Message boards :
Number crunching :
New Greenbank Files
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
Jimbocous Send message Joined: 1 Apr 13 Posts: 1853 Credit: 268,616,081 RAC: 1,349 |
Meow for the Lunatics! +1 |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Meow for the Lunatics! Amen to that, brother.... And another meow-out for them. A sometimes underapreciated bunch of dedicated souls. "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Interesting question is, will there be GPU GB work, as most GB stuff seems to be VLAR? Since VLAR work is currently turned off for GPUs on production (I have seen a good bit of GPU VLAR on SETI Beta) will that follow suit on GB? If so, will there be enough capacity to crunch the new volume of CPU-only work? Rather than speculating, keep an eye on what Eric says in the News area. http://setiathome.berkeley.edu/forum_thread.php?id=79411&postid=1778449#1778449 |
Jimbocous Send message Joined: 1 Apr 13 Posts: 1853 Credit: 268,616,081 RAC: 1,349 |
Interesting question is, will there be GPU GB work, as most GB stuff seems to be VLAR? Since VLAR work is currently turned off for GPUs on production (I have seen a good bit of GPU VLAR on SETI Beta) will that follow suit on GB? If so, will there be enough capacity to crunch the new volume of CPU-only work? Too many threads about the same subjects. Thanks :) |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
My first completed guppi - task 4854464991 - was deep in VLAR territory at AR=0.008804, but completed in roughly 75% (by time and flopcounter) of what an Arecibo VLAR needs. |
Brent Norman Send message Joined: 1 Dec 99 Posts: 2786 Credit: 685,657,289 RAC: 835 |
I have run 9 though, I'm seeing about 1:17 compared to 2:20 for MB VLAR |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
55 minutes for the CPU so far.. |
Cruncher-American Send message Joined: 25 Mar 02 Posts: 1513 Credit: 370,893,186 RAC: 340 |
Question about GB-VLARs: is it possible (has any testing been done?) that these new VLARs will behave differently when crunched on NVIDIA cards? Is it possible they won't take so long as the Arecibo ones do? Given that the data is said to be all VLAR (I think), it would be nice to know, so that the servers could be made to send them to GPUs... |
Mike Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80 |
Question about GB-VLARs: is it possible (has any testing been done?) that these new VLARs will behave differently when crunched on NVIDIA cards? If processed with OpenCL app yes. But no mechanism to distinguish OpenCL app from cuda so far. With each crime and every kindness we birth our future. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Question about GB-VLARs: is it possible (has any testing been done?) that these new VLARs will behave differently when crunched on NVIDIA cards? Talking of NVidia OpenCL apps, has anybody had any problems with the latest drivers, perhaps especially 364.72 for Windows 10? I've had an email from Jacob Klein, drawing my attention to this POEM project message board thread: Do not use Nvidia driver 364.47. I'm about to try and reproduce his NVidia SDK sample failures, but with the Windows 7 version of the drivers - if I have time, I'll run a SoG bench as well. |
Bill G Send message Joined: 1 Jun 01 Posts: 1282 Credit: 187,688,550 RAC: 182 |
I am running 364.72 on one of my computers. I have noticed nothing out of the ordinary so far. (I installed that driver last night as that computer updated (fast track) and I had to reinstall NVIDIA drivers, so I did the latest ones). Hoping all continues to work well. SETI@home classic workunits 4,019 SETI@home classic CPU time 34,348 hours |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Now we just need an update to the user task list page that shows the GBT work. It would also be nice to be nice to filter the view of tasks by the processor type. Such as CPU, ATI, iGPU, NVIDIA, & maybe ARM. When you are trying to find tack down an issue for a specific processor type and there are two or three in the host it can become time consuming unfiltered. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
WezH Send message Joined: 19 Aug 99 Posts: 576 Credit: 67,033,957 RAC: 95 |
What about AstroPulse task from GBT? Will they be available in future, or are AP's only from Arecibo? |
Mike Send message Joined: 17 Feb 01 Posts: 34258 Credit: 79,922,639 RAC: 80 |
Got anyone some GBT tasks for AMD GPU already ? I only got 9 for CPU so far. With each crime and every kindness we birth our future. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Ran the SoG testcase bench under the 364.72 driver Jacob was worried about - no problem found (though my test was Windows 7, his Windows 10). |
Jimbocous Send message Joined: 1 Apr 13 Posts: 1853 Credit: 268,616,081 RAC: 1,349 |
SSP shows 5 splitters active, but SSP only shows one tape getting split, and only 1 channel in progress on tape(dark green block) Should that not be 5 blocks, for consistency with Aricebo stuff? Now that more tapes are being split, it looks as though the above is incorrect, and that in fact a single channel is a very, very skinny green, barely visible. Makes sense, I guess, given the scale a larger number of channels per tape means each channel is a tinier piece of the bar chart. Tired eyes here? |
betreger Send message Joined: 29 Jun 99 Posts: 11361 Credit: 29,581,041 RAC: 66 |
Because AFAIK most of the GB data is VLAR and the paucity of APs I can foresee many of the big Nvidea GPUs will run out of data. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
So do we yet have good estimates on these new work units? I'm seeing a wide fluctuation on times 4854604170 angle 0.008804 Time to complete 53 min 4 sec 0 spike 1 autocorrect 4 pulse count 1 triplet 4854604378 angle 0.008804 Time to complete 2 min 11 sec 4854604345 angle 0.008804 Time to complete 11 min 2 sec 14 spike count 0 auto 16 pulse count 4854604470 angle 0.008804 Time to complete 51 min 27 sec 1 spike 5 pulse count So I'm not seeing any patterns yet. Anyone else? |
OTS Send message Joined: 6 Jan 08 Posts: 369 Credit: 20,533,537 RAC: 0 |
Because AFAIK most of the GB data is VLAR and the paucity of APs I can foresee many of the big Nvidea GPUs will run out of data. When both Green Bank and Arecibo WUs are being generated, I wonder if it would better to distribute the work load between GPU and CPU differently, i.e., limit CPUs to only VLAR work and GPUs to only non VLAR work. Right now CPUs can run both and GPUs, at least Nvidia GPUS, can only run non VLAR which heavily favors the CPUs. I would think it would be more productive to ensure there is work for both CPU and GPU by making sure that CPUs are doing work that only the CPU can do so that there is work for the GPU. |
Jimbocous Send message Joined: 1 Apr 13 Posts: 1853 Credit: 268,616,081 RAC: 1,349 |
Because AFAIK most of the GB data is VLAR and the paucity of APs I can foresee many of the big Nvidea GPUs will run out of data. Yeah, worried about that myself ... |
©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.