| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: Work unit availability |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 822
|
|
| Author |
|
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges: 
|
Hi,
----------------------------------------I have 4 GPUs that only one has receiving work units (rtx2070). The older intel and nvidia GPUs have not received a single GPU wu in a week. Is this expected? The rtx is receiving ~30-40 a day, all the machines has the same settings. The older GPUs are for example Intel HD Graphics 520 (laptop) and Nvidia Quadro 2000M (laptop). Are these simply too old for the task? |
||
|
|
sam6861
Advanced Cruncher Joined: Mar 31, 2020 Post Count: 107 Status: Offline Project Badges: 
|
My Ryzen 2700x computer containing both NVidia GT 1030 and AMD RX 5500 XT, I have found that most of OPNG tasks is 90% chance it is sent to NVidia GPU, leaving my AMD GPU mostly idling.
Probably this happens: 1. Client request work for multiple GPU in a single request. 2. Server checks NVidia GPU work first, and send a few. 3. Server checks AMD/intel GPU work next, and have none left. Server ran out of available GPU tasks. 4. Request completed. Client only got a few NVidia GPU tasks and most of the time, no AMD GPU work. There are a few invalid/bad work units. Have 1 bad one. https://www.worldcommunitygrid.org/ms/device/...s.do?workunitId=619082847 This work unit, OPNG_0002254_00139, is invalid, invalid, invalid, server aborted, and my computer's server aborted. |
||
|
|
Dayle Diamond
Senior Cruncher Joined: Jan 31, 2013 Post Count: 452 Status: Offline Project Badges:
|
Congratulations on this natural experiment, and thank you for upping the ongoing GPU work to 96,000 units a day.
I completely understand that WCG has to be sensitive to the needs of all projects, and how, if things go awry, one project can inadvertently harm the computation of other projects. The addition of COVID-19 CPU units in particular have caused catastrophic slowdowns. On November 22th, 2019, MIP processed over 228 years of work units. Yesterday, they processed 22 years. That same day, MCM processed 341 years of work. Yesterday, it was 265. I remember advocating for a change in work unit share at the time and am thankful that the other projects made room to accommodate high priority COVID calculations. Now that we're approaching the point where we can accommodate all projects at once, I look forward to the day when the vast majority of these disruptive CPU tasks are converted to run in their own slot so the other project timelines can return to their pre-pandemic estimates. Hopefully the back end handles this inadvertent stress test without too many issues, and that going forward the project scientists weigh any temporary disruptions to their science against the promise that their ongoing throttling does not remain the 'new normal'. |
|||
|
||||
|
maeax
Advanced Cruncher Joined: May 2, 2007 Post Count: 142 Status: Offline Project Badges: 
|
Have OPN+OPG and MCM running on 7 PC's.
----------------------------------------2/3 are MCM. See no problem in performance so long.
AMD Ryzen Threadripper PRO 3995WX 64-Cores/ AMD Radeon (TM) Pro W6600. OS Win11pro
|
||
|
|
erich56
Senior Cruncher Austria Joined: Feb 24, 2007 Post Count: 295 Status: Offline Project Badges:
|
from what I gather reading here, quite a number of people have received GPU tasks these days.
The latest ones I received came last week, just a few ones within several days, that was it. Have there been any changes in the background since then, so that I would have to amend my device profile settings? What is set now is: a check at "Open Pandemics - COVID19" and "YES" for Graphic Card use. |
||
|
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges:
|
Hi, I have 4 GPUs that only one has receiving work units (rtx2070). The older intel and nvidia GPUs have not received a single GPU wu in a week. Is this expected? The rtx is receiving ~30-40 a day, all the machines has the same settings. The older GPUs are for example Intel HD Graphics 520 (laptop) and Nvidia Quadro 2000M (laptop). Are these simply too old for the task? 14/04/2021 14:28:25 | | OpenCL: NVIDIA GPU 0: Quadro 2000M (driver version 377.83, device version OpenCL 1.1 CUDA, 2048MB, 1686MB available, 422 GFLOPS peak) this one atleast seems to be too old |
||
|
|
kittyman
Advanced Cruncher Joined: May 14, 2020 Post Count: 140 Status: Offline Project Badges: 
|
Unless perchance there is a newer driver for it?
----------------------------------------Meow?  |
||
|
|
goben_2003
Advanced Cruncher Joined: Jun 16, 2006 Post Count: 145 Status: Offline Project Badges:
|
Unless perchance there is a newer driver for it? Meow? Sorry kittyman, I think this is not an issue of driver support. The Quadro 2000M is a Fermi generation card, which appears to only supports openCL 1.1. Newer generations do support 1.2 as shown by the driver release notes. driver release notes 377.83  |
||
|
|
kittyman
Advanced Cruncher Joined: May 14, 2020 Post Count: 140 Status: Offline Project Badges:
|
Unless perchance there is a newer driver for it? Meow? Sorry kittyman, I think this is not an issue of driver support. The Quadro 2000M is a Fermi generation card, which appears to only supports openCL 1.1. Newer generations do support 1.2 as shown by the driver release notes. driver release notes 377.83 Ahhh...I was not sure about the hardware itself. No go for that one then. |
||
|
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges:
|
I dug up an old dell laptop with AMD and intel GPUs (both cl1.2+) and AMD got work instantly. I have now three cl1.2+ intel GPUs without work. Could it be, that there are so many intel GPUs around that there is simply no work for them ? (no work = not a single WU for a week)
---------------------------------------- |
||
|
|
|