Sieving for CRUS

[gd_barnes]

S596 is a new 1k base after BOINC found 4 primes for n=100K-300K. It now needs to be sieved for n=300K-1M.

[pepi37]

Originally Posted by gd_barnes

S596 is a new 1k base after BOINC found 4 primes for n=100K-300K. It now needs to be sieved for n=300K-1M.

No problem today will be sent to Yoyo

[pepi37]

Yoyo and I are willing to do some "test sieving" to depth 1e16 for anyone who needs the sieving done. There are two conditions:
First: It must be an npg file. No abcd files (so, one k remain in conjuncture)
Second: It must be sieved in at least two phases: First to 25e14 ( or 5e15) and then to 1e16.

So I would prefer to make this deep sieve for people that take a "hard sequence" to process where every candidate removed is a good thing. :)

For Reb : When you have many computers at your disposal 100-200 workunit more or less is not that significant but for one user that can save many days of unnecessary processing.

I will do additional sieving to 1e16 depth from time to time on already less-sieved files. If someone has a file that they would like us to work on, let me know here or in a PM.

P.S sieve depth for most cases is 1e15, and for npg is 5e15

Best regards to all

[rogue]

Originally Posted by pepi37

Yoyo and I are willing to do some "test sieving" to depth 1e16 for anyone who needs the sieving done. There are two conditions:
First: It must be an npg file. No abcd files (so, one k remain in conjuncture)
Second: It must be sieved in at least two phases: First to 25e14 ( or 5e15) and then to 1e16.

So I would prefer to make this deep sieve for people that take a "hard sequence" to process where every candidate removed is a good thing. :)

For Reb : When you have many computers at your disposal 100-200 workunit more or less is not that significant but for one user that can save many days of unnecessary processing.

I will do additional sieving to 1e16 depth from time to time on already less-sieved files. If someone has a file that they would like us to work on, let me know here or in a PM.

P.S sieve depth for most cases is 1e15, and for npg is 5e15

Why npg format? That makes absolutely no sense to me.

[pepi37]

Originally Posted by rogue

Why npg format? That makes absolutely no sense to me.

Because npg is output format from sr1sieve, and we talking about one sequence!

[rogue]

Originally Posted by pepi37

Because npg is output format from sr1sieve, and we talking about one sequence!

The latest sr1sieve supports ABC and ABCD formats for both input and output.

[pepi37]

Originally Posted by rogue

The latest sr1sieve supports ABC and ABCD formats for both input and output.

Since speed of latest sr1sieve and sieve 1.4.6 is same I dont think that YOYO will change server code and application just for support another input format.
One one thing is matter: that factors are ok, no missed factors, and to be stable.
1.4.6 give us all three conditions.
And at the end, it is irrelevant format : factors are more important



You are maker of srxsieve and mtsieve: you wrote those programs and I am very grateful for those programs, for your time and patience. Of all things only one is little dumbfounded me. For example: for many years npg format was ok, many use it , there is no problem with that. Now you start write mtsieve, and new versions of srxsieve and you decide that npg format is old and obsolete and you will abandon him. I cannot find real reason why now, after all those years,all of a sudden npg is no more desirable format.
On the same time when I in PM ask you that you add npg to twinsieve you add it without any problem. Does really few lines of codes is so much trouble for you?
I dont thinks so ( but that is my opinion)

[rogue]

Originally Posted by pepi37

Since speed of latest sr1sieve and sieve 1.4.6 is same I dont think that YOYO will change server code and application just for support another input format.
One one thing is matter: that factors are ok, no missed factors, and to be stable.
1.4.6 give us all three conditions.
And at the end, it is irrelevant format : factors are more important



You are maker of srxsieve and mtsieve: you wrote those programs and I am very grateful for those programs, for your time and patience. Of all things only one is little dumbfounded me. For example: for many years npg format was ok, many use it , there is no problem with that. Now you start write mtsieve, and new versions of srxsieve and you decide that npg format is old and obsolete and you will abandon him. I cannot find real reason why now, after all those years,all of a sudden npg is no more desirable format.
On the same time when I in PM ask you that you add npg to twinsieve you add it without any problem. Does really few lines of codes is so much trouble for you?
I dont thinks so ( but that is my opinion)

I wasn't aware that sieving was done in a client/server format where all required input files with the same format.

npg format is much harder to comprehend than ABC. Every time I write code for it, I have to look at the source code in order to figure out how to decipher the first line. This is the #1 reason why I avoid it.

I'm not saying that newpgen has no value. It supports sieving for other forms that no other program can do. I'm saying that there is little or no reason for new software to be compatible with that format.

[VBCurtis]

I can split a .npg file with a text editor and little knowledge. I don't know how to split an ABCD file for testing on multiple machines without using srfile or similar.

For human users of sieve outputs, .npg is the easiest format I know of to understand. The header explains what kind of sequence (to the testing software, not to me), and each line is an independent member of that sequence.

[pepi37]

Originally Posted by VBCurtis

I can split a .npg file with a text editor and little knowledge. I don't know how to split an ABCD file for testing on multiple machines without using srfile or similar.

For human users of sieve outputs, .npg is the easiest format I know of to understand. The header explains what kind of sequence (to the testing software, not to me), and each line is an independent member of that sequence.

Very nice words, explain perfect what I need to say.
Thanks VBCurtis

[rogue]

Originally Posted by VBCurtis

I can split a .npg file with a text editor and little knowledge. I don't know how to split an ABCD file for testing on multiple machines without using srfile or similar.

For human users of sieve outputs, .npg is the easiest format I know of to understand. The header explains what kind of sequence (to the testing software, not to me), and each line is an independent member of that sequence.

I fully agree with your statements regarding the ABCD format.

ABC is closest to npg, but is more flexible and doesn't take a lot of thought to understand it.