Big Progress in FAH Daily Production

I'm going backwards. my "BA" rig hasn't picked up a work unit in 2 days, I thought the room was cooler than usual - now to find out what the problem is

Yep, same thing happening on my 4P for the past few days as well. Just glad that it isn't my rig, and obviously a Stanford issue.
 
......Wonder what happened to Grandma. She dropped way down.

I am assuming that plext0r was a huge part of team Grandma. See above post where he states he is moving his rigs to another team.

Outside high temps here have only gotten up to the low 70's, but the basement is already hovering around 82F with all these 980Ti folders.
Looks like I will need to gradually bring the farm down now so I won't need to drink Gatorade just to sit in front of the PC.
 
I had to delete the current unit and start over- I forget the unit but it was hanging after a few frames... Lost 3 days :(
 
Don't think it's anything like that, cause my 4P system is straight Linux and it's not seen anything in a few days. I've already tried rebooting and deleting the work prior to firing F@H back up again.
in your case, can it be an empty work server ? assuming semi-BA might run short in supply ? I run my little cpu:3 and cpu:6 without any supply issue.
 
I've posted over at FF, it looks like its a lack of WU for machines with more than 24 cores. Kasson's lab is the only one that supports them and none of his servers are currently online :-(

I've installed v7 client and set up a pair of 24 core slots, PPD sucks and I can't get hfm to monitor it:mad::mad:
 
I am assuming that plext0r was a huge part of team Grandma. See above post where he states he is moving his rigs to another team.

Outside high temps here have only gotten up to the low 70's, but the basement is already hovering around 82F with all these 980Ti folders.
Looks like I will need to gradually bring the farm down now so I won't need to drink Gatorade just to sit in front of the PC.

Ahh seems you are right. I didn't even put 2 and 2 together.

That's not good. I think you'll still hit the 1B mark though. Assuming you haven't already since the stats have been down since sometime yesterday night.
 
do you have more details ? what indicate "hanging" ? CPU or GPU ? Windows 10 ? Over in FF there is a thread going on with hanging systems (mainly FAHControl); Folding Forum • View topic - FAH control panel hang

I have a Sandy Hex, Win10, no GPU folding (amd7970), it would fold a few frames (according to the log) and then just freeze, no ability to access the panel, no CPU cycles in use, couldn't connect to the web interface. After reboot, same; a few more frames and hung/paused with no ability to get into the program to see anything. Cleared the current work unit folder and it immediately grabbed a work unit and started folding again. If it happens again I'll write down the job number.

Ah- looks like I crashed again:

***************************************
*------------------------------*
Folding@Home Gromacs GB Core
Version 2.27 (Dec. 15, 2010)

Preparing to commence simulation
- Looking at optimizations...
- Created dyn
- Files status OK
- Expanded 825672 -> 1402860 (decompressed 169.9 percent)
Called DecompressByteArray: compressed_data_size=825672 data_size=1402860, decompressed_data_size=1402860 diff=0
- Digital signature verified

Project: 9038 (Run 887, Clone 2, Gen 24)

Assembly optimizations on if available.
Entering M.D.
Mapping NT from 10 to 10
Completed 0 out of 250000 steps (0%)
Completed 2500 out of 250000 steps (1%)
Completed 5000 out of 250000 steps (2%)
Client no longer detected. Shutting down core

Folding@home Core Shutdown: CLIENT_DIED

******************************************

That's on a freshly installed client (well, less than a week old)
 
Last edited:
Well, 2 24 core slots gives 120k ppd, 1 slot gives 100k - i'll just run the one I think.
 
Yea, sooo.... I was digging around the C drive, and noticed that permissions were jacked... No Mojo, but I think tinkering after a little liquid happiness might have been the issue. Wiped the SSD and reinstalled everything fresh (Win10 slow ring, etc), and all seems happy now.
 
The Mojo story and a pink Athlon would beg to differ with your assessment to not drink and fold.......
 
Hard to quote using a phone but here goes..... from relics post

Thanks to Sol, who saved the story from the great forum purge.
Here it is, in all it's painful, box-mashing glory:

The Mojo Story

..and so it began....
...sitting on my kitchen floor, building a new DC box while indulging in some of the finer versions of ethanol-based liquid refreshment. Halfway through the boxen building, I realized two things....

1. I was out of good scotch.
2. I hadn't started mixing the "mojo" for the party.

Now "mojo" is a particularly vile mixture of pure grain alcohol, Cherry CoolAid powder and chunks of citrus fruits. (Please note the lack of water or any other diluent)

Mojo recipe:
4 gallons (~16 litres if you care) of 97% ethanol.
8 packages of sweetened cherry Cool Aid.
various oranges, limes, lemons, old shoes...cut into large chunks
Mix thoroughly, with bare hand, while chanting "Nothing good can come of this."
Place outside in snow to cool. (keep animals away! This stuff may kill anything smaller than a camel!)

Somewhere around the "mix thoroughly" part, the whisky, which I'd been drinking to aid in building the new DC box, kicked me in the back of the head......Hard.

This scattered my data, and made my numbers go all random, causing a nasty chain reaction of stumbling, losing coordination and dumping 4 gallon of noxious red liquid into a brand new tbird.

I don't mean "splashing a little on the box". I mean pouring 4 gallons of mojo directly into an open case, a direct hit on the northbridge. Now, as we all know, cases are not watertight. The mojo started escaping into every corner of the kitchen. I sprang into action in an attempt to contain the dangerous stuff.

Unfortunately "springing into action" isn't very easy to do when you've just polished off a bottle of whisky. So I sort of "stumbled into mayhem" instead.

My left foot placed itself directly into the PC case, crunching parts galore, my right foot then decided it wanted no part of this and left for vacation. This had the unfortunate result of leaving me with no means of maintaining my upper body's position above the floor.

Please pause here for a visual reference.

relic, dumbfounded look on his face, stained red to his crotch with mojo, one foot in a PC case, the other slipping radiply away causing an awkward "splits" position...with floor awash in red liquid.

I did the only thing I could do. I fell forward, leading with my face, into the ocean of mojo on the floor.

The resulting splash was absolutely amazing.

Bright-red, ethanol-disolved coloring reached the ceiling.

Tendrils of mojo snaked past the cabinet doors and coated the clean dishes and food in the pantry. The telephone immediately took on a pastel pink color as the mojo ethched it's way into every surface.

The moral of the story? If you remember nothing else I've said....at least remember this....never build boxen on the kitchen floor. Happy new Year.
 
The boxes folded just fine, and over clocked like crazy, but everything was permanently stained pink.
 
Not sure about you guys but most of my farm is idle and think this thread is why. :mad:

Yeah the bigadv server is down, I've reconfigured my rig to smp 24 to at least fold something but I've had to install the v7 client to get any work:mad:

Kassons servers are all off line at the moment so this may take a while to get fixed
 
Thanks. The dead stats server, the very day I was predicted to cross 1B, had me thinking karma was getting me for some act I could not remember committing. :shifty:
 
Every once in a while I get the urge to drop some $$$ on new hardware just for F@H, but it's things like this little FUBAR that remind me why I've scaled back to just one 4P and another PC or two crunching data for Stanford. As someone who's contributed for many, many years, it's discouraging to see that some things never really change at the top tier of this entire project. I've not really been able to get behind any of the other worthwhile causes that you guys contribute to, I guess because of the personal ties to the things that F@H was been trying to help resolve. Guess I'll just keep on keeping on... <sigh>
 
keep the faith, 2 new CPU cores have just been released for very early internal testing, core A7 & A8. A8 is the long rumoured avx core.

Fully understand the not wanting to invest - apart from a GPU I'm not adding anything this year.
 
Nathan_P, I still have 40 idle hosts. What the heck is going on? Are they silently removing support for the 6.34 client and forcing people to V7? I have no desire to switch to V7 at this time. I've seen no site admins or professors comment on the thread about all servers down.
 
The professor is gone ... He has a new job ...

I think plext0r was on about Kasson, not VJ.

There is still no update, I'm not sure if there is any work still available for the v6 clients. I didn't get any when I switched to straight SMP, I had to install v7. If you are using the [H] appliance its a 5 minute job, but that's not ideal when you have 40 hosts. When my current WU is done i'll see if I can pick up any smp work on v6.

Edit:- nope nothing available, smp 48, smp 32 and smp 24 fail to pick up any work on the v6 client
 
Last edited:
I think plext0r was on about Kasson, not VJ.

There is still no update, I'm not sure if there is any work still available for the v6 clients. I didn't get any when I switched to straight SMP, I had to install v7. If you are using the [H] appliance its a 5 minute job, but that's not ideal when you have 40 hosts. When my current WU is done i'll see if I can pick up any smp work on v6.

Edit:- nope nothing available, smp 48, smp 32 and smp 24 fail to pick up any work on the v6 client

So Nathan_P - you're saying that things are working under V7 okay? I was running V7 under Win Server for a while but the PPD was horrible, and I could only utilize 32 cores under Windows. That's one of the primary reasons I set things back up under Linux again. I'm not fond of the idea of restricting my hardware down to 32 cores again, no sense wasting that kind of processing power.

Is there any support for V7 under Linux at all? I thought it was all Windoze based anyways.
 
V7 will run under Linux, I got it up and running in about 5 minutes on the [H] folding appliance, download F@H client, download F@Hcontrol and off you go. However I've not found work for anything more than 24 threads, so you would have to set up more than 1 slot to get the most PPD. I'm currently getting 120k from a single smp24 slot, for intel it probably makes sense to just fold on the physical cores. If you have an AMD 4p I'd set up 2 smp 24 slots and see how you get on - you may see 200k or more.

tpf on a project 903x is currently 22sec giving around 120-125k ppd on a pair of 12 core ivy xeons at 2.4ghz running on just 24 threads. bigadv was giving about 170k ppd so its a drop but better than nothing, what I can't currently measure is power draw - need to move the killawatt
 
/sarc After reading a portion of that article (and throwing up in my mouth a bit) I am not sure how F@H will ever continue without him. /endsarc
 
I moved my farm back to BOINC (POEM & WCG) until the BA servers are put back online or we figure out some other way to take advantage of 32-core+ systems.
 
Back
Top