Message boards :
Number crunching :
Don't know where it should go? Stick it here!
Message board moderation
Previous · 1 . . . 73 · 74 · 75 · 76 · 77 · 78 · 79 . . . 147 · Next
Author | Message |
---|---|
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I try to never think of it . . . . . that way I can never be disappointed. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Speedy Send message Joined: 26 Jun 04 Posts: 1643 Credit: 12,921,799 RAC: 89 |
Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at it |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at it I'd expect they're just put aside for another day. There are some WUs in there that do crunch, so once they get the servers sorted out so they can handle the load, they could release them here again. Grant Darwin NT |
Speedy Send message Joined: 26 Jun 04 Posts: 1643 Credit: 12,921,799 RAC: 89 |
Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at it Thanks Grant I thought that was possibly the case. If they are complete noise bombs go and end within 3 seconds as there useful information within that time? |
rob smith Send message Joined: 7 Mar 03 Posts: 22204 Credit: 416,307,556 RAC: 380 |
Noise bombs may contain useful data, but the current analysis tools don't let us get at it if it is there... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Speedy Send message Joined: 26 Jun 04 Posts: 1643 Credit: 12,921,799 RAC: 89 |
Noise bombs may contain useful data, but the current analysis tools don't let us get at it if it is there... Thank you for the information. This is good to know |
Shannon Lester Send message Joined: 27 Jul 09 Posts: 83 Credit: 12,388,119 RAC: 140 |
I figured I would just post here instead of making a new thread. I have 3 new work units I’ve never seen before for my gpu. Two cuda42 and one cuda50. What are these? All of this has happened before and will happen again -Battlestar Galactica |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Since you are running the normal platform decided by the schedulers, they have sent you some of the older applications. The scheduler is still trying to determine the best app for your host. The CUDA42 and CUDA50 applications are MUCH slower than the SoG application. You are just going to have to suffer through the CUDA42 and CUDA50 tasks until the APR's for those apps are established and the SoG app wins out in the end. The only way to circumvent this scenario is to run the anonymous platform where YOU get to decide which apps to run. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Shannon Lester Send message Joined: 27 Jul 09 Posts: 83 Credit: 12,388,119 RAC: 140 |
Since you are running the normal platform decided by the schedulers, they have sent you some of the older applications. The scheduler is still trying to determine the best app for your host. The CUDA42 and CUDA50 applications are MUCH slower than the SoG application. You are just going to have to suffer through the CUDA42 and CUDA50 tasks until the APR's for those apps are established and the SoG app wins out in the end. The only way to circumvent this scenario is to run the anonymous platform where YOU get to decide which apps to run. Thanks Keith! I’ll let it crunch whatever they send me until it gets used to the new host haha. Someone’s gotta do it. LOL! All of this has happened before and will happen again -Battlestar Galactica |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
You will need 11 validated, non-overflowed CUDA42 and 50 tasks to develop the APR for those applications. So much is determined by the turnaround time for your wingmen. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at itI would imagine they will run them through at some point here. I server can handle them - just not all at once! If they are split one at a time they will whistle their way right on through ... but takes some babysitting and the hard to find TIME. |
Speedy Send message Joined: 26 Jun 04 Posts: 1643 Credit: 12,921,799 RAC: 89 |
Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at itI would imagine they will run them through at some point here. I server can handle them - just not all at once! I'm aware that the noise bombs create a lot of pressure for the server but eventually it clears itself and the creation rate is able to come back to a reasonable level would it not be more sensible to just let them all run the same time? |
Wiggo Send message Joined: 24 Jan 00 Posts: 34754 Credit: 261,360,520 RAC: 489 |
We had a similar batch not long back that also got pulled, but they wound up letting a file or 2 process at a time a few days later until that batch was done. I imagine that they'll do the same thing again when they have the time to manually feed them back in.I'm aware that the noise bombs create a lot of pressure for the server but eventually it clears itself and the creation rate is able to come back to a reasonable level would it not be more sensible to just let them all run the same time?Anybody know what happens to the noisy data (noise bomb) data when it is pulled from the server because the server can't cope? I am wondering doesn't get discarded or does somebody else look at itI would imagine they will run them through at some point here. I server can handle them - just not all at once! Cheers. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
would it not be more sensible to just let them all run the same time? If the servers could handle it, yes. But the last two server system crashes were as a result of the shorty storms that had been occurring at the time. The servers just can't cope with that sort of sustained load. Grant Darwin NT |
Shannon Lester Send message Joined: 27 Jul 09 Posts: 83 Credit: 12,388,119 RAC: 140 |
GPU upgrade question... I was thinking about upgrading from my 1060 6gb at the end of the year and was wondering about some of the newer Nvidia cards. How many work units would a 1660Ti handle at one time? 3 or 4? Or would it be better to go with a 2070 with 8gb ram and run 4-5 work units on. My 1060 6gb is working very well with 2 at a time but I would love to be able to run more per card. hopefully by the end of the year some prices will come down and I can grab 2 or 3 cards to put in this Threadripper rig of mine. All of this has happened before and will happen again -Battlestar Galactica |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
You wouldn't be able to do that many at a time without using the 2080 Ti which has enough memory for 4 tasks at a time. The differences in product stack in relation to ability to do multiple work units per card has less to do with the number of SM's but rather with the amount of memory on the card. The 1080Ti, Titan and 2080Ti have 11GB of memory whereas the lesser cards are all limited to 8GB or less. The OpenCL SoG application has a limit of 27% of available gpu memory for Nvidia cards. Somewhere around 40-60% for ATI/AMD cards. Two concurrent tasks is about all you can manage with lesser cards. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Shannon Lester Send message Joined: 27 Jul 09 Posts: 83 Credit: 12,388,119 RAC: 140 |
You wouldn't be able to do that many at a time without using the 2080 Ti which has enough memory for 4 tasks at a time. The differences in product stack in relation to ability to do multiple work units per card has less to do with the number of SM's but rather with the amount of memory on the card. The 1080Ti, Titan and 2080Ti have 11GB of memory whereas the lesser cards are all limited to 8GB or less. Way to ruin my fun Keith LOL! Hahaha I'm kidding though! Yeah that 11gb of memory would be for the best but I don't have that kind of money for a 2080Ti unfortunately. So pretty much youre stuck at 2 concurrent wu per gpu with the 1660, 2060/70/80 right? They can just crunch faster than what I have right now... All of this has happened before and will happen again -Battlestar Galactica |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Way to ruin my fun Keith Dang it Keith, I wanted to burst his bubble!!! Bahahaha......... |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Haha ha LOL, no Zalster it was my turn. Yes, Shannon the higher sku stack cards would just crunch the tasks faster because they can parallel more CUDA cores in the calculations due their higher SM counts. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13736 Credit: 208,696,464 RAC: 304 |
GPU upgrade question... I was thinking about upgrading from my 1060 6gb at the end of the year and was wondering about some of the newer Nvidia cards. How many work units would a 1660Ti handle at one time? 3 or 4? It's not a question of how many WUs a card can handle, it's a question of how many WUs per hour can it process. Grant Darwin NT |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.