Computation error in 16e Lattice Sieve for smaller numbers

arakelov · 2021-05-29, 11:17

Hello all!

First of all, sorry if I'm posting in the wrong place. I'm very new on this world...

I recently joined NFS@Home through the BOINC client. Everything runs fine except for the application named "16e Lattice Sieve for smaller numbers" that shows "Computation error" in the status column (advanced view) lots of times.

I tried to get more info on these errors doing

$ grep -i "boinc" /var/log/syslog | grep -i "error"

but this returns nothing.

Can you give me more information about these kind of errors and how to prevent them? Are they usual?

Thanks a lot and best regards.

arakelov · 2021-05-29, 12:14

Here you can see a sample screenshot:

VBCurtis · 2021-05-29, 16:40

Welcome! You and your question are in exactly the right place.

I just tried to go the the NFS@home website to try to discover which job is failing, but the site appears to be down. So, perhaps it's a server issue rather than a job error. Whatever the problem is, it's not on your side- something went wrong either in the settings for this specific job that you received workunits for, or in the server itself.

From your side, the best action is to make a report like you made, and wait a couple days before running this particular BOINC client again. We really appreciate work on this client, so thank you!

arakelov · 2021-05-29, 17:45

Thanks for your response. So, your recommendation is to stop BOINC and run it again in two days?

You're right, escatter11.fullerton.edu seems down right now... Last time it was accessible, I found this recent thread with a reported bug that seems the same that I'm experiencing: https://escatter11.fullerton.edu/nfs...hp?id=795#2152

Maybe I should post also there?

Thanks for your guidance!

frmky · 2021-05-29, 18:34

The university is doing a disaster drill today which includes cutting off access to the internet and checking the failover of critical systems. So access today will be sporadic.

The error is caused by the use of an octic polynomial in the latest job. The Windows binaries support these but the Linux binaries don't. This job will pass soon, and the next job will use a smaller order polynomial. I have adding octic support to the Linux binaries on my to-do list.

arakelov · 2021-05-29, 19:03

Ok, thanks a lot for the info!

jyb · 2021-05-30, 01:03

Quote:

Originally Posted by frmky

The university is doing a disaster drill today which includes cutting off access to the internet and checking the failover of critical systems. So access today will be sporadic.

The error is caused by the use of an octic polynomial in the latest job. The Windows binaries support these but the Linux binaries don't. This job will pass soon, and the next job will use a smaller order polynomial. I have adding octic support to the Linux binaries on my to-do list.

I'm amazed that this hasn't caused noticeable problems before. There have been quite a few jobs with octic polynomials in the last year.

arakelov · 2021-05-31, 20:08

It's working again!

2021-05-29, 11:17	#1
arakelov Feb 2020 1001₂ Posts	Computation error in 16e Lattice Sieve for smaller numbers Hello all! First of all, sorry if I'm posting in the wrong place. I'm very new on this world... I recently joined NFS@Home through the BOINC client. Everything runs fine except for the application named "16e Lattice Sieve for smaller numbers" that shows "Computation error" in the status column (advanced view) lots of times. I tried to get more info on these errors doing $ grep -i "boinc" /var/log/syslog \| grep -i "error" but this returns nothing. Can you give me more information about these kind of errors and how to prevent them? Are they usual? Thanks a lot and best regards. Last fiddled with by arakelov on 2021-05-29 at 12:15

2021-05-31, 20:08	#8
arakelov Feb 2020 3² Posts	It's working again! Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
More NFS@Home 16e Lattice Sieve V5 wu's needed	pinhodecarlos	NFS@Home	46	2018-03-12 22:43
Using 16e on smaller numbers	fivemack	Factoring	3	2017-09-19 08:52
Advantage of lattice sieve over line sieve	binu	Factoring	3	2013-04-13 16:32
NFS on smaller numbers?	skan	YAFU	6	2013-02-26 13:57
Factoring Smaller Numbers	marc	Factoring	6	2004-10-09 14:17

2021-05-29, 12:14	#2
arakelov Feb 2020 9₁₆ Posts	Here you can see a sample screenshot: Attached Thumbnails

2021-05-29, 16:40	#3
VBCurtis "Curtis" Feb 2005 Riverside, CA 4,861 Posts	Welcome! You and your question are in exactly the right place. I just tried to go the the NFS@home website to try to discover which job is failing, but the site appears to be down. So, perhaps it's a server issue rather than a job error. Whatever the problem is, it's not on your side- something went wrong either in the settings for this specific job that you received workunits for, or in the server itself. From your side, the best action is to make a report like you made, and wait a couple days before running this particular BOINC client again. We really appreciate work on this client, so thank you!

2021-05-29, 17:45	#4
arakelov Feb 2020 3² Posts	Thanks for your response. So, your recommendation is to stop BOINC and run it again in two days? You're right, escatter11.fullerton.edu seems down right now... Last time it was accessible, I found this recent thread with a reported bug that seems the same that I'm experiencing: https://escatter11.fullerton.edu/nfs...hp?id=795#2152 Maybe I should post also there? Thanks for your guidance!

2021-05-29, 18:34	#5
frmky Jul 2003 So Cal 2106₁₀ Posts	The university is doing a disaster drill today which includes cutting off access to the internet and checking the failover of critical systems. So access today will be sporadic. The error is caused by the use of an octic polynomial in the latest job. The Windows binaries support these but the Linux binaries don't. This job will pass soon, and the next job will use a smaller order polynomial. I have adding octic support to the Linux binaries on my to-do list.

2021-05-29, 19:03	#6
arakelov Feb 2020 3² Posts	Ok, thanks a lot for the info!