LTT Folding Team's Emergency Response to Covid-19

[ObeliskAG]

1 hour ago, Kilrah said:

I was going to ask how long a WU takes on a 200LX

[KickedAbyss]

Late to the game on this, but is LTT still contributing / looking for support? I've a xeon laptop and quadro M sitting doing nothing at the moment. 

 

Haven't folded much since 2005, but at the time my team of two mysteriously had approximately the same number of folding computers as may have been in a school district we may have attended. I mean, those Prescott HT P4s weren't being pushed by teachers or students....

 

#HighTechHicks 

[TVwazhere]

7 minutes ago, KickedAbyss said:

Late to the game on this, but is LTT still contributing / looking for support? I've a xeon laptop and quadro M sitting doing nothing at the moment. 

The LTT Folding team folds 24/7 365 days a year. We're always contributing and looking for support  

[Grounded]

20 hours ago, DontLickDoorknobs said:

But why does your graph go down to zero at the end?  Must just be changing out some more hardware for even more points? 

lol you caught me.  I'm changing some stuff around and switching to Curecoin . Also after seeing LAR_Systems Creaser build I've got an idea for the rx580 i replaced and an AMD 2200G ITX build I have. Curious to see what kind of PPD onboard Vega can get to

[Mawortz]

Hello everybody,

This is my first time in the forum and first time posting. Even though I have been watching Linus videos since before he had 1M subscribers.

I just wanted to say thank you to the Linus team and the community for this great effort, making the videos and setting up a server for Folding@Home.

I work in a big software company and I have been trying to get them to support Folding@Home with compute power and storage but it's not easy.

Anyway, this is my contribution so far, I'm using a GTX 1070

Cheers

[KickedAbyss]

1 hour ago, TVwazhere said:

The LTT Folding team folds 24/7 365 days a year. We're always contributing and looking for support  

Okay I'm in!  Set it up on my desktop too for when i'm not working; that's only a ryzen 2600 and rx580 but it should do alright.  As soon as i get WU - laptop got some, but desktop is still waiting.  I'm guessing they have ones written for amd and some for intel (and amd/nvidia respectively) as it's erroring with non available for my configuration.

 

Any word on how LTT's server is handling things?  I haven't seen an update, and am not sure i want to dig through 500+ pages of the forum to find out haha.  I thought it was very cool setting up a server to help handle the ingest - I think I read somewhere they're up to 60tb+ a day now?  Insane.

[Pasi123]

I recently hit 5 million points

 

[TVwazhere]

1 minute ago, KickedAbyss said:

 I'm guessing they have ones written for amd and some for intel (and amd/nvidia respectively) as it's erroring with non available for my configuration.

It means they literally dont have WU's. Right now there's a shortage of jobs since a lot of people started folding after hearing that F@H was supporting COVID-19 projects. So it's all configurations are getting hit by a shortage really  if your systems dont get wu's after about 6-8 hours, pause and resume the slots, as each time the client looks for a job the wait time increases (I found one machine had not completed a wu in over a day and had 24 attempts)

[KickedAbyss]

Thanks, @TVwazhere - I followed up on what I read: roughly 6TB/hr which means close to 144TB a day... 

Now it's probably as much a human element as even server backend.  Talk about a good problem to have.

[sese317j]

Aditionally the main work server (fah1) which assigned round about 28000 jobs per hour ran out of disk space in the last days. Before that the overall assign rate was about 25000 WU's per hour higher. But in the last few days some new work servers running on AWS showed up in the list, so maybe it will get better soon.

https://apps.foldingathome.org/serverstats

[LAR_Systems]

3 minutes ago, sese317j said:

Aditionally the main work server (fah1) which assigned round about 28000 jobs per hour ran out of disk space in the last days. Before that the overall assign rate was about 25000 WU's per hour higher. But in the last few days some new work servers running on AWS showed up in the list, so maybe it will get better soon.

https://apps.foldingathome.org/serverstats

This is my hope.  AWS systems configured correctly should not have storage limits and could take the manual labour out of cleaning returned work units on collection servers etc. for the F@H team.

[LAR_Systems]

3 hours ago, Grounded said:

lol you caught me.  I'm changing some stuff around and switching to Curecoin . Also after seeing LAR_Systems Creaser build I've got an idea for the rx580 i replaced and an AMD 2200G ITX build I have. Curious to see what kind of PPD onboard Vega can get to

Creaser has my 3200G test CPU in it.  The APU claims to get up to 390000 PPD in the advanced client on certain work units, less on others.  This is with it only driving the tiny 720P display so mileage may vary with a 2200G and if it's used on a higher res display or headless.

[leadeater]

11 minutes ago, LAR_Systems said:

This is my hope.  AWS systems configured correctly should not have storage limits and could take the manual labour out of cleaning returned work units on collection servers etc. for the F@H team.

Researchers will always need to manually go in and look at the results, some of it can be automated but there's always going to be a manual element to it. FAH's biggest problem currently is human power, people have to create protein start points for the simulations (Projects/WUs) and have to check the results at the end.

 

Azure has the same storage capabilities and AWS does, FAH does not use any current methods to move storage around and you can bet both Microsoft and AWS are not giving FAH blank cheques to have unlimited storage sizes on more premium storage. Good will only goes so far 

[LAR_Systems]

Just now, leadeater said:

Researchers will always need to manually go in and look at the results, some of it can be automated but there's always going to be a manual element to it. FAH's biggest problem currently is human power, people have to create protein start points for the simulations (Projects/WUs) and have to check the results at the end.

 

Azure has the same storage capabilities and AWS does, FAH does not use any current methods to move storage around and you can get both Microsoft and AWS are not giving FAH blank cheques to have unlimited storage sizes on more premium storage. Good will only goes so far 

I was meaning more just elastic storage for collection of WU.   From what I gather the latest server / WU issue is that they are currently manually downloading and then freeing up space on the collection servers.    These day I would assume there would be corporate good will and perhaps a gates foundation / federal grant to not loose all this processing potential because of something as easy to solve as hard drives are full.   

Give the poor guys some elastic storage and let them empty it out regularly while still allowing them to sleep if they get behind.

[Bitter]

15 hours ago, Orecal said:

I saw a few reply's here with my same issue, I have been getting CPU WU's like crazy to day but GPU WU's are far and few my daily points are down over 50% while not losing any clients. Is this happening to many others or just a few of us?

Ditto!

[Grounded]

14 minutes ago, LAR_Systems said:

Creaser has my 3200G test CPU in it.  The APU claims to get up to 390000 PPD in the advanced client on certain work units, less on others.  This is with it only driving the tiny 720P display so mileage may vary with a 2200G and if it's used on a higher res display or headless.

Very Cool. What is the total system PPD with the 1060?

[walderston]

1 hour ago, LAR_Systems said:

I was meaning more just elastic storage for collection of WU.   From what I gather the latest server / WU issue is that they are currently manually downloading and then freeing up space on the collection servers.    These day I would assume there would be corporate good will and perhaps a gates foundation / federal grant to not loose all this processing potential because of something as easy to solve as hard drives are full.   

Give the poor guys some elastic storage and let them empty it out regularly while still allowing them to sleep if they get behind.

The main issue for them is generating new protein trajectories. This is a manual process which needs human intervention. 

The collection servers also just dump everything to work servers when the work servers are ‘quiet’.

[walderston]

'COVID-19' option coming soon

 

[LAR_Systems]

2 hours ago, Grounded said:

Very Cool. What is the total system PPD with the 1060?

That depends on when you ask

So currently the PPD on the APU is sucking lemons due to the work unit... so total PPD is 400k... which is terrible

The PPD of the 1060 is consistently around 320k on it's own...  so never bank on the APU it's luck of the draw, discrete GPUs are still the way to go.

[REDYul]

Close to 25M! and new badge

 

My progress is the board is now getting slow lol

my target is to reach the top 1000 and with  optimism top 500 in the LLT team.

 

good luck healthy and keep folding

[GrimDemeanor]

We are supposed to Pm GOTSpectrum for the steam keys if we won one correct?

 

[walderston]

2 minutes ago, GrimDemeanor said:

We are supposed to Pm GOTSpectrum for the steam keys if we won one correct?

 

From the post - https://linustechtips.com/main/blogs/entry/1830-award-ceremony/

 

Quote

The winners will have 7 working days from the publication of this post to claim the prize, if the prize goes unclaimed it will be added to the community pool for use in future events. In order to claim the Prize you will need to PM me (@GOTSpectrum) here on the forum, with the subject

 

'Give me the key that I am owed!'

 

Also please can you include your rank to make it quicker and easier for me to get it all sorted!

 

[Bitter]

I got my Linux/P106 issue sorted, it was a config issue with FAHControl. Now I just can't get GPU work units to come through at all for any of my PC's, did they run out again?

[05MilMachine]

16 minutes ago, Bitter said:

I got my Linux/P106 issue sorted, it was a config issue with FAHControl. Now I just can't get GPU work units to come through at all for any of my PC's, did they run out again?

Took a few pauses to get them, but I have two running right now. I sat GPU idle most of today though.

[Grounded]

1 hour ago, GrimDemeanor said:

We are supposed to Pm GOTSpectrum for the steam keys if we won one correct?

 

I messaged them yesterday but still haven't heard back yet.