#folding-and-boinc
1 messages ยท Page 60 of 1
the dream is all green ๐
regularly I find it like this
it appears if i get that server, im not going to get a WU
cool. I'll give that advanced thing a try:
<slot id='1' type='GPU'>
<client-type v='advanced'/>
</slot>
see if that'll get me any funny WUs
can't hurt I guess
I have been running advanced on two GPUs for a week
and beta on a third
I've failed one unit out of 100
and that was only on the R9 nano which also is loud as heck so I'm not likely to keep using it
do they have a list of all client-type values?
yeah i noticed
the biggest issue I've had with beta units is that I stop folding in the morning when it's time to work (it's too loud/hot in my office otherwise) and start again after 6pm
and the beta units can take 6-7 hours on a GTX 1070 or R9 nano
so it's been harder to do "finish" and have it finish its WU's
they seem to have a very efficient pause/resume checkpointing mechanism
sure, but the bonus points are based on finish time - start time
ah
so if you finish after a 9 hour break, it's now a 15 hour unit not a 6 hour unit from their metric
the bonus points are to incentivize people finishing units and not pausing forever, getting stuff done, not having work tied up on people's PC's for weeks
right, that makes sense
Hi all, just joined F@H and was wonder if have to do anything special in the web interface to change folding for Covid-19? since there are 4 other research options but not for Covid-19.
set it to "no preference"
there's basically 98% covid units currently
they are priority anyway
ok good, thought i had to do something specific
if you set it to cancer or something it will attempt to do cancer first if there is any work available, then otherwise do covid anyway
well I just got one after setting client type to advanced. small 23.6K atom unit, but 8 million steps lol
yeah the new thing they're doing now with 40x compute power is trying the same protein from more angles
Coming this weekend (or when I finally get around to it, maybe) The ATOM Supercomputer Super Duper Folding Monstrosity.
sheesh
looks intense
honestly doubtful of how well atoms will handle it. you'll need to grab a lot of WUs and they'll take a lot longer to finish than with say, a quad core i7
I'm going to build them into a HPC (Cluster)
also all that will probably be beat by a single 2080ti >_<
since we're often limited by how fast they can hand out WUs, a lot of those atoms might end up idle
Folding event blog now up https://linustechtips.com/main/blogs/entry/1827-day-eleven-the-fall-of-oppression/
Legs Eleven anyone? It looks like the fall of the king may be on the horizon... How will he respond? Is the king to take up arms or abdicate his throne? Todays drink, Comforting Old Fashioned Spoiler Double Southern Comfort 10ml Bar Syrup A few drops of bitters Slice of orange...
does FAH work on clusters?
I believe it will work on Pegasus with some tweaking
I am not sure if you can trick it to do that or not... yo ucould definitely run it one client on each system
can one CPU WU run across multiple nodes? I saw it was using MPIRUN so maybe...
This will be a 58 core, 116 thread ATOM CPU lol
@weary epoch I hope so
It's going to be an experiment at least
it seems like fun
well good luck with that. if you get it working, it'd be nice to know how much message passing it does (like how much network load there is)
29 mini ITX machines with dual core / quad thread 1.6ghz ATOMs lol
that said though... a 3950x 16c/32-thread is beat by a GTX 1050ti, and a 1070 puts out 4x the work... so chances are a single 1080 will beat the atom handily
MPI screams "I need a really fast really low latency interconnect"
well yeah, GPUs will win
I have them, so why not?
if it's for the fun of it, then go for it - it's just not that efficient
435 watts of pure ATOM goodness LOL
do atoms have AVX?
15W TDP each
it does sound like a xeon phi minus the big vector FPUs, fast memory, and fast mesh intercoonect
Yeah, they're gigabit NIC only... It's going to be painfully slow to distribute the load, but meh
what is the overhead of 29 motherboards and 29 PSU's tho ๐ค
which atom is it btw?
so, what power do you guys run, Medium or Full? i'm going to watching some ytube videos or twitch on this computer otherwise it should keep chugging on F@H.
@rich depot full, and it won't bother your youtube use - the folding work is lower priority
but maybe turn off CPU slot, or limit it to to N - 2 threads so you have some threads for your browsing to feel snappy
if i set it to medium it uses 7 cores instead of 8, that should be enough right
I was even gaming sea of thieves the other night while GPU folding on full - lost about 30 FPS so I went from 120 to 90 but it was completely playable
2GB on each board. Some have 4gb
They were "scrap" at work, so I snagged as many as I could put in a box lol
oh geez N2600 uses the saltwell atom core. that's horrible
if pegasus doesn't work.. you could just run a client on each
Never said they were good lol.. They're painfully slow machines for anything other than industrial applications (which is what we use them for at work)
They do all have a legit Windows 8 IOT license though
you might even time out if you ran a WU on each. saltwell is like, a pentium mmx
So I could do them individually lol
hmmm so if you don't have AVX (doesn't look like you do) you're going to go at 1/4 speed again oof
on higher clocks with tweaks of course
Yeah, that's why I wanted to try clustering lol
you could also fold on linux, no windows 8 needed, especially for CPU folding
Just for shits and giggles
you could make a very small image and network boot it
I wanted to see how many PPD I could get on a clustered bunch of junk cpus lol
I'm going to hazard a guess at about... 100,000
I already get over a million PPD on my gaming rig, so I'm not so much worried about the points, just something to play with on the weekend
Well, I would if I could get consistant WUs lol
My biggest holdup at the moment is physical space to lay them out and network them lol
And I'm waiting on a package with 30 barrel connectors to build a buss bar and connect to my PSU lol
so a 3950x is 200,000-240,000 PPD
divide by 4 roughly for having 1.6ghz cores and much worse IPC
multiplly by two for having more cores
I'm registering my bet at 100k ๐
But no AVX, so working at MMX instruction set
oh yeah
might be faster as just 29 individual 4-thread clients idk
Maybe.
If I can't get pegasus to do what I want, I might just set up a headless PXE with autostart of the CLI client
it does need some amount of local storage tho for the work files
couple hundred megs
RAMdisk
each saltwell core can do 2x128b floating point ops per cycle. in theory that's 716.8 GFLOPS, for 56 cores at 1.6 GHz
except the core is only 2-wide, and in-order. so it'll stall on memory access (unless the sibling thread isn't stalled), and FAH is a good mix of int and fp instructions (so you're very limited by the core being 2-wide and in-order)
aaand 1050ti is 2.1 tflops
and 250k ppd indicated
so yeah I think you're probably closer with 20-40k because unfortunately PPD is not linear
doubling your computing power gives you 4x the PPD
Heh, my FX6300 + GTX 750 gets about 85k
if even that - I don't think saltwell will get anywhere close to theoretical and probably be even worse than a GPU in that respect, because it can't hide memory latency as well
My R5 3600 + V64OC gets 1.2-1.4
if I had a pcie slot in which to put my 1050ti I'd be running it, but I don't
GPUs deal with memory latency by keeping tons of threads in flight. big CPU cores can speculate ahead by hundreds of instructions past a cache miss. saltwell atom can't do either
that's probably 1.1 million for the vega64 and 80k for the r5?
Hmmm.. These boards have PCI slots, I wonder what adding a few FX5200s would do (LOL!!)
I do have a stack of old PCI GPUs from my vintage computer collection. I doubt anything older than FX would do anything though.. I wonder if a Geforce MX440 would be able to do anything.
I have some Radeon HD cards in PCI as well
I think you need something from the Fermi generation or after
I gotcha
earlier GPUs have real limitations with GPGPU work
I think they all use opencl now
technically the tesla generation (G80) can support cuda and opencl, just with a ton of limitations. you'd have to use a very old compute model
I just mean if your nvidia GPU doesn't list CUDA cores, it doesn't work with folding
So pretty much 2009 or newer
and if it's older than a 9 series it probably can only do core 21 work
while 10 series and up can do core 21 or 22 work
just noticed we did 10b yesterday lol
I doubt pre-fermi can get any WUs tbh.
which generation is fermi
gtx 4xx
probably a fair bet
and that's 2010 timeframe I think
so a gpu from 2009 probably won't get anything
AMD picked up OpenCL in '09, that's why I mentioned that year
it wouldnt let you add it if there was no core for it
I believe core_22 can run on older NV card than 900 series...
I'm not sure the exact cutoff on core 22
because they have serious limitations. like, the G80 (8800 GTX) can't even do double precision
Someone should drop in a GTX 280ti and see what happens lol
but the core 22 work is pretty beefy anyway
Me either tbh... But I'm sure older cards than 900 can, seeing as 21 can run on 400 and newer
core 22 left beta in december 2019
gtx 980 ti can do core 22 btw
it's reeeeeeally new
first-gen gcn can also get core 22 work
as for the gtx 280, it can't do concurrent kernels. so it'll probably be quite slow too
well eta for a core 22 workload on the 980 ti is like 2-3 hours
not sure why the r9 nano is looking at 7 hours
I'm on client-type advanced
same
even the 2080ti takes 3+ hours on these
maybe I just got a smaller WU
I was just chuckling about having a Geforce FX 5700 with a Radeon HD 5700 and a Radeon RX 5700 in the same picture. Someone has to make that happen.
the r9 is on track for 500,000 PPD - I think a 980ti will be more like 800,000 ?
1.5 days WU on a gtx 650 xd
eh not in reality since I don't get enough WUs to keep it fed
@frozen oxide as for the guess on the R5 3600 solo, here you go
not sure what it can actually do
94.4K
I guessed 80k? not bad
Is there any way I can pause individual work units? I have one running on the GPU and one on the CPU but I'd like to pause the GPU unit
Yes, in advanced control
Lol
Wow, that's really obvious and I feel a bit dumb. Thank you guys ๐
No worries. You can also pause/restart individual slots to restart their timers if they haven't picked up a WU in a few hours
Anyway, off to other channels. Y'all have fun. I'll update on the Atom supercomputer when I get around to it lol
cool, let us know ๐
I found something by accident
so firstly that "top" card is thermal throttling, of course... but I had it running in another machine for the past 2 weeks until just now, and it was estimated 1.35 million PPD on that one... running now at 55% power limit dropped it from 1850mhz to 1600-ish mhz and in this setup estimated 1.1 7million PPD.... slightly over half the power usage gives me only what like a 20% reduction in PPD
so I might just run it at 55% power limit all the time now... not just because I'm forced to by thermals >___>
yup, power limits can give some pretty huge efficiency improvements. even stock, most things are clocked past their best efficiency points
because amd/nvidia really care about edging each other out by a few fps
yea
im really grateful for my garage being cold right now
i just went to check and the room is 10f warmer than ambient outside but its only like 48 degrees outside cause it's raining
especially because that room only has a 580x now but its about to get another and a 980ti
found a deal on an 850w gold seasonic semi-modular unit for $130 plus tax on amazon
@weary epoch 70% power limits seem to be the best balance for my 900, 10, and 20 series cards
you'll probably have to test on each, but sounds reasonable
75% might be a bit past optimal for the 980 ti since it's still well above base (and well above stock boost) at that point. and it's running at 1.1120 v
I've got a kill-a-watt on this thing now... it's an i3 8100 with two 250W GPU's, gold rated PSU, and I have the 2-fan zotac at 74% PL with +0 mhz, and the asus at 80% PL with a +50mhz OC
493-500w indicated at the plug
nice
4.3mil PPD indicated... it takes a while to update tho so in an hour I'll drop the power limits again and see how much it hurts it
watching the new saul rn
Wait... it's tuesday?
I just watch it on netflix
@frozen oxide Somehow both of these cards are running 58-62C while folding
it's possible that they're getting lighter WUs
Top: ASUS GTX 660 Ti
Bottom :EVGA GTX 970
i know some are less intense than others and only use 60-70% utilization
@hushed bobcat the total power draw of those two cards is about the same as one of mine
Afterburner says 99% on both GPU's
dang lol thats good then
my rx580 is getting 71-76c sitting in a cold room in the garage
thinking i might open the window once i cram another rx580 and a 980ti in there cause it'll get warm
They are both 150W TDP
don't look at task manager for folding utilization tho
use afterburner or hwinfo or literally anything else
yeah... I'm pulling 500 at the wall so probably 430-450 for the cards when efficiency is subtracted? I tested the CPU alone on igpu and it was around 30 watts
which 2 cards are those?
if you have a low-ish room temperature and an open case, then yeah temps should be pretty good
AMD needs to unify their driver for dif cards. on my RX5700xt i have a full power overclocking/temp monitor with power draw metering but on the rx580 it has none of that and the software is totally different lol
I want the power draw stats the software for my 5700xt has xD
Looks like F@H stats server is fixed and I think I've had constant WUs for the last 2 days ๐ฎ
its not 100% but its getting there
Feelsgoodman
they still need to add more capacity, people havent brought the full farm online yet
give it another month xD
i have parts coming
My folding@home web control page has broken. I started folding on both my PC and laptop using the same account. They were working fine, my PC had a job, and my laptop was waiting for one, then when I cane back the web control page was constantly reloading on both, but according to task manager it was still running
im thinking it might be an issue with the driver revision
i thought i used the most recent ones but im not sure
there's an update from yesterday though so im gonna install it
the software is so different i can't even check to see if they're the same or not
it's the installer i thought i downloaded for my 5700xt
yea
im using afterburner on that rig because i overclocked the gpu from 1365 to 1430
would just be nice not to have to run separate stuff
the AMD one has CPU stats and then i could replace intel XTU on my screen with it
yea thats not what i was having issues with in itself lol
i know afterburner has it
im just trying to have less stuff open in total monitoring because im running out of screen real estate
and also afterburner isn't resizeable but the other 2 are
i wish but im not code handy
learning disabilities in pattern recognition and languages
yea, the driver i used was out of date
new one has the new UI ๐
at least im p sure
waiting holding my breath as it loads, it looks like the same loader
already rebooted it's just finishing load, says 5 minutes but the timer is going up as the bar slides further to done xD
should be fine timer is going down now
lol
i think it was angry because f@H restarted when i had paused it before reboot
stopped that
yea its flying now lol
this new installer is super extra, it even has sound
and an animated graphic
yera looks kinda like an atom with electrons rotating but all red
lol
good
if not then IGP standard is a great thing
thankful that they made that a normal feature on every board
i had to use it to troubleshot my rx580 rig
after i rebuilt it it wouldn't post despite being the same bios/config ect
i literally removed the board with ram and CPU mounted and just reinstalled the GPU in slot, thats it
only difference was no 950 pro 512
using an old 840 evo 2.5"
120gb lol i got it for $70 when they were the hot model as a boot drive for rigs like this
cool i can see power consumption on both now
yup
well, the first thing you should do with any out of box product is update firmware tbh
you never know how long it's been sittin on the shelf
yea
depends though
old A+ certification recommends against it for any part that doesn't need it
when i got my 2nd gen ryzen and put it in an updated board and it just went "NO"
there's risk to break stuff but if it functions and isn't otherwise broken or has a security bug then it's fine
I ended up returning the chip based on amd support recommendations, turns out my b350 bios wasn't up to date
but yea, in these situatios, 100% update lol
yea, sometimes you just have to RMA to update or find a test chip
i had a friend do that once
no money for a CPU that would work with that bios so they sent it to intel (when they still served a ton of mb direct to us) and they fixed it
hay it takes 24 hours for a new user to be showing stats right?
yea it took a couple days for me with the servers where they are
it took 12 hours just to get my authentication ID
for a i5 and a Gtx 1070 whats a good points per day?
ya im folding now on med. why i play on stadia lol
not a lag or slowdown
#cloudgaming
ya they all are about the same, i have GFN too, i was a heavy user of Parsec Glad Ltt did a video on it
ya will be nice to get a EXE or Web play for Xcloud but i dont see that happening till after new Box comes out
I use anydesk
nah, it's all vcpus
yep, esxi vmware
I've done over 600 wu's in a week
me too, running the beta client
most of my work is being done at my dc's using free power
rip my laptop to folding
Why is the PPD so low on some of those clients @storm talon ?
idk that is a mystery
i saw like 400-600 on some
time of death 5:15am
maybe lack of ram?
idk I rarely check it it's just running, and not really caring about ppd since i missed the cutoff for the contest
idk, I'm up for any suggestions on how to tune my cluster. it's very fun watching my r810's get pinned to the floor on cpu usage
Honestly I'd like to help but I probably need more info like how exactly the VM's are set up (OS RAM CPU) and the ESXi CPU Monitoring Page
Might even be a storage limitation @storm talon
Rip my asus rog 2015-2020 died fighting coronavirus. Canโt even post right now after I smelled smoke.
F
F
F
do mote active cooling
i cant find any info on who they are really
is it just amazon flexing on everyone?
yea bezos is flexing
Amazon have also set up Work servers
the bandwidth needed for folding is next to nothing compared to assign and recieve
yea im wondering why AWS hasnt taken that up yet
ooh nvm they have
well then im surprised that it's still lagging
moar fiber
MOAR
inb4 all their crunching is all done on their LAN
fiber where i want fiber
meh
wait am i missing something I dont see amazon on the F@H server list
https://apps.foldingathome.org/serverstats
they don't have servers
i could've sworn i saw an AWS server a few days ago
yea
if they actually had them i bet they would not be having bandwidth issues
that's even more dumb then if bezos is flexing without being able to get work either
ait i got 3 hours to balls to the wall witcher 3 before going back to folding. Should I just continue with quests at level or should I flex and turn into super quen warrior 2 electric boogaloo and push through DLC to get the runewright quest done 25 levels early xD
I have no idea what any of those words mean
If you've played witcher 3 you'd know it's fine lol
I have 10,000 hours in Rust if you can translate
quen is a spell shield it absorbs damage, theres a potion that can cast it automatically if i take more than 1/3 health in a single hit, so i can avoid fast hits and regen through every other, plus casting it myself as able
theres a point where high level stuff becomes easier instead of harder when running it and its kind of broken and lulz
@brittle coral Plot twist:
Bezos asked for all their WU, and instead of putting it on the WAN, only feeds them his own computers to flex.
linus tried to talk about doing the same thing in their own video and decided they probably wouldn't let him xD
yeah but unlike LTT, he cant do all the processing in the same building
ayyyyy but also unlike LTT he has enough networking gear to make the server think it's on the internet when it's actually in their AWS WAN feeding it internally
at least, that's my extrapolation of your consipracy
if Amazon wasnt crushing it right now
non zero chance theres an active thread on reddit right now assuming LTT is being given priority for units
yea but if you look at the metrics they add up
PCMR only has 1/4 of their total member base folding rn
which is why they havent quite broken 1bppd
blame the games
yea i do lol
im still sad about my BOINC accounts tho rn
there's some kind of issue with them
the milkyway@home founder has the same issue too LOL
im #6 for team anandtech there all time on milkywy@home and i can't access my account to bump it up if i wanted ๐ฆ
i think the majority of them have trouble fine tuning the F@H settings
and like in the top 20 still all time
yea
ive been trying to see whats up with that and help them just for the benefit of everyone but their discord doesn't have a dedicated channel even
loke how of of them open up advanced clienct and wonder why 'its not connecting'
they have 10x the active members in discord as well
yeah i was wondering why they dont have one either.
v frustrating when you're trying to be altruistic
most of their issues is really simple too.
yea
i would sit in there all day like i have been in here and relay what ive learned
distributed knowledge is just as powerful as DC
40% of issues is
-Why is my 'web client' constantly refreshing
-Why isnt anything running
-Why is it after i restart my computer nothing is running (its because you started the advanced client and not the FAH client proper'
or because they didnt set it to run at startup when they installed it and don't know to run: shell:startup
and add a shortcut to that foldeer
ask me how i learned to fix this xD
15% is 'how do i set up remote monitoring'
There is no cloud monitoring. only in network monitoring. there is no official app. unless you know how to configure a home VPN it wont work
i think the default install is 'auto start at startup'
i have a monitor with 3 inputs out there that i can use to set up 3 computers with VNC
but most choose 'advanced+manual start' but dont know where the manual start option is
was really simple
yea lol thats what i did but i did know hwere it was, i just decided to change it after i saw it didnt murder my rig at startup like milkyway@home did
milkyway@home would start so fast and hit my system instantly and it would slow down, on a Q9650 at 4.7GHz with 16gb of DDR2 1066 xD
i just pinned it to my start menu.
F to that powerful warrior you were RMAed before your own death in battle for a 2600k cause frys was lit then
i worked at the palo alto store for 2 years
its closed now
lease expired in dec and they didnt renew ๐ฆ
COVID-19 is keeping us from going to its funeral.
yeah i saw the article.
it was the one in the strip mall, right?
i wanted to do a road trip one day and visit all the themed ones.
about bezos, azure's flexing on the serverstats page. look at all those eastus.cloudapp.azure.com servers
like the aztec one.
less than 3 minutes drive from the original HP campus and the garage
the themes made some people uncomfortable but i still loved store 3 growing up
i was a big gold rush nerd
at like 9yo
when we went to sutter's fort and then to the american river to pan i actually found a 1g nugget hidden inside a rock i cracked open
smart me recognized the rock type
MY theory: Fry's is running off consignment and burned all their contracts.
now they're forced to sell perfumes and junk cases.
They are literally holding off liquidation (the last 3 years has been slow motion liquidation) cause they were hoping property values would slowly rise to eventually sell off their stores in prime locations for high density housing for millenials
https://ktgy.com/wp-content/uploads/2017/07/The-Alexan-Webster.jpg
what i was told by inside sources is that was the case
but they were trying to negotiate new vendors but it fell through
just kept having issues
seems like. i was fed that like a lot prior to BF 2019
they wanted to go to a model of pay as inventory is sold
but the vendors wanted cash up front for inventory
isnt the prior basically consignment?
i have a friend who stille worked at the store in palo alto till it closed
yea
they wanted to move to consignment but couldn't find a supplier
at least one who had confidence in them
they're condensing
sunnyvale is going to ride out the storm
as long as possible at least
the bitwit video made a lot of waves
the hope is the company will still exist in some form but they may move to online, buy my god they have to modernize badly to do that
ive been superchatting him to do it for a while.
they'll need outside investors IMO
microcenter shows the model works
yea
but dammit if it isnt a commodity by now
and microcenter closed their bay area location ๐ฆ
central computers is our fall back
my friend also works there
upgrade cycle expanded from 18 months to 4-5 years.
they are doing similar stuff to MC but also the company is chinese so different culture for sure
thats what killed Frys.
you see the old videos, people basically got lost in frys.
in a good way
it was basically an 8 ring circus
i'd go there just to have a sandwich and socialize
when i was like 18 ๐ฆ
why i applied
"billy will be in electronics, dad will be in tvs, jenny will check out the CD,s and mom will go to the theather"
^^
displays to entertain people helped a ton
every time i set one up it was $$$ for me
because the other associates didn't know how to milk it like i did
the san Marcos location even had marks on the ground for lines
basically queues to watch a movie inside.
nice
for "mining carts"
yea the theater room at our store was just ah ome theater sales room
when ti was still a separate room, i think it got turned into a sales floor at some point ๐ฆ
by the time i worked there in 2009
the palo alto location looks prime for another Costco.
business center?
that store never has half the stuff people want
then convince microcenter to put a smaller location in there
its about the same size as the nearest central computers including the service center floor
years past there would be a line to try this out
2017 was not to long, and yet so very long ago
its that a racing sim in a supermarket?
yea i set up the one at the palo alto store with my friend so we could co-sell the whole inventory of GPUs through it
i'd make the $$ on the new custom system, we'd split the GPUs 50/50 and he sold monitors
and the chair table ect
that's in a fry's
yeah frys feels like its going through something similar to toys r us. lack of inventory and traffic.
it had a prebuilt rig and a part-picked rig of much higher power custom built via our service center
so people could pick and choose
same price lul
holy shit its taped off now
this will probably be the last picture every of this angle.
im just glad people documented this prior to COVID-19
It went from
-The best place to buy electronics
-the crowded place to buy electronics
-the expensive place to buy electronics
-the place with no electronics made in the last 3 years
-the place that sells made for TV perfumes.
btw i leveled that +20lvl quest in witcher 2 with that build i was lulzing about earlier. absolutely flattened it o_O
at least i did so far, i went into a room full of enemies and AOEd the lot to death in about a minute
Hey all am I in the right channel? I'm having an issue with the folding at home thing.
pausing a slot and resuming it seems to reset the TPF timer, really screws up the estimate PPD lol
yea it does lead
Oh good. I'm having trouble getting my gpu to start. CPU is fine but gpu seems to be stuck at download?
the servers have been very overwhelmed
work units may be intermittent till they bring more bandwidth online to issue them
ah okay so if I just leave it it'll resolve itself?
but you can potentially get beta work if you're okay with monitoring it some
may lower your PPD a bit but its better than sitting idle for days
monitoring it?
yes occasionally the work units in beta are unstable
oof
the main issue i've noticed is poor optimizatino tho
lower PD
but way more PPD than zero ๐
way way more
i'm largely away from the desktop so i may have to leave it idle
ah well. thanks for the help though!
as long as the rig itself isn't overclocked it should be pretty solid
I've not had any problems getting WUs for a while now, advanced flag seems fine
i can explain how to set it that way or you can set the client-type advanced parameter in your config in slots
if you wanna do it yourself without help lol
that's what it is, name: client-type flag:advanced
advanced != beta but yea not going to talk about that one ๐
yea not worth explaining
should be renamed alpha tbh lol
hmm so 4930k @ 4.1GHz -nt6 is getting 34k PPD, swear that's slower than it used to be a year ago
yea thats lower than my 8350k 4c4t 4ghz
at 4.3 its getting 33614 on the current WU but i've seen it as high as 44000
G Skill TridentX F3-2400C10-4GTX running in XMP profile (unless it's not)
Quad channel is active
@ 666 MHz 9-9-9-24 (CL-RCD-RP-RAS) / 33-174-4-11-5-5-20 (RC-RFC-RRD-WR-WTR-RTP-FAW)
WU00:FS01:Connecting to 18.218.241.186:80
WARNING:WU00:FS01:Failed to get assignment from '18.218.241.186:80': No WUs available for this configuration
ERROR:WU00:FS01:Exception: Could not get an assignment
@delicate pivot
If you look at the logs, when you see something like this, it means they dont have any work for you at the moment.
actually these are the running timings: @ 1200 MHz 10-12-12-31 (CL-RCD-RP-RAS) / 43-313-3-7-16-10-10-26-7 (RC-RFC-CR-RRD-WR-WTR-RTP-FAW-WCL)
1200 is 2400 (double data rate)
holy balls these myosin WUs give a ton of PPD
Myosin? Which number?
Yup it's that problem. thanks @stark bough !
my rx580 is doing one and its getting 557360ppd but my rx5700 was getting 630000ish on the WU it just got before I paused it for a covid protein
er 5700xt
they switched to assign?
ah
i wonder if each server handles assign and accept individually, or is it all centrally managed.
at least there is one that is
my 8350k was crunching one an hour when it got 3 in a row lol
managed by the AS servers
cpu one i ahve rn is 13829
14409 is a faster one i think but my CPU is dropping from being paused
trying -nt 8 now, see if the PPD goes up a lot
whats that do
that's the actual parameter for threads in the slots setting
mdrun -s frame2.tpr -o frame2.trr -x frame2.xtc -cpi state.cpt -cpt 15 -nt 8
Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
109 is fine but 111 is not
๐
Core_a7.fah/FahCore_a7.exe -dir 01 -suffix 01 -version 705 -lifeline 27348 -checkpoint 15 -np 8
nah just gui, this is my gaming pc so can be lazy
i considered turning down my power plan, but then i realized how cold it would get
๐
was about to take a nap, and its raining outside
hahaha
I power cap my CPU
if its planetside 2 maybe i let it run and let them figure it out on their own
291 W of heat
witcher 3 i pause
capped to 1000W ๐คฃ
This UPS I got can only do 810W anyway
that sounds right
need to wait till next complete step to know
ppl with 4930k were getting that earlier at tilt
Whatโs PSU on a dual card setup for say a 2070s? 800W ok?
yes
800w could run 2 probably
Yeah I figure I got a lot of Wattage overhead if I add in a different card
How about 750W?
probably
2070 doesn't use a lot
whats the draw on one card
if its like my 5700xt im sipping on power at 180w peak rn folding
i have room for another on my 750w for sure
even with 300W per card (which it's not) and 150W for CPU that's 750W
max with Monitor on in my UPS? 340/810W
thats 10w less than gaming
think thatโs 6 core on my 3700x, and my 1070 running full tilt with a mild OC
OC 4930k + 2x 290X folding is around 800W
oooooh
you ahve the answers i seek
how much are you getting out of that 290x
ppd
just out of one
14543 @ 380K each
Kinda glad I spent the extra $10 and got a 750 instead of a 650 then
Probably wouldnโt have been worth it to spend an extra $40 for the plat
thats not awful, power draw per card?
Anyone here ordered one of the F@H LTT shirts?
PPD is now 43k btw
I considered it
i have the spare PSU for it if i get a splitter for the rig i put it into
seems low still
it is
@fierce steppe You know, I might get the hoodie and the shirt
16420
Been too lazy to measure myself though
for how much power it uses it is anyway
its about the same as an RX580 at probably double the power
at least 50% more
he looks like a medium but if he's really tall prob a large
Heโs 5โ6โ
hah
well AIDA64 tells me the power of the GPU but I 100% do not believe it lol
medium master race
it has size chart on the website
it seems accurate on my systems
i just updated both drivers to yesterday's release though
the values are read from within the card
they'll measure voltage @ amps
that data doesn't lie lol
@fierce steppe yeah but body type matters too
that sounds right
rest is PC and monitor
i assumed it would be around the same power draw per card as a 5700xt
volts math looks okay
you're on 50hz 240v input?
are you not?
240V 50Hz master race
@fierce steppe Thats the REAL Play
TMU
ect
just SP
none of the fixed function parts are really active
so they're all power gated
no the power states on my UPS does not match what AIDA64 is saying
so I do not believe AIDA
I hate that the favicon for the LTTstore.com is the same as the forums, makes it so hard to find the store page in my 6000 tabs
sry i missed that i drank some of the 190proof ethanol i bought for extractions and cleaning since ISO was out of stock everywhere xD
lol
It's nice being able to sanity check that sort of stuff though
PPD still 43K
wonder if it's just this project
yea i used to leave my system running off a kill-a-watt but i melted the kill-a-watt because it wasnt rated for 20a draw xD
and i had a 1050w PSU running off it and another rig
i have an AXi 1200w i have to RMA
cool but yea not accurate
it died 6m into its life
ouch
still got 3 years or so on warranty
mines old
i have been very very lazy
EVGA NEX1500
Hey guys! Thanks for folding! You are all doing a great job, keep up the good work!
ayyy
I've made 3 planes and 4 cranes just today ๐
๐คฃ
can i have a 2080ti to keep up with people
self high-five
It was a good one
one of my friends is giving me a 980ti to fold on cause he doesn't have room xD
the rig in the garage only has a 600w PSU as primary, i have a 750w enermax china branded one i pulled out of an ancient bitcoin miner with 2 8 pins on it plus some cpu 8 pins i can convert to GPU with the right molex ends, and i have an 850w seasonic unit i just ordered otw
whatever rig ends up with the most cards is getting the paired PSUs
@brittle coral Thats why I like the UPS. I was hoping iCUE on my 650i would indicate real time power draw like the Asus, but it only reports on rail mode, temps, fan speed, and efficiency
next one is getting the 980 and whatever else i put with it
yeaaa
i just dont have a UPS
no good UPS pushes what i need
and i don't have room in my rack for one there either
all 12u are used up
Just get one with failed batteries and leave it in bypass mode
i need to get rid of that desk its under, extend my GALANT with another leaf, and put a 16u in
Iโm going to grab another one of those Costco 1500 VA ones
still get the power monitoring
space is the issue
True sine wave isnโt needed for modern PSUโs, right?
you can actually get really tiny ones
rn my audio rack takes up 12u
i need one that can do 20a/2400w tho
xD
or a bank of se veral
but if it ever goes in to inverter mode it'll shut off lol
PCs don't mind simulated sine
Anything with a transformer block iirc
Are double conversions sub 70% efficient?
Always been curious
no
to bad i cant just build one
92% to 96%
i have a pile of torroidal transformers 6" wide from amps that are bad
crown k1/k2
So about the same as line interactive?
very close
each one about 20lb
Iโve been thinking of getting a double conversion for my Hard drives/NAS
Figure line interactive is best for gaming anyway
Eaton 9130 UPS โ Features
Cuts energy usage and costs with up to 95% efficiency in online double conversion mode and 98% in high efficiency mode.
Those power conditioners that were real popular in 2006, those were also Line interactive, right?
those don't do anything heh
Fancy surge protectors made of gold
pretty much
any power event where you'd need protection those don't really work
so it's like, why bother
what does work is platting the conductors of the power cable like Cat ethernet
that does reduce noise
Costco version and others I tried seemed to just be 10/100 for sub $140 when on a UPS
the part of the industry i work in that name is a running gag sry
i rent klipsch and turbosound gear
rented* now i guess
My old man insists theyโre still good
Bose as like 1 product I like, for one specific task
Despite being 17 year old or something at that point
yea
that is about right
they have a couple things that are worth the price but the rest is meh
Bose 301 for surround speakers
im not a fan tbh
heard them, wasnt impressed lol
sensitivity is just to low
signal sounds distorted no matter how you set em up
they are only good for surrounds lol
yea basically
just because he spent a lot of money on it doesnโt mean itโs good. The receiver unit is acting up, and unlike say bookshelf speakers I canโt just wire in the sub to something else to make up for a failed component
and even then there's better surrounds for the price
and I mean that in the they are the best I've used for that task way
Bose systems are tuned to their specific setup, nothing else
hahaha i mean
you can
i did that to a bunch of m-audio subs
checked the specso n the amp, found a 2ch one comparable, fixed 2 subs in 1 go under my desk
I like the 8 in wide range on them and the reflecting tweeter
I got a companion 5 from the base. Theyโre awesome for my computer setup
But I cannot see them as good for anything else
front in my theater are Martin Logan ๐
free is free
i have enough klipsch pro speakers for a 12ch klipsch system with 6000w of sub@ 105db sensitivity in array, 8x18 4x15
The control puck feels premium AF
i just dont have anywhere to run them like that right now xD
So Iโm definetly not complaining there
12ch+bass
lol
i figured i'd LCR my KP-301s (klipsch chorus) and put the 4th as a direct rear and then set up 6 channels of KP-250 (heresy) surrounds and maybe move stuff around, set up 4 atmos and 4 surround ect. but just a dream rn
i also have a pair of KLF20s just sitting lmao
maybe use klf20 x2 and 301 in rear
dream dream dream blah blah blah xD
@silver wing From your remarks about your 4930k earlier
My 4790k @ 4.4 GHz with 6 threads is pulling 36k PPD
WU 16400, Base Credit 437
hmm so the same really, output has change since last year
i took a reading a few days ago and it was 43k PPD
thats about right
didn't write down the WU or Credit though
the 2 extra threads doesnt make a big difference
my 8350k 4c4t at 4.3ghz is about the same
ah yea
i keep spacing on that
4790k is a 4c8t tho
i had one lol
F soldier
you died in battle before your time
you're running 8 threads right @silver wing ?
just out of warranty
8 before, just upped ot 9
My pfp okay lead?
meanwhile ryzen is over here with 8c16t like PPD GO BURRRRRRRRR
@jade spade yea, better than the other one lol
I have a Ryzen R5 3600 in a ESXi VM with 6 threads pulling 75k PPD
WU 14613, Base Credit 892
My 3800 pulls 120kppd easily when i let it run all day
if i game for 3-4 hours it still gets 80k or so
im seriously considering selling it though and getting a 4950x when they are released
well when i was messing with my 8700k the log said something about "In order to ensure WU stability reducing thread count to (insert number of threads it was set at before the WU started)
I can't wait for used EPYCs to hit ebay on mass
mine mine mine mine mine mine mine mine mine mine mine mine mine mine mine mine
yea i had trouble getting wUs at start that i attributed to maybe my thread count but then i set it to 12 and it was fine, so i upped it to 15 and here i am
is there a way to set up multiple pc's on the same account?
would I just set the same username and password?
aww i want an EPYC server soo bad, my experience with the Ryzen "Server" platform has been less than Ideal
@jade spade yes just set up each clients identity to be the same
dont forget your passphrase
yep, it's doing it
the thing that has too many stuff to remember
oh yeah I had to reset an admin password to get F@H on my pc hides
weren't you the guy who's mom thought the F@H client made the PC vulnerable to the NET?
nooowww I need a workload..
yeah
she still doesn't like it
hehe
even after I explained it
I need a WU waaaa
I've been in similar situations, try explaining to a friends parents that you need to log into their router to forward some ports.
jeez
ugh this guy on craigslist sold the RX580 i was gonna buy without telling me till after i said i'd buy it and that i was buying new PSU to go with Y_Y
and gooodbyyeee qustoodiioo!!
@jade spade you can get an authentication key and then you can link your PPD to a single account yea
It is linkkeeddd eesh
it's done ^
well, done logging in
D:
still no WU!
I will take even the smalles of units!
i cant remember your system specs, but if they are pretty beast you can try setting advanced mode in the client
my sytem specs is 8GB, 8 cores, 3GB GTX 1060
short version ^
I set the checking frequency higher
waet
nu
hmm
eh?
well i think you want a 1070 to set the advanced client
am confuse
atleast that's what's been recommended here
@jade spade Setting the Advanced Client Flag in the Expert settings opens your PC up to more WU's, However the WU's are usually more intensive so you want a fairly powerful system to crunch them
ah, well my computer sits here all day wasting power anyways, so it will have time, dunno how to do it either way
am looking
can't figure it out
Advanced Client > Configure > Expert > Extra Client Options (Add)
yeah nooo...
that's scary
(I know, big baby, but I don't like to mess with stuff that's that complicated)
Reboot client
the advanced flag is fine if you have any kind of rig targeted at 2k gaming from within the last 5 years tbh
It's a 4+ year old system, targeted to be a "Gaming PC!!" from ibuypower
was a prebuilt
You can always change it back later
If your GPU constantly gets WU's that take a day or more then you dont have enough oomph
yea the worst iv'e seen is 4-5 hours on my 580x
@jade spade look at the pic from a few min ago i posted
oke, jsut set it the same as that?
yea
when I close it, it vanishes completely from all of my computer
advanced control
no...
its not in system tray?
if you closed the advanced or web control its still running
Start Menu
unless you closed the actual work process
no.. @brittle coral... you're not listening
click = Folding @ Home
you click "Quit" it closes F@H completely
that should open the client and the web interface
and I can't reopen it, it's like it never existed
I once clicked quit and had to dowload it all again
i'm gonna be honest that sounds like an installation problem lol
I have twitchy fingers ๐
i dont even know of a "quit" button
the only one i see is the exit button that doesnt stop the WU
and the x in the top right which does the same thing basically
ooh down there, yea i dont touch that xD
it is literally the first option
i hadn't looked past the advanced control
yeah,.. really don't
I use that all the time to restart my client
WHAT
it should restart when you reboot
yea
if you set it to start on boot
even if you dont know where you installed it you can run it again that way
but for the future
remember what drive it's on and go to the folder it's in
you can make a shortcut to the core executable
waidaminit
ooooor, if you set it up to start on boot, you can find tthe shortcut to it in the start folder
huh...
I recommend to set the client-type per slot, not for the entire program
that's how i set it up when I did it
but how do I do that?
am not know how
go to slots
configure -> slots -> choose slot -> edit -> extra slot options (expert only)-> add
10 threads 52k PPD
that'll change in about 3 minutes
yep ๐
24 hours is the default "I have no idea so 24 hours"
14 minutes a fold (on my laptop) is pretty fast
I'm also on day two of a WU
nearing the end though
seven hours
oof no QRB for you then
huh?
quick return bonus
no, it's about 3x the minimum points
hmm, would of thought nearly 48 hours wouldn't get any
If the job takes more than 24 hours it should reward Base Credit
dont know if it takes pausing into account
would of thought yes due to it's name, but no idea
