Author |
Message |
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
Post here for general life stuff that doesn't belong anywhere else :). |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
Morning boiz, another 29C bank holiday here. i3 host 500597 CPU @ 80C last night so reduced to 50% CPU, same for 248648, so reduced performance this end with this currently ambient. Won't do too much damage to progress just for a day or 2. 33C in the den yesterday evening. |
|
|
|
I've enjoyed working on primegrid.
Curious for those of you who are more experienced roughly how often do you find a prime of any size?
I'm almost to my first 1million points and haven't found a single one. Hoping some year to find the largest woodall prime. |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
Oh it can take a while. This can be the ultimate 'test' of patience. Enjoy the journey though, not just the destination.
Work on SGS on CPU & GFN15 on GPU if you want your first find, although these are not big enough for T5k classification. PPSE on CPU & GFN16 on GPU is the first entry into T5k-land.
E.g here are mine (each user page has a link saying Primes/APs with a ratio of initial to double-check): https://www.primegrid.com/primes/?section=primelist&userid=130544. |
|
|
Monkeydee Volunteer tester
 Send message
Joined: 8 Dec 13 Posts: 548 ID: 284516 Credit: 1,736,929,573 RAC: 3,178,961
                            
|
The best projects for finding primes are SGS and PPSE on the CPU and GFN15/GFN16 on the GPU. Basically the smaller the number the easier it is to find a prime.
My first prime was as a double checker in June of 2014. I had been here for 6 months at the time and that was an SGS project prime.
Primegrid as a project has only discovered four Woodall primes between 2007 and today. The most recent of which was on March 21st, 2018. If you really want to find a Woodall prime then it would be best to turn your CPUs to that project and let it rip. But don't be surprised if it takes years for anyone to find one.
Primegrid is a game of patience and luck.
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*1 + 8*11 + 9*1 + 11*3 + 12*1 = 169
|
|
|
|
Does anyone know how Project Staging Area works in terms of setup and grinding?
I don't see it as an option in work/project preferences.
|
|
|
Monkeydee Volunteer tester
 Send message
Joined: 8 Dec 13 Posts: 548 ID: 284516 Credit: 1,736,929,573 RAC: 3,178,961
                            
|
Does anyone know how Project Staging Area works in terms of setup and grinding?
I don't see it as an option in work/project preferences.
PSA is a very different beast than the rest of PG. PSA is not something you can get credit for through BOINC apps.
PSA consists of two different kinds of work. Manual Sieving and PRPNet.
Manual Sieving can be found here: http://www.primegrid.com/show_sieving.php
PRPNet can be found here: http://prpnet.primegrid.com/
I have never done manual sieving, so I'll leave that to others here to explain in detail.
PRPNet can be a little intimidating to set up at first.
You download the package from the link above and then you edit several ini files with the settings you want. Instructions can be found in the Project Staging Area section on the forums here: http://www.primegrid.com/forum_thread.php?id=1215
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*1 + 8*11 + 9*1 + 11*3 + 12*1 = 169
|
|
|
|
Does anyone know how Project Staging Area works in terms of setup and grinding?
I don't see it as an option in work/project preferences.
Keith explained about PRPNet.
I'll give an explanation of manual sieving...
MSRs, or Manual Sieving Reservations are the other way to earn PSA credit. These can earn you A LOT of credit.
It works like this: you reserve a range to sieve on a certain GFN range like GFN18 Sieving or GFN23 sieving. You reserve what is called a P. Basically a certain block size of numbers you will work on. When you start you can only reserve small amounts of P because they want to make sure you can do them successfully before working on larger ranges which can take some time. The more ranges you complete successfully the more and more P you will be able to reserve.
Now, after you have reserved your range you get taken to a page where it gives you the command line to run. You need to save the Linux or Windows app to your machine. Always use opencl, never the CUDA one. The CUDA one is very very slow in comparison. Save this in a directory of your choice.
To get the proper command line you can set options like conserve CPU which I always tend to use and the B or block size. This is how much data gets sent to the GPU. Bigger block sizes means it often will run faster. Some GPUs can not handle larger block sizes. B7 is the dafault. To max it out use B13. This can cause screen lag on some machines with older GPUs. Highlight and copy your command line
Then, open a terminal or command prompt window. Change to the directory where you saved the sieving program. Run that program with the command line given on the webpage for your reserved range. I like to rename the sieving program to something something like gfnsocl so I can type it quick. Your command like will be something like "gfnsocl 23 15000 15100 B13 W1"
Once it starts it will give you an ETA of how long it's expected to take. Do not reserve too much. There is no time limit but often JimB gives a warning after 2 weeks. I usually don't reserve more than a week at most. Sometimes overnight or a 1 day range.
You will see a P/day which is how many P/day your GPU is expected to crunch..This will vary on different GFN ranges as the CPU becomes more of a limiting factor. To use the least CPU and sieve really quickly as a result run the higher GFN ranges such as GFN22 and GFN23. These require the least CPU power. If you run GFN17 or GFN18 you may need to run more than one instance of the sieving program at a time because the CPU needs to do a lot of work. The sieving program only uses 1 CPU core per instance. To do this reserve more than one range, and using multiple command prompt or terminal windows run your sieve program for each range in those separate windows.
Once you see how fast your GPU can do at the upper ranges you can adjust these multiple instances to work for the lower GFN ranges and get the max P/day total.
If the CPU is not a limiting factor then running more instances won't increase your P/day but it will increase your time to complete a range. Sometimes if you can only reserve a certain size you can reserve multiple ranges, run them all at once, and increase the run time to something like the following morning instead of having them only take 4hrs at 1 instance.
You can stop and restart your sieving program. It checkpoints to a file and saves all it's work. There is the sieving programs, checkpoint file, and fount factor files. For each new range it will create new or reuse if you are resuming, checkpoint files and factor files.
When your range is completed it will tell you so. If you try to rerun that range it will check the checkpoint file and know it has been completed. You must now upload the factor file to receive credit and complete your reservation. Go to the manual sieving reservations under your account. Find your open reservations, find the range you have completed, and select upload. Browse your files to your sieving program directory and select the 'f' file. The one beginning with f that corresponds to your completed range. If you happen to upload a wrong file the system will know this and you will have to upload the correct one .
Range uploaded... that's the end of your reservation. Next JimB will verify and approve the ranges, if all is well they will be marked as completed and you will be granted credit.
Hope this gets some team members started. As I said at the beginning.. MSRs can generate A LOT of credit. You won't find primes at all, only factors which eliminate numbers which have to be checked in the future. |
|
|
|
Thank you Penguin! |
|
|
|
Thank you Penguin!
You're welcome! |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
What's the opposite of an orgasm?
...
An XORgasm. |
|
|
|
An XORgasm.
[/img] |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
A very Merry Christmas or whatever you observe, hope you have a good one. |
|
|
|
Happy New Years everyone!!
|
|
|
|
Hope everyone is safe and healthy during this pandemic. Happy prime hunting! 😁 |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
Likewise. Revision to UK rules up & coming w/c 11/5, relieving the exercise-once-a-day restriction, although I fear this will spike traffic emissions. I expect to be returning to work this month now. |
|
|
|
Happy Pi Day 2021 Everyone! |
|
|
|
Happy Pi Day 2021 Everyone!
Yes! Happy Pi Day! |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
You're a bit, as they say, late. |
|
|
|
Watch out for Dave and his new "toy". :) He'll be moving up the stats ladders super fast now. Congrats on your new machine Dave! |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
Thank you Penguin. It's taking some getting my head around tbh, but so far a very clean commissioning period. Keeping a close eye on as many graphs as I can find. |
|
|
|
Thank you Penguin. It's taking some getting my head around tbh, but so far a very clean commissioning period. Keeping a close eye on as many graphs as I can find.
The larger sized LLR projects are where it will excel vs your old cpu more so than elsewhere. Cache cache cache
|
|
|
|
GPU temperature 103, I wonder how long it will last. |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
GPU temperature 103, I wonder how long it will last.
103 what???!!! Stop processing, clean out the gunk & reconsider its installation position. Come to Discord. |
|
|
|
Просто у меня система охлаждения самодельная, сильно зависит от окружающей температуры, заводскую пришлось снять - слишком шумная. Сегодня прохладно в комнате так выше 98 не видел, вообще всё норм, видеокарта уже проверена временем, в самом корпусе холодно, текстолит охлаждается вентиляторами, ничего не воняет :) Просто написал, может кому интересно в каком состоянии работает железо, но спасибо за помощь. Думаю единственное что ей поможет так это "arctic cooling accelero mono plus", но для неё (видеокарты) это роскошь. |
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
English only please. |
|
|
Bertle Volunteer tester Send message
Joined: 27 Jan 20 Posts: 28 ID: 1236175 Credit: 455,000,416 RAC: 99,996
                   
|
Hi people
Hope you're all doing well.
____________
|
|
|
Monkeydee Volunteer tester
 Send message
Joined: 8 Dec 13 Posts: 548 ID: 284516 Credit: 1,736,929,573 RAC: 3,178,961
                            
|
Heya Bertle!
Always glad to hear from you! I hope all is going well :)
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*1 + 8*11 + 9*1 + 11*3 + 12*1 = 169
|
|
|
|
Hi people
Hope you're all doing well.
Hi Bertle! :wave:
|
|
|
Dave  Send message
Joined: 13 Feb 12 Posts: 3258 ID: 130544 Credit: 2,476,706,884 RAC: 4,456,157
                           
|
A very happy new year from myself to the rest of the team. I wish you prosperous developments both in terms of PG & elsewhere. |
|
|
|
A very happy new year from myself to the rest of the team. I wish you prosperous developments both in terms of PG & elsewhere.
Right on Dave. May it be a year with much development and prosperity.
|
|
|
|
Aha!!! Youse guy's/gal's brought me luck joining this team this challenge as I just got a 17 mega!! :) |
|
|
Monkeydee Volunteer tester
 Send message
Joined: 8 Dec 13 Posts: 548 ID: 284516 Credit: 1,736,929,573 RAC: 3,178,961
                            
|
Aha!!! Youse guy's/gal's brought me luck joining this team this challenge as I just got a 17 mega!! :)
That's awesome! Congrats!!!
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*1 + 8*11 + 9*1 + 11*3 + 12*1 = 169
|
|
|
|
Not to sound like a dumb a$$ or anything ha ha but..... challenge cleanup... So doesthat occur just like normal crunching? As i just keep crunching gfn17 and the server will just hand out the cleanuo tasks like regular tasks? |
|
|
Michael Goetz Volunteer moderator Project administrator
 Send message
Joined: 21 Jan 10 Posts: 14045 ID: 53948 Credit: 485,740,737 RAC: 668,833
                               
|
Not to sound like a dumb a$$ or anything ha ha but..... challenge cleanup... So doesthat occur just like normal crunching? As i just keep crunching gfn17 and the server will just hand out the cleanuo tasks like regular tasks?
Correct.
But...
The truth about cleanups is that mostly they're about waiting for abandoned tasks to time out and be resent by the server. Mostly, all the cleanup tasks have already been sent out to be crunched. Most of them will be crunched successfully, but a few won't.
So there's usually not a whole lot of cleanup tasks available. Often, none at all. Cleanup tasks, when available, are shown on the front page like this:
* Only 20 tasks, 12 affecting scoring positions, of Geminids Shower Challenge (GFN-21) cleanup work are currently available! *
If the message is red, there's a lot of them available. If it's not red, then the number is fairly insignificant.
At the moment, there's no GFN-17 cleanup tasks available.
These numbers are updated every 15 minutes.
____________
My lucky number is 75898524288+1 |
|
|
|
Not to sound like a dumb a$$ or anything ha ha but..... challenge cleanup... So doesthat occur just like normal crunching? As i just keep crunching gfn17 and the server will just hand out the cleanuo tasks like regular tasks?
Correct.
But...
The truth about cleanups is that mostly they're about waiting for abandoned tasks to time out and be resent by the server. Mostly, all the cleanup tasks have already been sent out to be crunched. Most of them will be crunched successfully, but a few won't.
So there's usually not a whole lot of cleanup tasks available. Often, none at all. Cleanup tasks, when available, are shown on the front page like this:
* Only 20 tasks, 12 affecting scoring positions, of Geminids Shower Challenge (GFN-21) cleanup work are currently available! *
If the message is red, there's a lot of them available. If it's not red, then the number is fairly insignificant.
At the moment, there's no GFN-17 cleanup tasks available.
These numbers are updated every 15 minutes.
|
|
|
|
Thank you very much for the explanation Michael. :-) |
|
|
|
Aha!!! Youse guy's/gal's brought me luck joining this team this challenge as I just got a 17 mega!! :)
You are doing very well. Not just 1, but 2 Mega Prime Finds!
Welcome to the team. |
|
|
Renix Send message
Joined: 26 Aug 16 Posts: 349 ID: 455383 Credit: 4,243,245,463 RAC: 14,509
                      
|
Well the hot hot hot weather has arrived, so going to take a breather from the crunching. And also that's not to mention that our elec prices are increasing too. So gonna sign off. It's been a kick!! :-) |
|
|