Join the MicroCenterOfficial Folding@Home Team: #257944

MightyMayfield

cine_chris said:

@MightyMayfield it`s good to see you back to your normal self.

Still at about 70% capacity but the machines are starting to come back online. Still a bit shaky 

JADO

Cine Chris  Thanks for the help.  I added my onboard GPU and a couple cpu's and changes to Parkinsons.  Will let those run for a day or so to see if that helps.

cine_chris

JADO said:

Cine Chris  Thanks for the help.  I added my onboard GPU and a couple cpu's and changes to Parkinsons.  Will let those run for a day or so to see if that helps.

It can be frustrating.  Hope it improves for you.
What`s crazy is the GTX cards love these WUs, my 1660Ti is currently sitting at 1M PPD & the 1050Ti run at 150% of normal.   I don`t have the 1050s running now.  I`ve got to be very aggressive to keep the RTX cards fully utilized.  I delete the slot and reload it.  I`ve seen the 2080Ti go from 1.4M to 4.2M, but even being retired, it kills way too much time.

JADO

I agree thanks for the help

cine_chris

Project 171## is a series of short WUs to benchmark Core x22 on various GPUs.
It might appear that your machine is behaving strangely. 
To me, it looked like the GPU was trying to load CPU WUs.
So, check the proj#  prior to hair extraction.

JADO

Chris, are you able to see what my machine is doing?  I have no issue with that but I would like to understand better.  If my GPU's are trying to load CPU WU's how do I stop it.  I have been getting a LOT of Cor_22 FAH errors the last couple weeks. I suspect it is related to the last update.  In the past I would get a couple a day now I am getting 7-10.  Also, I have no idea what a hair extraction is.  Thanks 

cine_chris

JADO said:

Chris, are you able to see what my machine is doing?  I have no issue with that but I would like to understand better.  If my GPU's are trying to load CPU WU's how do I stop it.  I have been getting a LOT of Cor_22 FAH errors the last couple weeks. I suspect it is related to the last update.  In the past I would get a couple a day now I am getting 7-10.  Also, I have no idea what a hair extraction is.  Thanks 

LOL hair-extraction == pulling your hair out in frustration
@JADO I'd post your question on the FAH forums.  Probably in the AMD driver section.
https://foldingforum.org/index.php?sid=576c0ee8bc7a3d38d36e78ca637b2172
I don't have any AMD GPUs and your problems might be something they've seen.  Right now, I'm all Windoz & Nvidia.  
Searching their site for your GPU model might be a good start.  HTH

cine_chris

@JADO are you folding on the CPU?
If so, try removing the CPU slot from Folding.  I just spotted a problem on my X299 system with CPU folding enabled.  I don`t normally fold on CPUs.  One of the problem symptoms was connecting to the machine remotely while folding on the CPU.  The client was slow to update and timing out when connected to my central client.  As soon as I dropped the CPU I saw a spike in GPU utilization & the comm issue was gone.  Thinking this might be a bug in a recent update.
 

cine_chris

Interesting anomaly.
Anonymous spike, note their chart is 1,000s.
That`s a daily 5B PPD spike over their already huge 10B PPD.

ryau

That's 2,000 2080TIs, assuming they each do 2.5M PPD. If we only knew who they were and could get them to join our team!

MightyMayfield

ryau said:

That's 2,000 2080TIs, assuming they each do 2.5M PPD. If we only knew who they were and could get them to join our team!

Haven't see the hourly graph, but I wonder if they are backported points from WU uploaded to the nonstandard collection servers. Pretty sure we got at least a few hundred thousand points last night from older WU that way.

cine_chris

MightyMayfield said:

ryau said:

That's 2,000 2080TIs, assuming they each do 2.5M PPD. If we only knew who they were and could get them to join our team!

Haven't see the hourly graph, but I wonder if they are backported points from WU uploaded to the nonstandard collection servers. Pretty sure we got at least a few hundred thousand points last night from older WU that way.

Looking at todays data, both Anonymous & Nvidia's SaturnV were way over their avg production.  Guess they must have some secret pipelines of WUs out there.  I've turned 200M pts in Aug and didn't see 1 waiting for load.  

teamtempest

Or it might be many long-term WUs finishing at roughly the same time. The one on my dinky little GPU still has about a day to go on one that's taking four days total. Once it finishes my personal graph on overclockers will show a big jump for that day. Probably below average for several days now with no GPU contribution recorded.

ryau

Mine suddenly jumped up, I guess there was a backlog of processing WUs. I usually do 40-50 WUs per day, but yesterday it shows 142.

cine_chris

It appears to be back to a near 'normal' 6-7min update cycle.  I noticed it yesterday, think worst case for me was a 3-5 WU back-up as it seemed to jump nearly 1M PPD.    

teamtempest

Different WUs put different demands on a GPU. The non-covid related one that just finished after four days dragged down video performance on my machine the whole time. Right now it's working on a free energy calculation for the moonshot, which will take only 3 or 4 hours and video performance is much snappier. Not as nearly many points as the hard one, but if I keep getting free energy calculations, I'll show about 20 WUs instead of just one for the same time period.

cine_chris

49!! A few more over the next couple of months, but It'll take about a month to get there.
It looks like some of the high flyer teams have faded drastically over the last several weeks.   Looking at the PPD rate difference (circled), compared to our current 119M PPD, some groups of donors are abandoning en masse.
Compare that to anonymous, it would be an interesting study. 
 

Tramp78

Seems the secret to moving up the list is to not stop. The overall production of everyone globally has fallen over time from the peak a few months back. Guiess people got bored or lost interest. I'm still going to keep doing this until a vaccine pops up (and then I'll probably switch to 50% for whatever else they want). The problem hasn't gone away yet so I'm still going to do what little I can to help out. I honestly don't know if folding helps at all but it has to be more productive than sitting at home doing nothing. 

teamtempest

Got my prize in the mail. Thank you, Ryau!

ryau

teamtempest said:

Got my prize in the mail. Thank you, Ryau!

You're welcome! Thank you for your contribution to the team!

cine_chris

Latest Moonshot Info https://foldingathome.org/2020/08/24/covid-moonshot-sprint-3/
Some interesting graphics, but some problems with Sprint#3 compounds
"We detected an abnormally high rate of failures with the first preparation of Sprint 3 compounds, so we’ve temporarily rolled back to Sprint 2 until we can investigate. Thanks for your patience!"

cine_chris

Currently, looking at Team 24hrAvg,  MCO@#35 marks an interesting gap in team PPD.  115M > 95M.
I also noticed that many teams are stabilizing at current levels after the March-April surge.
Note that small clusters are moving into the Top50, like Element AI, Turing Group, EVL, & UoK in this chart.  
Updated: Looking at this chart makes me think my 3-4M PPD dip had a negative impact.  Recruiting just a few dedicated Folders can make a huge difference over the long haul.

ryau

I'm interested to see what the new nvidia and amd graphics cards will produce. We might start to see some steady gains as people start upgrading their machines. I did think about putting another 2070 Super in my machine, but I'm waiting to see what's going to be launching.

Kirito

Got my card!  Planning my next trip to Cambridge!

Thank you!

Kirito

ryau

Kirito said:

Got my card!  Planning my next trip to Cambridge!

Thank you!

Kirito

You're welcome! Thank you again for your contributions to the team

cine_chris

Sustain mode, waiting to see what Nvidia & AMD will deliver in the fall.  This was the best I've seen, 8.1M PPD, from the current setup.
I try to maintain a target level overall of 7.5-8M PPD.  When it dips out of range, I start tweaking.
Local, my monitor/working system, old Z97 mobo, E3 Xeon, GTX1660Ti.
Target range: 900K-1M PPD
frame, old Q87 mobo, i7-4770k, RTX2060KO, RTX2060sup
Target range: 3M PPD
giga, new X570 mobo, Ryzen5 3600, RTX2060sup, RTX2070sup
Target range: 3.6-4M PPD
My current goal is 300M for August, 800M overall, my spreadsheet shows an 8/31AM arrival target.

kmiller922

Mail came today thank you  @ryau

ryau

kmiller922 said:

Mail came today thank you  @ryau

You're very welcome, thank you!

cine_chris

Thx @ryau I received a mysterious envelope with no return address, I disappointedly opened the accompanying  MC credit card bill first hoping to find a surprise.       LOL.
Also received notice from PantherX on the FAH GitHub team, that my whining about GPU WU allocation is still in their sites. 
tl;dr FAH currently doesn't allocate WUs to match the hardware you are running, hence your gear appears that it is being throttled with resultant low PPD yields. 

MightyMayfield

Upcoming Tuesday should be pretty cool. Looking forward to continued progress for folding!