#folding-and-boinc
1 messages Β· Page 18 of 1
i think ill just leave it for a bit because my points per day on a 2080 and 8700k are currently 4714
lol
i think something like this would be sufficient
You need to have the username, team number and PASSKEY entered to get decent points, you get bonus points after 10 successful WU uploads
@buoyant valve Dont use PCIE riser with folding
well a few hours ago i was getting 260k estimated per day just from my gpu
you will get like 50-75% less points @buoyant valve
now its using 1.4% of gpu lol
what is the point of points anyhow?
Without being able to see what's happening you have no idea how to sort the problwm
@buoyant valve to make you fold harder than your friend
are they not a pointer of how effective your shit is?
They are a numerical value awarded to you based on the amount of work your systems have done. They exist as humans are so competitive its unreal
considering its just gonna be a server for that alone its gonna be set it and forget it
Like I said, don't use risers, especially 1x risers
you need 8 lanes per card minimum
that just means itll be limited to what 4 cards? wouldnt using 8x risers for an additional 3 remote cards be useful?
no because you will lose 50-70% production on the cards over 1x
i said 8x risers not 1x
You can use 8x risers but FAH is so latency sensitive its unreal. I personally wouldn't trust it for Core 22 especially. It should work most of the time but the likelyhood of slots failing is much higher. With that said, it should be fine for the most part.
well the original thought was to have 3 mounted to the mobo in the 16x slots and to populate the rest of the slots to the remote bays using 8x risers but if that wont work then i guess ill just get a shorter case and go with just populating what i can
so which is more effective? boinc or FAH?
So im using the FAH thing, but its not able to pick up my gpu, any help?
@buoyant valve Depends on what area of science you ar e wanting to help
I'm the folding team leader so I am obviously going to say F@H
a 24U case hmm
so anyone else find themselves with the "No WUs available" log error?
We are hammering the servers that hard that they are running out of WUs to offer.
Not we as in LTT but we as in everyone
mmhmm which I guess is a good thing
This is the default teams graph
also a good thing as this is a participation drive and not a contest :p
And this is our graph.
Quite a number of other teams are seeing similar gains so yes, the PC community has answered the call of the scientists
glad to see a lot of non-affiliated peeps folding
If you have any gamer friends please do ask them to join team LTT
i cant get my gpu running either
Lets see how hard we can push this
whats the team number for LTT?
for sure I already have my group on board
The forum died apparently
joined
Its been down for the better part of a day now
I gotta go guys sorry, @rich turtle can you help?
yeah I was having issues from last night so I'm assuming that a lot of machines got put online
thanks spectrum
doesnt even pick it up
you should be able to set it up under configure in the advanced control client
i tried
delete all the slots and add it manually maybe?
have you updated the nvidia drivers
yes
it's all under the slots tab in the configure window
very weird. other than those basics I don't have much more
@stark bough you could check craigslist, thats where i got my racks from
@stark bough mine were $40 each
@buoyant valve Someone was literally giving away a full size rack but I had nowhere to put it
yeah a full size rack wont fit in my room, short ceilings suck
mine are 30u
and they arent really deep enough
I'm showing no WUs for my CPU now
@warm plaza they are out of WUs so their server keeps crashing with all the requests for work
Okay, alternate question, can a Radeon HD 6950 do anything folding-wise? It shows up in the list of devices but I get errors when it tries to init
20:25:51:ERROR:WU01:FS01:Failed to start core: OpenCL device matching slot 1 not found, try setting 'opencl-index' manually
ive no idea, ive only just got my rx580 working on my server a few hours ago avter months of fiddling
personally i just have both boinc and FAH running right now since FAH has insufficient work available
Forum is back !
@spare glen You got a pic of the βNo WU errorβ
20:38:38:WU02:FS00:Connecting to 65.254.110.245:8080
20:38:38:WARNING:WU02:FS00:Failed to get assignment from '65.254.110.245:8080': No WUs available for this configuration
20:38:38:WU02:FS00:Connecting to 18.218.241.186:80
20:38:39:WARNING:WU02:FS00:Failed to get assignment from '18.218.241.186:80': No WUs available for this configuration
20:38:39:ERROR:WU02:FS00:Exception: Could not get an assignment
their servers have nothing available
Just installed it F@H, it was able to get one for the cpu but the gpu keeps trying to conect a 155.247.166.220:8080 and its returning a HTTP_SERVICE_UNAVAILABLE anyone else having this issue?
Ahh, no WU's guess ill just let it run π Happy Folding.
Hey all o/
hi
reading up I guess I'm not the only one struggling to get GPU work for F@H
no i have a gpu sitting around waiting for work
so that aside, I do still have a question if I may.. I've added LTT's team number and it seems registered to it however on the panel it doesn't load the info. Pic:
Been like that for 6 hours now
CPU isn't struggling to find work, only GPU and of course the stat pages and above ^
Does the actual program open, if you're on windows.
Yip, opens, been working with cpu, got some work for gpu earlier today. 4-6 hours ago
Yeh the website https://client.foldingathome.org/ just in a refresh loop for me also
22:21:56:WU00:FS00:0xa7:Completed 255000 out of 500000 steps (51%) is where im at
just shut everything down and reopened and still just CPU. oh well, will leave it running on CPU for now. Just concerned that there's no loading of the info above.
hmm 20 sec's a fold doesn't seem bad....
My specs could be better but I'm willing to throw everything to it so.. will also setup my other rigs to fold tomorrow
Was getting COVID-19 projects this afternoon, now I see it's giving me Project 13826, which is help with Usher syndrome
i wish they would make a covid-19 selection so we can just work on that
Yip.. I understand it would take time to push an update (cause apparently the drop down selection is coded in, not refreshable/changeable at will without an update) and they wanted to get us on covid-19 as fast as possible but really... it's been long enough now
well, F@H main stats page finally updated for me
3 LinusTechTips_Team 46,464,684,723 1,076,313 223518 +5
anyone know why my gpu isnt fully utilized?
what gpu is it
do you have the folding power set the medium or full?
it does - different projects do different utilization
though I've had some use 40% utilization according to windows and still my GPU was at the maximum temperature and clock speed sooo it was not indicative of "working hard"
oh and, im trying to connect my other pc to a client, and it seems like its stuck on connecting
that's happening to everyone - folding has had a 4x surge in contribution from a week ago - their servers for downloading new work units are not always responding
it eventually gets one
i mean, adding a client to the client list
guess I was just lucky to get my 4 Covid WU's today
i got a few yesterday with no problems
Oh well, gonna lave it overnight and will see what it's up to in the morning
Is there a LTT stats page for Folding clients using the team number?
https://folding.extremeoverclocking.com/user_list.php?s=&srt=4&t=223518
This is the one I stare at
heard about folding for a while and like many people I guess decided now was the time to see what it's all about, if anybody experienced has time time to school me, please send me a DM
got a bunch of
got f@h going on my PC and my wife's machine but I want to figure out how to optimize it, the PPD I see is way lower than what I see people talk about
seems my original setup used the wrong team number, so some of my score went to that team π¦
and seems I need to work on getting much more score on LTT team
rip
yeah so after finishing latest cpu unit, no CPU or GPU wu's for the last hour
did we win folding? did we fold all the things? or is it just their servers are derping due to the load
@frozen oxide can I pick your brain on this?
maybe, but I literally folded 16 years ago, then started again yesterday, so I don' t know that much
but technically there IS a post from one of the Dr.s at the University stating that we'd taken ALL work units.. and that they were working on adding more
so kinda? guess we won in that weird way
most peoples PPD will be lower than other peoples due to them having better systems and how many work units you do will affect you PPD
oh and the ones which may not finish, are being held for 6 days or whatever right
guys imma crush yall with my one cpu and 2 gpu's https://img.thaun.dev/nlmja.png
@timber grove I mean PPD that is tiny fractions of what I see other people talking about, on similar hardware
my PPD estimate was varying wildly from 70k to 1.9 million yesterday depending on if I got WU's or not
also it says my FahCore is 0xa7, and I keep seeing people say the newest one is 22?
a7 is CPU avx
yip Crast
22 is GPU's 9 series and up
wonder how many are being held and not worked on π¦
I have an RX 5700.
that too then
The distributed-computing project Folding@home uses scientific computer programs, referred to as "cores" or "fahcores", to perform calculations. Folding@home's cores are based on modified and optimized versions of molecular simulation programs for calculation, including TINKER...
many forum threads talking about that being the one that Navi cards work on, because old ones didn't
a7 is definitely a CPU worker
well only my CPU is working right now
my GPU got one WU, finished it, and now is waiting a long time
so I guess I'll wait and see what it gets
see above - it seems like most of the WU's have been taken
like there's no point in repeating work someone else might do, unless they crash or shut down and never do it
right
It's just that 6.99 day waiting period that sucks if they do abandon
yeah... I hope that they don't get tons of WU's showing up like that
in a week while we sit idle for the next roughly six days
Yip
And issue is, a lot of people shut down for the night etc. so that's entire night sessions of nada
ooh.. it wont be as good as GPU but.. I have an entire server room of CPU's I can throw at it
OK, my wife's machine has a GPU WU (on her 1050 ti...) and it's on fahcore 21
21 is older GPU's or something
rip, thats my gpu
but my 1080ti was running 22 soo that's weird that two different 10 series use different cores
when trying to look at my stats on f@h the website is either giving a bad gateway or no results, does anyone know why that is? Is it just an issue with their servers rn?
Their servers are being flooded
understandable
so.. if you hit refresh a lot it does eventually load
every 2-3 minutes I Just refresh and it loads up
I guess their servers being walloped is a "good problem"
yep.
Just sad that it always takes something terrible to bring people together
ok I'm out for now. Later o/
tbh I think its really cool. definately something I am going to talk with my biology teacher from last year about whenever my school opens again, she'd find it interesting
oh hey, the stats page loaded!
the GPU workloads also take like 30x as long to download for me as the CPU ones even when I'm not just looping forever waiting for a workload
my internet isn't that fast
so my GPU has spent a lot more time idle than my CPU
the funny part of that meme is that i have both an amd and a nivida gpu in my rig rn
how do you have that working? won't one go to broken driver when the other one gets a working driver?
unless one card is like... suuuuper old
the amd card is a r9 270x 4g and nvidia card is rtx 2070
man i wish i could just plug in another gpu in, but that wont matter since its draining my cpu anyway
I just got some core 21 GPU work... just as my CPU just finished a unit
Woke up with computer stuck on a BSOD screen.
Thankfully looks like it was running for about 6 hours before it crashed so I didn't lose too much
oof
Got a one man army going on here
https://folding.extremeoverclocking.com/team_summary.php?s=&t=236584
Oh, this is why...
https://www.coreweave.com/
yeah
On Tuesday night we began donating the compute power of over 6,000 GPUs to the @foldingathome projects researching Coronavirus (COVID-19). We are proud to be able to donate our flexible GPU compute to a project like this.
6,000 GPU's
well, GPU Cloud, holy crap π
At least now we know where all the WUs went π
same here, turned it up and got me a GPU WU. so watching a chill rimworld stream on twitch while my gpu seems to lift off any moment π
my GPU has a work unit and it's a core22 test benchmark?
I got one of those after my undervolt crashed my GPU yesterday
Yeah I've got a Core22 test benchmark WU as well
I wonder if they're just sending those out when there isn't any other available projects
i am also currently working on project 14310 on my GPU
same
oh great... now i need another isolating transformer
side note
(39, 64, 0)
if you right click the tray icon and go to advanced control
try and do the "finish" button to finish current workloads and not leave them hanging I guess
yea
will do that when i go to sleep, so it will finish current WU's but not get new ones
I just booted my machine which crashed last night, limited it to half the cores, so it can finish the workloads it committed to (it still needs tweaking it's not a stable machine, but that's a project for another day)
it had both cpu and gpu workloads from last night, while my beefy video editing machine is like... plz give me work plz for hours today
I'm still waiting on assignment too
stopped after my assignment this morning and my CPU got a new job, but my GPUs never got new ones and they aren't getting anything :[
anyone know the workaround, or where I can find it to get the viewer working? idk I just think it'd be fun to watch
(for f@h
From a bit of googling it seems the viewer hasn't worked right for a while, though I thought I could remember it working a few years ago
yeah, I heard there might be some kind of workaround, but I haven't found it. The default protein is somewhat interesting though XD
its time for Linus to help f@h build a storinator
Oh lord haha
could be a good way for them to spread the message about their team for the f@h project
Techlinked and WAN show already boosted it pretty healthily
when I was watching WAN last night is when I redownloaded it. Been too long
OH
one of my GPUs just got a job!!
yeah I'm surprised I never used it before. I am pretty interested in tech, chemistry, and biology so this is pretty spot on for me. I guess I just never had anyone explain what it actually did until now.
my cpu got a job after a restart just now and my second gpu is trying to get one
my gpu is at it rn with project 14310
I've never measured my computer's consumption, and I already gotta pay a high electric bill for my security deposit because I moved apartments/jobs, figured I'd let er rip tater chip anyway
Perfect way to christen attaining my dream job/company
if I pause my fold is it safe to shut down my pc for a few minutes?
Yes
my computer pulls a little less than 600watts
@pastel nacelle My second GPU also got 14310
I have 2 1080s and a 4790k, with a 1000w gold power supply. It heats my apartment lolol
#space heater
eyyup
rip
ok well before I make a big mistake, is it worth it to let a radeon hd 3450 that I have run some folding or should I not even try?
no
I don't know... does that even have compute API's?
idk, so ima leave it out
does anyone else see PPD differences between 14310 jobs and human malware related jobs?
On my Radeon VII i'm seeing like half the PPD estimate doing only 14310 WUs
It could be because I'm getting them so infrequently though
different gpus like different work units
I just got a job on my gpu!
the PPD estimate is hurt a lot yeah by the idle time - when I wasn't getting units for a few hours the PPD dropped to like 50k
anyone know where I can figure out what 11762 is
hax
Ok another question. Does it take special configuration to fold on multiple computers in the same network? My gaming PC is getting jobs way more frequently than the one in my guest bedroom.
My old roommate sent me that pun when I told him F@H was doing covid lmao
no it doesn't - that's just a bit of luck tho
but many people aren't getting new jobs at all for hours at a time on either their CPU or GPU or sometimes both the last 6 hours
I managed to get a few covid wu's yesterday
afaik F@H does not benefit from SLI - there's a post about it
and instead run it on each GPU individually as separate cores
...presuming you get WU's
Yeah, SLI doesn't work...
Run 2 slots one for each GPU and they will work on projects independently
make sure there's 2 GPU slots
then you probably just don't have a work assignment for the other GPU - most people aren't getting them at all the last few hours
no I have on on both at the moment
check your status page it says what each slot is doing
@pastel nacelle I think the majority of my WU were covid related for the past couple days until today's drought
nice
only started yesterday but I did some covid CPU and some covid GPU and a bunch of cancer GPU
I have one cancer on GPU, and a covid on the other GPU and the CPU
Is it bad to have PBO2 on while doing CPU folding?
if your system overheats, then yes, otherwise, no
when my recent wu loaded it said 24 hours at first, but now that it is going at it, the time is down to 1.5 hours so thats nice
tbh its pretty cool that the tech community can be a part of solutions like this
Ok. Looking at logs the 3600 in my guest room did some CPU WU in the past couple hours and I don't want to send out garbage data lol
well also a bunch more big companies made posts on their official twitter in the last day heh
also a lot of discords and subreddits did their own teams
If Popey from the Linux and podcasting community folded with all his Thinkpads in his horde, I think it would bring about the end of the world lmao
Seems like big companies joining could easily do all the GPU and CPU work quickly
also probably bring down the power grid xD
I like the tweet where the folding project dared google stadia to donate some GPU's
not like they are doing anything else
now that the 90 days of stadia early adopters has ended, I'm sure they have plenty 
well all the preorder people got 90 days of stadia included
i have a pc with one of those wish 1050s and a core 2 quad XD I should not leave it running some folds XD
(i got the gpu from a friend who was unfortunately a victim of the scam)
do it
tbh I would, but I dont trust the psu in that pc
I just built it because I realized I had enough junk parts to actually make a functional pc
just got a work unit for my amd gpu
I just got a mystery 11762 GPU workload on my mini-ITX system @pastel nacelle
nice XD
same
hmm I turned off SLI and it didn't really change, must just be a difference in the workload for the different WUs
and a 14304 on my cpu
I found a F@H / PCMR post on imgur now
@void lynx my single GPU 1080ti is at "5%" utilization in windows and hitting 76C so I don't think it's doing "little" work it's just little as how windows chooses to interpret it - compute is different from rasterization
Imma have to agree with you on that
one is at 40% and the other is at 1% lol
but afterburner is clearly showing activity lol
you can use cpuid hwmonitor
I should really move one of these rigs to the bedroom where I run a space heater starting at 9pm... except now they're both on WU's and I don't wanna jinx it
(I posted the scrap rig in the my-rig channel
actually have the window cracked in my office, and in 2 hours I'm normally gonna start a space heater in my bedroom, seems counter-productive
I have a dual machine, dual xeon server at my old apartment I'm going to put in my parents basement to fold on... used to be my plex server, but 60db at idle was, a bit much in an apartment
oof
@pastel nacelle I have a Dell Precision T3400 with a Core 2 Quad Q9650, and I tried to use a Titan RTX in it for some folding but it won't output an image in a machine this old. π¦ My 1050 Ti works in it tho, and the Titan works in my other newer machines.
@void lynx waiiiiit
yis
it would be pretty cool if the titan actually worked in it XD, do you think it was a psu issue? or compatibility
this is my GTX1080ti system... look at how much afterburner and windows disagree on utilization
its the cuba
here's my mini-ITX (gtx 1070) - again disagree but by a different amount
Mine is the same
TLDR we found the culprit - task manager knows jack
windows, big friggin surprise
Not sure if it is a PSU issue... I replaced the stock Dell PSU with a 500W CoolerMaster 80 Plus Bronze unit.
if I wasn't bound to windows because of oculus my rig would be on Linux again
I've tested a Radeon VII in it so idk... the Radeon VII worked.
Plenty o thinkpads with linux though π
i see, thas too bad, it's a cool concept though!
cuda goes high when it runs
Guess I could try one of my Titan Xp Empire edition cards lol
XD
omg my mini-ITX has wifi I forgot... derp BRB
Windows be like "you can use hotkeys to launch the snippet tool!!" And I'm like, uh, no, my keyboard was made before Windows existed but thanks anyway.
No Windurs key for me!
I too have a discrepancy
Yeah Windows is dumb... at least multigpus works in it though. Linux hasn't had proper SLI support since, uh, a long time. Using the x server with more than one GPU is a nightmare...
i considered buying a second 1080 when I saw someone posted one that was the same model used for a good price, but I honestly dont need sli rn (though i lowkey want it XD
Space heater from 1920... meet space heater from 2020
nice
Also I know I shouldβve set it up for remote access to not drag a monitor into my bedroom
Just joined the folding team π

just a single 2070super, but doing ok.
Played w/ underpowering it today to keep it cool / quiet in my office
put a waterblock on it
suboptimal without the other parts π
I fall asleep to the sound of a Strix 1080 and an Evga 1080 running at 70% fan speed lol
I think you know what I mean
the underpower of it made this actually doable long term.
also, every skin of afterburner is trying too hard.
just get a bigger monitor π
Guys if you want Task Manager to show your GPU usage while folding you need to change one of the graphs to show CUDA usage
are all cuda cores the same? Or perhaps same inside a single 20xx line of gpus?
No
ie. my 1070 probably has different chips, so diff performance? Separate from count.
Well, on the same GPU die yes, though different cards might have different amounts of CUDA cores enabled
It'll show cuda but still not a higher utilization overall, at least on mine
have to click on it
how many points per day does a 1070 do? Worth plucking it from its plex case and slotting it into this system? (basically unused in plex...)
Don't do a lot of transcoding? Or does your iGPU do most of it haha
just don't transcode much. 95% of what I have is 1080p and just streamed inside the house
it's a sad 3200g chip, if I ask it to transcode some 4k or something it gets sad
I have one particular heavy user that never remembers to transcode and I'm like, bruh, you are chewing through upload data haha
force em? Isn't that a setting
I just have it for my own use really as a nice interface, not so much sharing it w/ friends/family
Yeah I've tried it's not easy,
Oh yes the interface is lovely, and he is pretty much the only one who uses it outside the LAN so it's not a big deal
But yeah, forcing transcoding is basically not a thing
It's annoying
If it is a thing, it's buried and a PITA to use
@pastel nacelle I kept messing with it and eventually got it to work!
nice
nice!
Got a GPU WU almost immediately lol
not too shabby
I didn't realize the 2080 vs 2080ti was so different w/ compute tasks. I expected that jump going to titan
As you might have noticed work units fluctuate quite a bit in PPD output, so depending on the unit you are folding you might get highly different
@green dove a 2080ti is 92% of a titan in compute tasks presuming 11GB of VRAM is enough
yeah, guess I'm surprised by the drastic jump from 2080 to 2080ti.
Excited to buy whatever the 3080ti is whenever it comes out....
it's almost entirely in memory bandwidth
oh really?
256 bit vs 384bit
interesting
372* sorry
the 2080super bridges the gap a little bit by using faster clocked memory
I've never really dug into how gpu compute works. I should do some machine learning or something and learn
Okay guys wanna help out in defeating coronavrus?
jump on in
was the advertisement from Nvida posted earlier?
Oh yeah, it was.
However I wonder why AMD is silent
Probably because Navi cards suck ass for folding
it matches an RTX 2080
Their CPUS can be used
The 5700XT? From what I've seen people getting they've been getting less than 1M PPD
Though it was only recently that Navi cards didn't work at all so maybe they've made progress?
there's no super in that list buut I imagine they're not terribly far off
That result for the 5700XT is based off 1 user self reporting...
ah
There will be people around who are using a 5700XT who will know what they get
my 3950x is only on track for 243,000 PPD and 3600x for 43,212 PPD so don't get your hopes up on CPU folding making you points
especially a laptop CPU >___<
Metallus is reliable...
He's getting 1-1.3M with his 5700XT with +50% power target
I was wondering what the current state of Navi was on F@h, last time I checked it wasnt working for OpenCL or something around that.
So that's a lot better than what I thought it was... I was seeing other comments around 700-800K mark
Mine's a gaming laptop cpu
@lilac matrix it's still pretty rough... only 4 cores...
the 1660 ti is a 1080p beast
and throttle city
the 1660ti and your 9300h are on different planets when it comes to folding
4 cores 8 strings
guy
I'd just go GPU folding... CPU folding isn't really worth it
the 1660ti will probably do 450,000 units or something if it doesn't throttle in the laptop... the 9300h will do like... 20,000 if it doesn't throttle
Unless you have a bunch of cores
my 3950x is only on track for 243,000 PPD and 3600x for 43,212 PPD so don't get your hopes up on CPU folding making you points
32 cores is still not going to be half of a 1660ti in points
how do I make it use the GPU?
You can right click the CPU slot and set it to Finish, then when it is finished go in to the configuration settings and remove the slot
it's worth noting that it takes a whole CPU core almost to feed the GPU worker with enough data
which is why some people with 4 cores but 2 GPU's installed were seeing 50% usage earlier with no CPU slots running
^ Yep
@frozen oxide so I should buy the 3990X instead of the 3970X? Got it π
you do you - but you'd make about 20x as much points for the same money I think with a 3970x and 2 2080ti's
Yeah I'm joking. I've already got a pair of RVII and a Titan RTX doing folding right now lol
Just a shot in the dark but does anyone have any idea when they are gonna be adding a new cause preference for the virus? I was hoping they would have one but heard that I just have to set it on "Any"
no clue
Apparently it can only be changed by changing the software... They're already working on a new version but its been in the works for a while
I would imagine that with the severity of COVID-19, keeping things running as is takes priority over major changes.
no more CPU Wus?
oof
The servers are being slammed at the moment. Expect WUs to come and go.
yeah, thats a good problem to have
inb4 i install boinc cause i want my space heater
imagine being the team that has to test the WU's to benchmark them for PPD
pretty sure at least 1/3rd of folders dont even care about the points right now
AFAIK they have a dedicated bench rig they use as a reference point and then if you complete it faster than that you get bonus points
So, I've just learned about folding today but I am curious as to the purpose of the points. Are they just a high score or do they equate to something. Not that I actually care, I just want to actually contribute towards tangible results instead of the insane amount of armchair doctors I've listened to over the past few weeks.
Hi Score it is. I am cool with that. π
if someone didn't know what work units were for covid-19
there's a few more
they must have updated it
https://apps.foldingathome.org/project?p=14328 these are CPU projects
the description even links back to the same list of projects which it's not on
they're adding both projects and workloads for existing ones as fast as they can now that they have so much compute
ok
We hope to have preliminary drug screening simulations on CPUs released in the next week. dang those bio/med/biomed students must be working it to get these out
definitely
Hey guys, I'm trying to get a passkey for F@H but I haven't gotten an email yet after going through the passkey creation form
took a few minutes for me, but they've been having all sorts of server issues w/ the influx of new people
Ok
takes me many attempts over several hours to get a work unit sometimes... which is not great
Right now I'm on Project 14328, protease - potential drug target for covid-19. If I'm reading it correct, it would seem they are starting drug screening simulations now which is a good thing.
looking for help setting up toward the common goal
I loaded BOINC, am I doing it wrong?
@lilac chasm you can still contribute without a passkey though anyone could contribute to your username or steal your username potentially in the future - that's the only thing it really does
help
oh wait
advanced viewer says updating
is this the update we were looking for?
annoying the servers are down. My gpu is IDLE!
the jobs are trickling in
my GPU's have been more on than off the last 6 hours, but the prior 6 hours they were doing almost nothing
hurry up and bring more WU's im cold
at 9pm I put my mini-ITX in my bedroom to fold... at 10pm I put another PC in there... things is working like a legit space heater >___>
HA chart?
yea
Is F@H aware of a 2 socket systems. My friend has a older system with dual xeon's and wants to use it to do some folding while heating his apartment.
For a second I thought that was 68C LOL
Oh god that would be really hot for a bedroom.
as far as I know f@H supports multi socket and even NUMA but it might be worth testing running two separate slots one for each
Not sure if this helps, but I just set up a VM on my dual E5-2620v2 server using 16 out of 24 threads on it and F@H is running great on it right now
Were you not able to get it to use all cores/threads without a VM?
Didn't try to as it's running on my freenas machine and didn't want to botch the existing setup
I can test it on my Mac Pro. It has two Intel Xeon X5690 CPUs, but it is running Ubuntu not Windows or macOS...
damn, i just checked my guest room pc and it stopped folding a few hours ago. So I paused it for a minute and then restarted, and it got a GPU job immediately. This software is kinda janky lol.
Hey hey, if I am creating a second rig to fold, would I setup remote client to connect to my main rig IP with the same username and passkey?
Local IP?
Cause itβs not cloud based
still no luck on getting the GPU to run
yeah local
same network same subnet
They just aren't connecting π¦
My main rig is folding fine. Client wont connect via the remote client feature
made sure port and pass was the same
@cloud flower I found I needed to disabled windows firewall to get this working. I don't recommend running this way long term though
I tried everything I could think off with eth remote connection, letting the app though the firewall. Changing various settings, ect. My degree is in networking but I still could never get it to work properly without the firewall disabled. I just gave up in the end
local firewall isn't an issue. I have an edge firewall server to replace windows firewall in any case.. Disabling Win firewall now to test
both rigs win firewall is disabled. still nada.
lemme close and reopen the app
on the remote side
I have found a reboot is needed if you change the remote access settings in both sides for it to even have a chance of working.
In the end I went with teamviewer instead to manage them remotely. There is HFM.net too that you can use
yeah nada. Oh well. TV it is
Sadly it seems the one rig is just getting HTTP_SERVICE_UNAVAILABLE errors
I'm back and with a cup of coffee I figured it out
Think it is worth trying to rig a k5000 into rig to get it running?
eh, IMO, not really. It's from 2012. I mean, every little bit counts I suppose, but for comparison, my 1080 crushes it.
I will forgo the K5000
same, my 1080 is doing great for gaming and folding XD
1536 Cuda cores on the k5000, the 1080s have 2560 cuda cores, for comparison
It's all in addition
yeah
Working from home is a foreign concept for me, I turn wrenches for a living lol
Indeed it is, like I said every bit counts
π
I think I am going to go fire up my r730 with linux to see how she does
My only desktop rig runs winblows. All the ThonkPads are Linux π
Still have a server with two dual xeon machines in it to deploy for folding, still need to drag it out of the other apartment...
So far I've got the gaming desktop going, the Mac going, figured my friend would love me to actually get the Dell going
um, whats the difference between fahcore 21 and fahcore 22?
cause fahcore 22 seems to be using alot more gpu than fahcore 21
and is like 3x faster
common Stanford, give my 16 thread ryzen some WU's!
WARNING:WU01:FS00:Failed to get assignment from '65.254.110.245:8080': No WUs available for this configuration
Seems like they've been running dry on them for the last 8 hours or more
@dusky loom yes. currently there is a major shortage on WU's. Just wait.
@upper shuttle Core 22 is the newest core relieased and is basically designed to max out the new super-massive GPU cores found in turing and the like. Basically it is meant to run hard as all hell and get more work done
It also uses a newer version of openMM the library used for folding and other molecular simulations
How are you seeing GPU useage?
task manager?
Don't use task manager whatever you do for CPU reporting
GPU*
Use GPUz
Task manager does a terrible job at reporting GPU utilisation when you are doing none rendering work loads such as folding
ah nvm, afterburner shows 89%
yeah thats weird for them to do
nice joke
cant upload my wu
ouch
Same here. I've got a work unit on both my computers that won't send.
At this rate, I'm just gonna have a backlog of completed results without being able to send them...
They need to sort out their backend to allow more connections. It's bottlenecking progress
guess I got lucky on the upload
My understanding is it's not so much not being able to connect to the servers, but rather that the servers have run out of WUs to hand out
"Several servers ran out of WU's overnight. The issue is being addressed as fast as the project's owner can get to it. Unfortunately the server(s) may have to be taken off-line while the new WUs are being generated."
That is one of the issues
the F@H team is working on it. if they don't do it carefully the system might get more unstable or even unuseable. so pelase just give it time
so they ran out of WUs, their backends can't handle the connections, etc
It's kind of like a DDOS. but a good DDOS π
we're flooding them with connections, same effect as a bad DDOS but yeah
yes, kind of π
I didn't think there was a good kind of DDOS but apparently we've done it π
PCMR team is up to almost 10,000 active users and just did 110M points at last update
Well, well done to everyone for coming together. It's great π
(for perspective LTT has 3,000 active users and did 69M last update)
so we are better quality but pcmr has the quantitys advantage
lets see what happens at the event start :>
Well we do have a 90 billion point lead over PCMR team
So we are expecting for jobs to start getting assigned again soon?
As soon as they have more WU for us π
Work Units?
yip
Good time to dust your system and whatnot
I built another rig today just for it
if I've got the parts, and the electricity, why not
the drought of work units continues
uhhh 17:15:08:Enabled folding slot 00: READY gpu:0:[ES1000] 17:15:08:Enabled folding slot 01: READY gpu:1:GP106 [P106-090] 17:15:08:ERROR:No compute devices matched GPU #0 AMD:0 [ES1000]. You may need to update your graphics drivers. 17:15:08:ERROR:Exception: Could not get IP address for true: No address associated with hostname 17:15:08:WU00:FS00:Connecting to 65.254.110.245:8080 17:15:08:WU01:FS01:Connecting to 65.254.110.245:8080 17:15:09:WU00:FS00:Assigned to work server 192.0.2.1 17:15:09:WU00:FS00:Requesting new work unit for slot 00: READY gpu:0:[ES1000] from 192.0.2.1 17:15:09:WU00:FS00:Connecting to 192.0.2.1:8080 17:15:09:WU01:FS01:Assigned to work server 140.163.4.241 17:15:09:WU01:FS01:Requesting new work unit for slot 01: READY gpu:1:GP106 [P106-090] from 140.163.4.241 17:15:09:WU01:FS01:Connecting to 140.163.4.241:8080
do i just let it hang here?
assuming y'all also are having problems getting WUs?
yep
glad its not just me. got all these cycles and nothing to put it towards
i wish i could do this in docker but all the containers i could find are CPU only...
mmm u can definitely do it in docker with GPU.
i haven't found a working container yet for it though
Damn the folding servers seem to be STRUGGLING
oh nice!
I set it up to run on my home PC and it's been "Downloading" for the last hour
i cant find anything on mobiusnine/foldingathome online, what options do i add to it?
same
they really are! no WUs. i have to keep restarting my container for it to pick up something
are u using unraid?
im using Arch Linux atm
ah
The 6 Editors machine only did 3m PPD yesterday because it got like 0 Work Units
tbh its in a way a good thing, means that we are getting everything they need done quickly I guess
@unique sigil having trouble downloading for which OS?
Windows
I have their client dw, just not getting any work
indeed
all of those mining machines?
6 Titan Vs + 2 2080 TIs D:
I heard their servers were running out of space or something as well
we have achieved maximum fold!
yeah i thought it was an issue with pfblocker but nope. its them
nice!
why does the bot gotta mute me for logs
how long does it take for your system to rip through a wu?
i got project 14310 on both my gpu's my gtx750 will take more then a day for it at least my 1070ti is almost done with it
so @north basalt i ran the container with nvidia-docker but i got this, left is bare system
It depends on the WU @pastel nacelle
I have my 18 core NAS going, 6 Editors at work, and two 2080 Ti's at home
how can i disable it on my internal GPU aaa
thats sick @unique sigil
sad my cpu doesn't get a task tho but meh
Well, only one 2080 Ti atm, need to throw the second one in my NAS so I can fold on it too π
rip
oh shoot, how do I reset the attemtp timer. its saying that it isn't going to do another attempt for 5 hours
how do i exclude a GPU?
From what I can see, CPU WUs have been pretty constant in availability, but GPU based one (at least for Vega64) have not.
gpu has been available for me but cpu isn't available
@pastel nacelle Pause and restart the slot
RIP GPU jobs assigned to me π
how can i get this working without a gui?
@split cipher on what OS
there we go!
I have to use edge to get the web control to work
i just need to somehow force gpu0 to not run
@split cipher 1) /etc/fahclient/config.xml do you have web-allow and allow sections saying your remote IP range ?
its set to v='true'
- open http://fah_client_ip:7396/ in browser
- alternatively use the fahcontrol tool to see the progress and configure the client
i know my folding at home web control used to work but now it broke so yeah
- <gpu v='true'/> is one thing, but also needs something like <slot id='0' type='GPU'/> for it to work
- it also needs opencl lib to be installed in OS
(for the GPU folding to work)
how can i exclude gpu0 but not gpu1?
@split cipher can you use fahcontrol ?
if yes, then it has options like this
so opencl-index and cuda-index might help with that
no idea how that translates to config file, let me check
can i install fahcontrol without installing anything related to working?
<cuda-index v='1'/>
<opencl-index v='1'/>
</slot>```
looks like that
fahcontrol is a separate dpkg for debian systems, no idea about arch
so you will add a GPU slot and then you set cuda/opencl index to GPU you want to use
im using a mac for gui
see above XML from the f@h xml config file
i think windows and mac has only full installer, but you can then just disable the client, so it won't run
GPU units will be an issue for long time
i dont want this one doing anything at all due to it being actually useless
in any case, as i said, you can just configure your fahclient via XML file
the GPU slot like that will force the client to use the 2nd GPU (if it is zero-based numbering)
@split cipher 1) first you need to enable remote connections. for that open /etc/fahclient/config.xml in editor and add lines like this (change the IP range if needed) :
<web-allow v='127.0.0.1 192.168.0.0-192.168.0.255'/>
<password v='password_of_your_choice'/>```
all that while fahclient is stopped
In 2014, the world avoided a horrific global outbreak of Ebola, thanks to thousands of selfless health workers β plus, frankly, thanks to some very good luck. In hindsight, we know what we should have done better. So, now's the time, Bill Gates suggests, to put all our good id...
I remember watching this a few years ago...
@split cipher 2) start fahclient after those changes
- click on add button at bottom of the left list
- give it a name, set the IP of server and password you set in the config file
and then it is done and you can configure it via the configure button, or control on right side
not getting anything :\
@dusky loom CPU or GPU ? GPU tasks are rare and mostly exhausted
its sitting on connecting
Mix: CPU, CUDA, and OpenCL
@dusky loom CPU you should be getting, GPU is mostly dead as they do not have projects to distribute https://greenfoldingathome.com/2020/03/14/thanks-covid-19-now-there-are-no-more-work-units/
i really want to see the r730 get a job
@split cipher do you have port 36330 enabled on firewall ?
(on either side - arch or laptop)
i dont see why it would get blocked
i was able to access the web interface without fw changes
@split cipher and config.xml you changed ?
(and restarted fahclient after the change)
if you do not have the allow line, you won't be able to connect
10.254.254.254.254 ?
one too many .254
got it
now you might want to push the slider to full, i seen it not getting work if you only had medium
or just wait a bit and watch the log tab
well now i wait
only my nahalem computer has a job
should i enable cpu?
as i said, GPU jobs are rare now, they got too much horsepower in last few days
so if you want to help with random projects, sure, you can do CPU jobs too
i added cpu
but the server infrastructure of F@H is under strain for sure
no passkey email for me, stats servers are kinda dead
229500 is LMG right?
223518 i thought
just waiting...
223518 is LTT
Nice!
rip π
should it look like this while waiting?
π I got an estimated 84133 point one. then when I handed it in, I only got 32k π¦
in any case, i am not going to keep the CPU working on my 2950x,i might keep the GPU tasks running tho. For CPU i will use my server with 3900x
how can you tell how many points you got?
@split cipher ready is when you got no jobs, yes. Look at log what does it say
im using my spaceheater x5650s lol
@pastel nacelle either by monitoring your points earned or in theory via stats ?
but stats is dying all around
ah, the stats page is struggling to load
ok, so it'll show me points when it finishes in the log?
sometimes it doesn't
"No WUs available for this configuration" is the most typical message right now π
Glad to see it wasn't just me not able to load stats. They got all the horse power they need now lol
@stark bough i have one CPU WU from 2 CPUs and 2 GPUs
I have a stupid question
out of 2 CPUS and 2 GPUS I have 2 GPU and 1 CPU running
What is the difference between BOINC and Folding at Home?
@mystic hollow Thanks to the 3900x, I think on thinking of my 3700x as a Ryzen 5
Is it just the program used?
folding is protein specific, and a bit easier to configure. BOINC is a general distributed computing project where you have to select things thats not necessarly just proteins
I knew one of you could answer that. Thanks! Can you run both programs at once?
@stark bough thats sweet. Haven't ever really looked into it
uhhhh
That is what my 2700X aspires to be
Just got one aswell on my cpu:16 VM
spikey
do i need OpenCL for the cuda one to work?
@split cipher what are you running on?
GPU0 is paused
@unique sigil https://imgflip.com/s/meme/Yall-Got-Any-More-Of-That.jpg
You got any of them work units?
there we go
Try restarting your client, if that doesn't get you one then there must not be any left for your config
on 2080 Ti π
WRU GPU units T_T
14% GPU usage, very light load on my 2080 Ti
We allocated 20,000 Wus for Intel and 80,000 for Ryzen, and now we have to configure the remaining 15,000 unused Intel WU's for Ryzen
Guys as of last update we are 3rd best producing folding team in world, out producing EVGA and curecoin by a good margin, keep it up!!!
Tell your friends, your parents, your siblings, hell even tell your dog about Folding for Team LTT! Lets show the world what we can do! Let's get everyone involved!
@frozen oxide yeah, 90% in GPU-Z. OK then
my system is either broken or working but i have no idea since im not getting WUs
I installed the 2080ti 30 minutes ago and it's gotten no GPU jerbs c'mooon
@frozen oxide Freaking Radeons. They took ur jerbs.
Feed me
took er jerbs
please gib job

jerrrrrrbs!!
I keep getting CPU jerbs but not GPU ones... just wanna see what PPD the 2080ti can do lol
nothing of mine is getting jobs
well "keep" is relative I guess... of the three rigs running, only one or the other gets a CPU job occasionally but they're not even long ones
Thank you for bringing this up. Why you might ask? We are sending so many passkeys out that @Yahoo and @Google are treating us as a spam email server. It's also in the FAQ https://t.co/ZF7GJ3h7qZ
lets just break into CPu headquarters and steal all their jerbs
Also anyone waiting on a passkey^
i got my passkey in like 2018 or something lol
my pass key took 12 hours to get to my inbox yesterday
Team LinusTechTips_Team's points are 115,341,250,909.
should i be worried about this? I have cuda
please dont die
how do i fi this????
its a cuda device
@stark bough ETA π
does anyone know how to fix my issue?
i know we knock em out so fast.
@split cipher i'll message you
thanks
@stark bough you misunderstood me. a GT 710 takes exactly one day to do it's work unit π (tho it is still enough, as deadline is 8.2 days)
@mystic hollow thought i was loking at TPF
doh
last folding month i was running things on a netbook. took 4 days to complete a unit
was in the middle of eating lunch, lol
@stark bough vega64, stealin yer jeeerbz
external rad wut?
i guess im using an external gpu so π€·ββοΈ
J E R B S
welp, F@H still not giving me anything. The 1950X is happily crunching on Rosetta though.
the poor FAH servers...I just got a passkey request that I did yesterday morning
oof, at least you got it
i ended up finding my email from years ago before it came lol
@tough trout yup, they are completely overwhelmed
and then in 2 weeks, they will be back to original performance after everyone stops
I guess this is all a net positive, lots of power to get the much needed work done
Regular folders = Theybtook our jervs
@stark bough lolz sorry
Hmm, ETA 1 hour for my GPU job.. hope they got more! π
Iβm cold. I need jerbs
@solid anchor sounds like an easy job, mine was a 3 hour job
3 hours on 2080 Ti for me π
Donβt expect me to run benchmarks like a pleb, Iβm an American and I expect jerbs

anyway, i dropped GT710 after first job, because it was too slow π
might move the GTX1060 or R9 380X to the server, but unsure
yeah, i 'cancelled it' in official way, so the WU is back to pool
@mystic hollow does it matter how fast the jerbs get done at all
@mystic hollow ehhh
@tough trout if you don't do it before Timeout, then the work will get assigned to someone else, so two people will do the same work
Methinks either you have a way harder job, or my vega64 is faster
@tough trout Not really, but as long as it hits the deadline. Itβs all about that score though
ahh nice, is it always 5 days like that?
They send out several copies of the WU.
Usually if multiple people can replicate the simulation independently, they can refine their models and predictions on what drugs they can use to change the folds
@mystic hollow
@tough trout expiration depends on how hard the task is
@solid anchor base credit 38000 for me
ahh i see
@mystic hollow Ah, gotchya
@stark bough On the blower cooler, 102degC
On the waterblock, 34degC
Well, more like 40ish under load
β€οΈ
On my Define C build, I was feeling the exhaust and I was like βwow, it must be super hot in that caseβ
But Mainboard is 26, CPU/GPU is 48, and ai realized that probably 90% of all the heat generated is leaving through that one rad
still wondering why it sais "cpu" project while running on my GPu but... well π
Intake was 21-23
is it worth it to try to fold on an awful ancient ES1000 GPU?
was wondering then what would happen if I tried to cool both CPU/GPU (no VRMs) through a single 120mm rad

Cook bagel bites via convection through the rad
@split cipher es1000 doesn't have open cl so probably not
my CPU finished a job, now back into the eternal wait it goes
@split cipher it's fine CPU jobs are just bonus fluff - even a 16c/32t can't beat even a GTX 1050 for points - gotta get dem GPU jobs
gotta get GPUs π¦
im throwing ewaste at it lol
there are a few production jobs that run on CPU's that produce data that eventually makes GPU jobs if I understood yesterday's update
thats really cool
how does the whole job system work?
imagine there's a cake eating context and our team has to eat the whole cake- they cut the cake into small pieces so our team can finish the cake - each person is working on finishing a cake piece
woot woot!
and once one cake is finished what happens, how is the next "cake" made?
scientists / researchers have to come up with problems/experiments to be solved
then split it up into small workable chunks
which is what happened yesterday basically they ran out of problems to hand out - and a lot of people had cake pieces they hadn't finished but
cake is not the best analogy now that I think about it because we have to hand work back and I don't want to extend this analogy to stuff we'd get out of processing cake
lmfao
π€
okay saving my analogy - they hand out lots of flour&eggs and we have to all go home and bake cupcakes in our own ovens
for a big cupcake sale at the end they'll have 10,000 cupcakes
(oh I just realized why we are talking about cake XD
I will be sad if the 108k points arent accurate
depends - they are dividing it by time so if you get it done in the same amount of time as it thinks it will now then it will be close
just joined the LTT folding team was on the PCMR team all day yesterday
I've been getting what looks like 80,000 to 150,000 points per GPU job so π€·ββοΈ
π
I see, well rn I'm only folding bc I need to get my room clean for when my school begins distance learning XD
so the eta should be accurate
my current job is 48k
@amber wraith Thanks for joining us!!
indeed, welcome!
i wish i could put my desktop on folding, but it doesnt detect my GPU π’
i guess this is what i get for hackintoshing lol
it's a bit weird points system really because faster GPU's not only get more points for finishing the job faster but also can finish more jobs - so a 1080ti vs say, a 1050ti not only gets double the points per job but finishes each job in 2 hrs instead of 4 so it actually makes 4x the points per 4 hours - I think that the time-based bonus should be much lower TBH

