Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Generalized Fermat Prime Search :
GFN-1x Small Primes search starts 2021-01-21
| Author |
Message |
|---|
streamVolunteer moderator
Project administrator
Volunteer developer
Volunteer tester
Send message
Joined: 1 Mar 14
Posts: 860
ID: 301928
Credit: 504,204,908
RAC: 201,569
                        
|
|
To continue efforts to create complete list of small GFN primes up to b=2G, a "Search for Small Primes" project will be started January 21, 2021, 09:00 UTC on "Private GFN server".
Note this is not a PrimeGrid project. It's running on a separate server, server address is http://boincvm.proxyma.ru:30080/test4vm/ (also note the non-standard port). If you're a new user here, you must create your account on the site first, using invitation code PrimeGrid. Automatic registration from Boinc client most probably will not work.
Project description
The goal is to create full list of GFN primes as far as software limit and computation resources allows. Historically, each sequence is tested, depending on it's complexity, either up to b=400M (old software limit) either up to b=2G. Current testing status is:
GFN-1 - GFN-11 - tested manually up to b=2G by a group of PrimeGrid users;
GFN-12 - wasn't tested to significant depth;
GFN-13, GFN-14 - tested on Private GFN server up to b=400M;
GFN-15 and above - currently tested on PrimeGrid.
In the first phase, we will test GFN-12 up to b=2G. There are plans to run second phase, extending GFN-13 search up to b=2G, but it'll require additional sieving, so it'll be a delay between first and second phase.
Primes found during this search will be relatively small and cannot be reported to any "top" list. We will find thousand of them. Although they'll be collected mostly for statistical purposes, these primes can be used as a starting points searching for "chains" or "twin" primes, which may lead to another interesting and reportable discoveries.
Technical details
The project is GPU only.
We'll use new generation of Yves Gallot' program, "Genefer 2.0". Designed exclusively to test small numbers (when each test takes just a few seconds), it can test many candidates at once, fully utilizing your GPU and requiring almost no CPU time. It also implements Gerbicz error checking to detect hardware errors in GPU calculations (unfortunately, fast double-checking is not possible yet).
Task size has been selected to be about 30 minutes on a 1660-class GPU. Since task runtime is proportional to log2(b), first tasks will be completed faster. They'll reach expected length around b=300M and continue to grow slowly until end of project.
Since it'll be first time when Genefer 2.0 will be used in Boinc environment and on many different hosts, all tasks will be fully double-checked.
Known issues
** No graceful error handling **
This program does not tolerates even slightest of GPU errors. If case of any calculation error, program will immediately abort with "Gerbicz failed" message. There are no restarts. Check GPU temperatures, underclock the GPU.
** A problem with multi-GPU systems **
A bug in version 1.00 causes program to always use first GPU in the system. It'll be fixed soon.
Update 1.01 fixed some multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
** Note **
Please avoid running this project on hosts which can interrupt GPU calculations often, especially on slow GPUs. For example, on systems which are configured to run Boinc or GPU tasks "only when user is not active", and user interrupts too often. Currently implemented checkpoint system saves only index of last fully processed candidate, but not a current state of each test. Although each individual test can be short (for example, 2 seconds), there are up to 256 test running in parallel. If a program will be interrupted again within 256*2 seconds, no progress can be recorded. | |
|
|
|
|
Looking forward to it! Thanks for making it possible :D | |
|
|
|
|
Very excited to see Yves new app in action! | |
|
 robish Volunteer moderator
Volunteer tester
Send message
Joined: 7 Jan 12
Posts: 1861
ID: 126266
Credit: 5,286,425,849
RAC: 2,798,366
                        
|
|
Ready.....
____________
My lucky numbers 10590941048576+1 and 224584605939537911+81292139*23#*n for n=0..26 | |
|
|
|
|
Set ....
____________
 | |
|
|
|
|
Go!!!
____________
SHSIDElectronicsGroup@outlook.com
waiting for a TdP prime...
Proth "SoB": 44243*2^440969+1
| |
|
|
|
Please avoid running this project on hosts which can interrupt GPU calculations often, especially on slow GPUs. For example, on systems which are configured to run Boinc or GPU tasks "only when user is not active", and user interrupts too often. Currently implemented checkpoint system saves only index of last fully processed candidate, but not a current state of each test. Although each individual test can be short (for example, 2 seconds), there are up to 256 test running in parallel. If a program will be interrupted again within 256*2 seconds, no progress can be recorded.
Hmm I've just noticed that there are 2048 tasks in total, in 8 batches which means 256 per batch, runtime on a GTX1650 is around 22 minutes.
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size? Runtime is around 5-6 hours.
In fact, I can see the stderr output in the project folder and see if I found a prime in these 2048 tests 😁
What are your times?
____________
SHSIDElectronicsGroup@outlook.com
waiting for a TdP prime...
Proth "SoB": 44243*2^440969+1
| |
|
streamVolunteer moderator
Project administrator
Volunteer developer
Volunteer tester
Send message
Joined: 1 Mar 14
Posts: 860
ID: 301928
Credit: 504,204,908
RAC: 201,569
|
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size?
It does a benchmark in the beginning of test (takes about 40 seconds on my system) and selects fastest parameters for given GPU (not only batch size, there are few others).
Tasks sizes are too different now because candidates are too small. They'll stabilize around b=300M.
| |
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 660
ID: 164101
Credit: 305,042,960
RAC: 56
|
Hmm I've just noticed that there are 2048 tasks in total, in 8 batches which means 256 per batch, runtime on a GTX1650 is around 22 minutes.
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size? Runtime is around 5-6 hours.
The number of tests per batch depends on the maximum work-group size. On Nvidia CL_KERNEL_WORK_GROUP_SIZE=1024 => 256 batches and on AMD CL_KERNEL_WORK_GROUP_SIZE=256 => 64 batches.
It is too bad this parameter is so small on AMD. This is a limitation of the driver, not of hardware. If this parameter was 1024, AMD GPUs would be more efficient for computations.
| |
|
|
|
Hmm I've just noticed that there are 2048 tasks in total, in 8 batches which means 256 per batch, runtime on a GTX1650 is around 22 minutes.
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size? Runtime is around 5-6 hours.
The number of tests per batch depends on the maximum work-group size. On Nvidia CL_KERNEL_WORK_GROUP_SIZE=1024 => 256 batches and on AMD CL_KERNEL_WORK_GROUP_SIZE=256 => 64 batches.
It is too bad this parameter is so small on AMD. This is a limitation of the driver, not of hardware. If this parameter was 1024, AMD GPUs would be more efficient for computations.
Ahh I see. big oops for AMD then...
And @stream could we have a status page for this project like before? Would there be too many primes?
____________
SHSIDElectronicsGroup@outlook.com
waiting for a TdP prime...
Proth "SoB": 44243*2^440969+1
| |
|
streamVolunteer moderator
Project administrator
Volunteer developer
Volunteer tester
Send message
Joined: 1 Mar 14
Posts: 860
ID: 301928
Credit: 504,204,908
RAC: 201,569
|
And @stream could we have a status page for this project like before? Would there be too many primes?
It's at usual place (GFN-12-13-14 statistic).
Individual primes will be not shown - we'll have tens of thousands of them.
| |
|
|
|
Hmm I've just noticed that there are 2048 tasks in total, in 8 batches which means 256 per batch, runtime on a GTX1650 is around 22 minutes.
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size? Runtime is around 5-6 hours.
The number of tests per batch depends on the maximum work-group size. On Nvidia CL_KERNEL_WORK_GROUP_SIZE=1024 => 256 batches and on AMD CL_KERNEL_WORK_GROUP_SIZE=256 => 64 batches.
It is too bad this parameter is so small on AMD. This is a limitation of the driver, not of hardware. If this parameter was 1024, AMD GPUs would be more efficient for computations.
What happens if you run more than one at a time on the GPU? | |
|
|
|
On a Vega 8 iGPU, there are 32 batches in total, with 64 tests per btach, I think the app detects the GPU and adjusts the batch size?
It does a benchmark in the beginning of test (takes about 40 seconds on my system) and selects fastest parameters for given GPU (not only batch size, there are few others).
Tasks sizes are too different now because candidates are too small. They'll stabilize around b=300M.
Is that what was happening at the start? Mine seemed to be doing nothing, I got 12 seconds of CPU usage, very little GPU usage, then nothing for quite some time, then it got going at full GPU usage.
By the way, AVG is (presumably falsely!) detecting the program as a virus. It sent it off for checking at their end, so presumably once they ok it, it won't bother any AVG users. I just clicked ignore and it worked :-) | |
|
|
|
** A problem with multi-GPU systems **
A bug in version 1.00 causes program to always use first GPU in the system. It'll be fixed soon.
Ah. I missed this note the first time I read the post. That knocks out my two dual-1660 Ti machines.
When is "soon" roughly? Sometime during the challenge, or after?
____________
Reno, NV
 | |
|
|
|
** A problem with multi-GPU systems **
A bug in version 1.00 causes program to always use first GPU in the system. It'll be fixed soon.
Ah. I missed this note the first time I read the post. That knocks out my two dual-1660 Ti machines.
When is "soon" roughly? Sometime during the challenge, or after?
You could always forbid the second GPU to run that project in the app config file. Then at least you can run one GPU per machine on it. | |
|
|
|
You could always forbid the second GPU to run that project in the app config file. Then at least you can run one GPU per machine on it.
Yep. Setting that up now.
____________
Reno, NV
| |
|
Yves GallotVolunteer developer
Project scientist
Send message
Joined: 19 Aug 12
Posts: 660
ID: 164101
Credit: 305,042,960
RAC: 56
|
** A problem with multi-GPU systems **
A bug in version 1.00 causes program to always use first GPU in the system. It'll be fixed soon.
When is "soon" roughly? Sometime during the challenge, or after?
Soon is today.
| |
|
|
|
** A problem with multi-GPU systems **
A bug in version 1.00 causes program to always use first GPU in the system. It'll be fixed soon.
When is "soon" roughly? Sometime during the challenge, or after?
Soon is today.
Yes, it looks like the updated 1.01 app version you rolled out has fixed this issue on my multi-gpu machines. Thanks!
____________
| |
|
streamVolunteer moderator
Project administrator
Volunteer developer
Volunteer tester
Send message
Joined: 1 Mar 14
Posts: 860
ID: 301928
Credit: 504,204,908
RAC: 201,569
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
| |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Working ok here with two different ATIs in the same machine. (R9 280X and RX 560).
Also working if I double up the tasks per GPU, it makes them run at 100% instead of 85%.
Good luck getting Nvidia to work alongside ATI, that used to be impossible with any project.
| |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Works like a charm. Thanks!
____________
Reno, NV
| |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Good luck getting Nvidia to work alongside ATI, that used to be impossible with any project.
Was totally fine on my Win7 machine with GTX 970 in Slot 1 and Radeon 7750 in Slot 2 on several projects, after I stopped running Einstein on the AMD only and my other projects on the Nvidia only.
Especially on PrimeGrid I could run GFN-something on the Nvidia and PPS-Sieve on the AMD. Or different GFNs on both. Or the same ones.
Took the 970 out some time ago, so atm I don't mix GPUs.
____________
Greetings, Jens
92914140^65536+1 | |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Good luck getting Nvidia to work alongside ATI, that used to be impossible with any project.
Was totally fine on my Win7 machine with GTX 970 in Slot 1 and Radeon 7750 in Slot 2 on several projects, after I stopped running Einstein on the AMD only and my other projects on the Nvidia only.
Especially on PrimeGrid I could run GFN-something on the Nvidia and PPS-Sieve on the AMD. Or different GFNs on both. Or the same ones.
Took the 970 out some time ago, so atm I don't mix GPUs.
So you're saying you usually had to run different projects on each?
I've actually heard of people not being able to get Windows to work with 2 different makes of card, nevermind Boinc, but that was a long time ago, perhaps drivers are better now.
I've got more PCs than cards, so it's easy enough to keep them grouped by similar card. | |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Good luck getting Nvidia to work alongside ATI, that used to be impossible with any project.
Was totally fine on my Win7 machine with GTX 970 in Slot 1 and Radeon 7750 in Slot 2 on several projects, after I stopped running Einstein on the AMD only and my other projects on the Nvidia only.
Especially on PrimeGrid I could run GFN-something on the Nvidia and PPS-Sieve on the AMD. Or different GFNs on both. Or the same ones.
Took the 970 out some time ago, so atm I don't mix GPUs.
So you're saying you usually had to run different projects on each?
I've actually heard of people not being able to get Windows to work with 2 different makes of card, nevermind Boinc, but that was a long time ago, perhaps drivers are better now.
I've got more PCs than cards, so it's easy enough to keep them grouped by similar card.
As I think AMD cards are a PITA on Linux, I put my 7750 into the Windows box that also had a 970.
I used this combo to be able to run Einstein full-time on the Radeon, thus preventing difficulties with the amount of threads used by GPUs.
Most of the time this worked flawlessly. If not, Einstein had forgotten my preferences, and I had to set them anew.
As Einstein would get more out of my Nvidias, I changed my settings at some point in time and let Einstein use Nvidia, but set its GPU work to 0.45 CPU use instead of 1.
Played with venues on several projects and had the 7750 get work from Moo, Collatz, MilkyWay, and sometimes PrimeGrid.
I never had issues with Boinc mistaking cards and having work run on the wrong one.
____________
Greetings, Jens
92914140^65536+1 | |
|
|
|
|
Hello!
Is it only my pc's not able to upload work? No problem getting new work and server status
seems ok, all is up and running, although browsing the site is slow.
A lot of users with heavy equipement, my guess db starting to fill up?
Thank You again Stream for making another interesting project availible☺
With regards,
Hans Sveen
Oslo
____________
MyStats
My Badges | |
|
|
|
Hello!
Is it only my pc's not able to upload work? No problem getting new work and server status
seems ok, all is up and running, although browsing the site is slow.
A lot of users with heavy equipement, my guess db starting to fill up?
Thank You again Stream for making another interesting project availible☺
With regards,
Hans Sveen
Oslo
Not having a problem here, but I only have 3 cards on it. My 4th card has (I think) a RAM fault and will only run smaller programs.
| |
|
|
|
Update 1.01 fixed most critical multi-GPU issues, but it'll work only if your cards are of same type (i.e. all NVidia or all ATI). A combo of NVidia+ATI will not work. Correct fix is more complex. Hopefully it'll be done later.
Good luck getting Nvidia to work alongside ATI, that used to be impossible with any project.
Was totally fine on my Win7 machine with GTX 970 in Slot 1 and Radeon 7750 in Slot 2 on several projects, after I stopped running Einstein on the AMD only and my other projects on the Nvidia only.
Especially on PrimeGrid I could run GFN-something on the Nvidia and PPS-Sieve on the AMD. Or different GFNs on both. Or the same ones.
Took the 970 out some time ago, so atm I don't mix GPUs.
So you're saying you usually had to run different projects on each?
I've actually heard of people not being able to get Windows to work with 2 different makes of card, nevermind Boinc, but that was a long time ago, perhaps drivers are better now.
I've got more PCs than cards, so it's easy enough to keep them grouped by similar card.
As I think AMD cards are a PITA on Linux, I put my 7750 into the Windows box that also had a 970.
I used this combo to be able to run Einstein full-time on the Radeon, thus preventing difficulties with the amount of threads used by GPUs.
Most of the time this worked flawlessly. If not, Einstein had forgotten my preferences, and I had to set them anew.
As Einstein would get more out of my Nvidias, I changed my settings at some point in time and let Einstein use Nvidia, but set its GPU work to 0.45 CPU use instead of 1.
Played with venues on several projects and had the 7750 get work from Moo, Collatz, MilkyWay, and sometimes PrimeGrid.
I never had issues with Boinc mistaking cards and having work run on the wrong one.
I set everything to use 0.01 CPUs, as in zero. Most of the time it can get enough CPU by squeezing in with the other tasks, certain combinations need tweaking, but I only do that when necessary, first I try running two tasks at once on the GPU, if that fails I change it to need a core free. I go by the GPU usage in MSI Afterburner (you can also use things like GPU-Z, something with a graph is nice).
The only problems I've had recently with Boinc on GPUs is:
1) If you run two or tasks per card, it doesn't realise they will take twice or more as long, and downloads too much work.
2) For a short while after I added a 2nd card (both AMD) to this windows machine, it swapped over what it called d0 and d1 a few times. Strangely it's now settled to have them in the same order as everything else (as in Windows, utilities, etc). Historically Boinc always listed them in reverse order which was confusing when trying to work out which was the fastest one.
| |
|
streamVolunteer moderator
Project administrator
Volunteer developer
Volunteer tester
Send message
Joined: 1 Mar 14
Posts: 860
ID: 301928
Credit: 504,204,908
RAC: 201,569
|
Is it only my pc's not able to upload work? No problem getting new work and server status
seems ok, all is up and running, although browsing the site is slow.
We noticed some network problems in Europe today, may be your traffic comes through one of affected paths. Connectivity was restored within a hour. No problems from other locations.
A lot of users with heavy equipement, my guess db starting to fill up?
No. This project is just a toy comparing to GFN-13/14. Only 102657 workunits for whole project. A GFN-13, for example, had 100K workunits just for 1M of search space. I only need to monitor disk space, it generates 1 Gb of results every day.
| |
|
|
|
Is it only my pc's not able to upload work? No problem getting new work and server status seems ok, all is up and running, although browsing the site is slow.
We noticed some network problems in Europe today, may be your traffic comes through one of affected paths. Connectivity was restored within a hour. No problems from other locations.
I'm in the UK and had difficulty using the Primegrid forums yesterday and the Steam (for games) forum, but everything else was ok. They were both timing out when trying to post.
A lot of users with heavy equipement, my guess db starting to fill up?
No. This project is just a toy comparing to GFN-13/14. Only 102657 workunits for whole project. A GFN-13, for example, had 100K workunits just for 1M of search space. I only need to monitor disk space, it generates 1 Gb of results every day.
That doesn't sound like much, this desktop has a 4000GB disk....
| |
|
|
|
|
Hello!
Thank You for giving me
clear answers!☺
With regards,
H Sveen
Oslo
____________
MyStats
My Badges | |
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2863
ID: 130544
Credit: 1,021,782,057
RAC: 1,184,135
            
|
|
Apologies if I'm missing it somewhere, but what is the app name for app_config? | |
|
|
|
Apologies if I'm missing it somewhere, but what is the app name for app_config?
gfn1x_small - to find these names, you have to look in the file "C:\ProgramData\BOINC\client_state.xml", find the section for this project, then look for names.
I use this to make it run 3 at once, otherwise the GPU isn't fully used:
<app_config>
<app>
<name>gfn1x_small</name>
<gpu_versions>
<gpu_usage>0.333</gpu_usage>
<cpu_usage>0.01</cpu_usage>
</gpu_versions>
</app>
</app_config> | |
|
Dave
Send message
Joined: 13 Feb 12
Posts: 2863
ID: 130544
Credit: 1,021,782,057
RAC: 1,184,135
|
|
Thanks Peter that's got it working! | |
|
Post to thread
Message boards :
Generalized Fermat Prime Search :
GFN-1x Small Primes search starts 2021-01-21 |
