| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: OpenPandemics - Monthly Calls with Researchers |
| No member browsing this thread |
|
Thread Status: Locked Total posts in this thread: 58
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I've got immense respect for the COVID Moonshot folks. At the same time, I am immensely jealous of their abilities :p
I'm a total amateur in the computational chemistry space, and those "matter-hackers" are practicing pure black magic! |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
As an "amateur", assuming you are in the U.S., the best thing to do is contact you local Congressman. If they sense some possible favorable publicity to themselves, they can work wonders.
|
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges: 
|
Why posit it if you know the answer... would you tell "go take a hike" you Android / Linux Arm crunchers :? I'm not 100% sure on ADADELTA. I really doubt they'll re-run work but I'd like to know more because they say that the ADADELTA method is significantly better. As for the CPU part, it was mostly a "joke". I know WCG won't cut-off CPU-only volunteers from participating on a project. They wouldn't have even bothered with Android in the first place. Jim1438, Who knows. Like you said, FAH and FAH2 did that. Maybe they'll run a new software eventually for certain compounds/targets. But maybe thanks to ADADELTA, that won't even be necessary. I've long wished projects that offer GPU apps stopped offering CPU apps so those CPU's can be used on other projects. But I fully understand why that doesn't happen. subcosmos, Hello, we've talked once before on Twitter. Very cool that they are using your results from Quarantine@home on Folding@home. Like I said then, I hope your job situation improves.  - AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz [Edit 4 times, last edit by Falconet at Jun 24, 2020 7:50:09 PM] |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
We just finished our June call with the research team. Great news, thanks for the update !!!CJSL Crunch, crunch, crunch.... |
||
|
|
uplinger
Former World Community Grid Tech Joined: May 23, 2005 Post Count: 3952 Status: Offline Project Badges: 
|
There are some fun thoughts about GPU coming from the simple statement. :) We are VERY early (we just got a look at the code yesterday) in the GPU version of AutoDock for OpenPandemics. How the project will work on the grid will depend on what we see when we perform a beta. Since the last time we ran a GPU project on WCG, our systems and databases have been greatly improved upon. This means, we are going to explore every option as to the best way of distributing the work. If it means a separate project with different batches, then that will be the case. But with our backend being stronger, we may be able to just send the same work units and not have two different paths. We are excited on our end as well for a GPU project, but steps to get us there need to happen first on the backend and with an alpha/beta test. When we get to beta, we would love for everyone to help and give us a great idea on the path for production, but that is again in the future. When we get closer, we will post, but for right now, it's way too early.
Thanks, -Uplinger |
||
|
|
BladeD
Ace Cruncher USA Joined: Nov 17, 2004 Post Count: 28976 Status: Offline Project Badges: 
|
Nice thoughts on the GPU part of the project!
---------------------------------------- |
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3315 Status: Offline Project Badges:
|
Good luck, uplinger and team.
----------------------------------------- AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W - AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W - AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges: 
|
In Device manager you can set a device profile (school, work...) to use only GPU tasks and set this profile for a Device to use GPU only. That does not address what I'm talking about.  ...KRI please cancel all shadow-banning |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges:
|
...I'd like to know more because they say that the ADADELTA method is significantly better. This looks very interesting:https://ruder.io/optimizing-gradient-descent/ ...KRI please cancel all shadow-banning |
||
|
|
Aurum
Master Cruncher The Great Basin Joined: Dec 24, 2017 Post Count: 2391 Status: Offline Project Badges:
|
Since the last time we ran a GPU project on WCG, our systems and databases have been greatly improved upon. This means, we are going to explore every option as to the best way of distributing the work. If it means a separate project with different batches, then that will be the case. But with our backend being stronger, we may be able to just send the same work units and not have two different paths. It's never too early to plan for success. It's not about whether you can do something it's whether you should do it. If you hamstring the clients by giving us only one way to control both a CPU WU and a GPU WU then we'll be forced to pick one or the other. It's likely that the GPU version will run five times or more faster than the CPU version. The alternative is that over 80% of CPU WUs will be aborted in order to maintain the GPU work flow. If IBM does not think about this from the beginning how can you possibly produce an efficient work flow? Just because your servers can throw down a flood of WUs does not mean our computers can deal with them. It's also very likely that when Scripps get a GPU project fielded they'll redirect the CPU project to serve a different purpose. Some mathematical techniques and approaches may be better served by a CPU than a GPU. ...KRI please cancel all shadow-banning |
||
|
|
|