LTT Folding Team's Emergency Response to Covid-19

[walderston]

1 minute ago, Jordan Johnson-Verburg said:

Is the 3D visualization viewer broken?

I only see it cycling through Demo snapshots and not what is currently being worked on.  I don't need it of course, but thought it'd be cool to see what protein simulation was being worked on similarly to Berkeley's Boinc - Rosetta@home viewer/screensaver. 

https://gist.github.com/alfanhui/fd69de8354047609fad5ebbef829eb94

 

Follow that for a fix

[Inkertus]

Just now, jctappel67 said:

It doesn't work atm

I believe they mentioned it wouldn't work for certain jobs but assured us important calculations were being done anyway.  Think it was regarding chemical interactions instead of atomic-level motion calcs.

[ShortRouter]

10 minutes ago, walderston said:

you don't need to set it to static. 

 

The IP will only change if the VM restarts and most likely you'll get the same IP back. 

I don't see any point in making it more difficult. Just set static IP and forget it. Anyhow, you would also need to add VPC network firewall rules in Google Cloud settings.

[ShortRouter]

9 minutes ago, Guellan said:

So I found if you go to iCloud store, there is a nvidia quadro virtual workstation image...no need to install drivers and it runs on ubuntu.

 

Now for the question...how do I set the "Advanced" flag in the command line?

On Ubuntu, you modify the FAH config file:

 

sudo nano /etc/fahclient/config.xml

so it looks like this:

<!-- Client Control -->
  <fold-anon v='false'/>
  <!-- Folding Slot Configuration -->
  <gpu v='true'/>
  <client-type v='advanced'/>
  <!-- Slot Control -->
  <power v='full'/>
  <!-- Work Unit Control -->
  <next-unit-percentage v='90'/>
  <!-- User Information -->
  <passkey v='yourpasskey_should_already_be_here'/>
  <team v='223518'/>
  <user v='YourUsername'/>
  <!-- Folding Slots -->
  <slot id='0' type='CPU'/>
  <slot id='1' type='GPU'/>
</config>

next-unit-percentage I recommend to set to 90 with how slow the downloads are from many servers. It will download the next WU when current reaches 90% completetion. Second GPU is added via    <slot id='2' type='GPU'/>    etc. Save file to same name, same location.

[Kilrah]

1 hour ago, ShortRouter said:

Well since some of the top folders are using the ifrastructure of some datacenter they work at for folding and folding in the cloud is not prohibited...

Not against folding in the cloud, but violating the cloud providers' terms to get more then you're allowed to..

1 hour ago, Macaw2000 said:

My hope is that people here who are dabbling in cloud, use this time to learn better cloud skills and up their game. If you want to make the big bucks, learn to cloud and learn to automate the cloud!

Yup, but you don't need multiple accounts for that.

 

The terms of use of folding@home actually require that you have the permission of the hardware owner to use it before you fold on it. If multiple accounts are against the TOS of the provider then that would equate as not having the permission of the hardware owner.

 

Definitely should be taken into account for eligibility to the event rewards and ranking.

[ShortRouter]

7 minutes ago, Kilrah said:

Not against folding in the cloud, but violating the cloud providers' terms to get more then you're allowed to..

Yup, but you don't need multiple accounts for that.

 

The terms of use of folding@home actually require that you have the permission of the hardware owner to use it before you fold on it. If multiple accounts are against the TOS of the provider then that would equate as not having the permission of the hardware owner.

 

Definitely should be taken into account for eligibility to the event rewards and ranking.

I'm about to violate LTT forum account rules to upvote this more.

[Guellan]

4 minutes ago, ShortRouter said:

On Ubuntu, you modify the FAH config file:

 

sudo nano /etc/fahclient/config.xml

so it looks like this:

<!-- Client Control -->
  <fold-anon v='false'/>
  <!-- Folding Slot Configuration -->
  <gpu v='true'/>
  <client-type v='advanced'/>
  <!-- Slot Control -->
  <power v='full'/>
  <!-- Work Unit Control -->
  <next-unit-percentage v='90'/>
  <!-- User Information -->
  <passkey v='yourpasskey_should_already_be_here'/>
  <team v='223518'/>
  <user v='YourUsername'/>
  <!-- Folding Slots -->
  <slot id='0' type='CPU'/>
  <slot id='1' type='GPU'/>
</config>

next-unit-percentage I recommend to set to 90 with how slow the downloads are from many servers. It will download the next WU when current reaches 90% completetion. Second GPU is added via    <slot id='2' type='GPU'/>    etc.

Thanks. I have the 4 T4 GPU's added already, was just wanting to make sure they stay fed till friday. Looking like $1.44/hr.

[HaRdLy007]

I guess i have to start another instance to have 2 v100 or 4 P4's ?

[ShortRouter]

5 minutes ago, Guellan said:

Thanks. I have the 4 T4 GPU's added already, was just wanting to make sure they stay fed till friday. Looking like $1.44/hr.

I'm sure it's awesome PPD for the money. What region did have those T4s available? I gave up and went with 2x P100 but it's roundabout 2.5 € per hour.. I'd probably have to remove the dual P100 first to migrate because I only requested max 4 GPU total.

[Inkertus]

@El_Psy_CoNGR00 The panels are expected to show up by Tuesday.   Looking forward to those relatively frigid temps.  Might even try stock settings instead of the underclocked/volted I'm running now.

[Guellan]

8 minutes ago, ShortRouter said:

I'm sure it's awesome PPD for the money. What region did have those T4s available? I gave up and went with 2x P100 but it's roundabout 2.5 € per hour.. I'd probably have to remove the dual P100 first to migrate because I only requested max 4 GPU total.

us-east1-c

[Zberg]

Guys a positive note:

 

My hospital is turning a corner it seems.  The curve seems to be flattening (our town social distanced early and the town shut down pretty well), we have a good buffer of ICU beds, and it seems like the hospital is not overwhelmed and will not be as our cases are not jumping up the way they were at first.  We have enough PPE, tests, and vents to sustain this current rate.

 

We have a backup make-shift ICU that we havent even had to use because the actual ICU still has a few beds and vents (and there is a second good size ICU that has some more vents and beds)

 

If everything stays on track in my city and we keep it up, we likely will NOT be overwhelmed and have vent beds for all who need it.

 

Dont get me wrong, there are a lot of covid cases, and some of them are pretty nasty, and they have pushed our resources a bit.  But we havent been pushed to the point of being overwhelmed due to everyone jumping in roles to fill needs.

 

To be seeing it improve from the inside there, and the outside folding with you all has been awesome.  Lets keep it up, and keep it going after the event (there will still be plenty of research needed and proteins to fold with coronavirus when this event ends!

 

 

[Zberg]

Just now, Zberg said:

Guys a positive note:

 

My hospital is turning a corner it seems.  The curve seems to be flattening (our town social distanced early and the town shut down pretty well), we have a good buffer of ICU beds, and it seems like the hospital is not overwhelmed and will not be as our cases are not jumping up the way they were at first.  We have enough PPE, tests, and vents to sustain this current rate.

 

We have a backup make-shift ICU that we havent even had to use because the actual ICU still has a few beds and vents (and there is a second good size ICU that has some more vents and beds)

 

If everything stays on track in my city and we keep it up, we likely will NOT be overwhelmed and have vent beds for all who need it.

 

Dont get me wrong, there are a lot of covid cases, and some of them are pretty nasty, and they have pushed our resources a bit.  But we havent been pushed to the point of being overwhelmed due to everyone jumping in roles to fill needs.

 

To be seeing it improve from the inside there, and the outside folding with you all has been awesome.  Lets keep it up, and keep it going after the event (there will still be plenty of research needed and proteins to fold with coronavirus when this event ends!

 

 

a caveat, this is one city, and they are ALL different.  Be careful, social distance, and wash your damn hands!

[marknd59]

Damn Default just overtook for points today.   

[Takmikill]

doing 2.2 to 2.3 mil ppd.. I came in late but i got from 960 to 606 in a day, which is nice. I'm nearly at 100 mil overall now

[Zberg]

4 minutes ago, marknd59 said:

Damn Default just overtook for points today.   

I have noticed today is the first day in about 4 days I have had to pause/unpause so much to keep WU going.  For a good while it was going non stop always starting the next one as soon as it finished.  Today I have come up 3 times to the PC sitting idle with no WU and had to refresh 5 times to get a WU.  My points are down a little.  Maybe the WU are drying up a little

[marknd59]

2 minutes ago, Zberg said:

I have noticed today is the first day in about 4 days I have had to pause/unpause so much to keep WU going.  For a good while it was going non stop always starting the next one as soon as it finished.  Today I have come up 3 times to the PC sitting idle with no WU and had to refresh 5 times to get a WU.  My points are down a little.  Maybe the WU are drying up a little

The servers say the assign rate is 133,488.00/hr 

 

[marknd59]

Having said that when I look again 2/3 of those are CPU

[Zberg]

Just now, marknd59 said:

Having said that when I look again 2/3 of those are CPU

I dont know if that rate is good or bad.  My CPU has always been able to get WU but they are always tiny.  My GPU (both systems) have been going non stop for a good bit of days, but seems like today that stopped.  Maybe ill change client-type to advanceD?

[Alex Atkin UK]

Now were really cooking:

[GalaxyNetworks]

8 hours ago, Alex Atkin UK said:

Yeah I was struggling to get CPUs too, it said 0 availability even when I selected 16.  They make it WAY too hard to just setup a VM when it honestly doesn't matter WHERE its located.

If you don't want to upgrade your account or look at quotas, then yes... It's hard.

But it's rather easy if you just give it a few seconds. Region usually makes a lot of sense. In this situation you just want any GPU in any region, and they have limited GPUs right now, I know that's annoying. But this kind of usage is not what Google Cloud or any cloud is about.

Try Azure, you'd be just as annoyed with them! Just looked, I can't even find all regions in the quota tab, so I'd have to make a support ticket and wait for human response.

 

I just tried to make a Google Cloud preemptible instance with 4 T4's... Failed in US, succeeded in EU. And I could even clone it a few times... I don't have credits though, so I'll be terminating them again.

[Alex Atkin UK]

3 minutes ago, GalaxyNetworks said:

If you don't want to upgrade your account or look at quotas, then yes... It's hard.

But it's rather easy if you just give it a few seconds. Region usually makes a lot of sense. In this situation you just want any GPU in any region, and they have limited GPUs right now, I know that's annoying. But this kind of usage is not what Google Cloud or any cloud is about.

Try Azure, you'd be just as annoyed with them! Just looked, I can't even find all regions in the quota tab, so I'd have to make a support ticket and wait for human response.

 

I just tried to make a Google Cloud preemptible instance with 4 T4's... Failed in US, succeeded in EU. And I could even clone it a few times... I don't have credits though, so I'll be terminating them again.

As you see above, I folded (pun intended).

[Ayren]

I didnt dare to look at the rankings since the weekend but im pleasantly suprised \ ^ -^/
also @Macaw2000 you crazy [expletive redacted]
you werent kidding when you said you where going to flex your AWS muscles you werent kidding, huh?

[HaRdLy007]

I'm only able to have 1 GPU per instance, is there anyway of changing that?

[Alex Atkin UK]

Just now, HaRdLy007 said:

I'm only able to have 1 GPU per instance, is there anyway of changing that?

See a few pages back.