URL="http://www.amphoria.co.uk/megabit-twin"]9800-10100T[/URL] complete.

Taking 11600-11900T.

I have uploaded a new sieve file [URL="http://www.amphoria.co.uk/megabit-twin"]0.1-120T_nohdr@10100T.7z[/URL] sieved to p=10100T. There are 57,187,394 k's remaining after removing 0-0.1T which was loaded into PRPNet.