Join the MicroCenterOfficial Folding@Home Team: #257944

Kirito

You can get GUI temp sensor monitoring from all system sensors using any of these apps.

https://www.tecmint.com/monitor-cpu-and-gpu-temperature-in-ubuntu/

I use lm-sensors.

Kirito

ChrisHarshman

@Kirito, thanks I'll check those out, maybe the command line ones will let me look at the GPU temps through a ssh session instead of logging into the actual box.

kmiller922

@MightyMayfield Top 300! congrats

MightyMayfield

kmiller922 said:

@MightyMayfield Top 300! congrats

Don't forget to congratulate yourself, too!



But really, thank you to everyone who has continued to support the team. I really appreciate it!

cine_chris

MCO was #33 for the week with 115,136,488 PPD avg
Team Stats for the week 24hr Avg Top 20  

Kirito

I had been hoping to break into the top 100 but oh well.

Sadly, I had to turn one of my servers off.  Apartment was getting too hot to operate it and my electric bill went up $50 on that machine alone from the power draw of all the fans in it and the heat it produced.  The saddest part was the amount of heat coming off the RAM,

which has its own cooling fans.  I have nine fans in the case and it wasn't enough.  Repasted the CPU's, made sure the laminar airflow through the case was functioning properly and still couldn't keep it cool enough.  It's a Lenovo C20x, both Xeon processors and 48 GB of RAM in 12 slots.  My normal electrical before FAH was $80 to $90.  This month was $197.  This one C20x was drawing $50 a month.  If I still have it this winter I will turn it back on to heat my apartment again. 

Kirito

cine_chris

Perhaps this isn`t appropriate for the MC forum, but I thought it was interesting?
OMG, Nvidia has more than doubled the PPD output of SaturnV!!!!
I was admiring their output in the 400M PPD range just a few days ago, but today...

2

NVIDIA_Saturn_V

NVIDIA Corp

1,026,374,441 24Hr/Avg

Could this be an Ampere upgrade???  I think the host count is in the same range @1456!!!
If I recall accurately this is 217% jump, then I did more checking and 6/28 was a 375% spike.
Autonomous is much more, but with 169351 hosts
I see they put in a massive spike several days ago also, nearly 4B PPD.  Next time WU get slow, I know where to look.

 

ChrisHarshman

Wow, much lower score yesterday.  I noticed at one point all 4 of my GPUs were waiting to get a new assignment.  Oh well. 

cine_chris

ChrisHarshman said:

Wow, much lower score yesterday.  I noticed at one point all 4 of my GPUs were waiting to get a new assignment.  Oh well. 

Apparently it was a server problem, mentioned on the EVGA forum.  As I`m retired & cowering-in-place I noticed the prob early.  Pausing the slot, with a right-click:pause, then right-click:fold resets the wait timer.  Had to do that many times yesterday.  The CPU heavy guys fared much better I suspect as MM was >30M for the day.   Anonymous took a huge hit, >5B PPD.  I also noted a similar spiked traffic pattern to Nvidia`s Saturn_V which I posted above. 

cine_chris

Guess who`s now a Top50 Folder (24hr avg) @MightyMayfield
After noticing that I was pleased to see that I was in the Top200, but I was also disappointed that it indicates enthusiasm & support is waning, but noting the heavy hitters, like Anonymous, pick-up slack capacity & their average slope is positive.  

0legravy_leg

anyone else notice much slower GPU processing (or longer TPF) with some of the WU the last couple of days?  I have had a few on my rx5700 cards that would run really long TPF and reduce PPD to well under 1m when they were normally closer to 1.2m even on 2080ti some of them would reduce PPD to under 3m due to the longer run times.  Also wondering if the recent windows updates impacted performance? 

cine_chris

0legravy_leg said:

anyone else notice much slower GPU processing (or longer TPF) with some of the WU the last couple of days? 

Yes, see comments above.
Also, I recently looked into what I call `toxic WUs`. My RTX cards were choking on them. What I found out: these are lower atom count models. Apparently, projects like Covid-Moonshot have many of these. Checking into GPU utilization & power they were both down to match the low PPD yield. Turns out based on a very recent GitHub thread discussion... there is no WU allocation moderation to match the slot device capability. Based on Chodera`s comment "Just a heads-up that we're working on extensive benchmarks of a variety of systems to be able to gather data to be able to predict performance on individual GPUs based on some easily-computed features of any project." So, they`re working on methods to improve the process. Anyway, I felt better realizing lowerPPD was lower hardware burn and my whining about toxic WUs catalyzed a GitHub enhancement request to improve WU allocation.

0legravy_leg

cine_chris said:

0legravy_leg said:

anyone else notice much slower GPU processing (or longer TPF) with some of the WU the last couple of days? 

Yes, see comments above. I looked into what I call `toxic WUs`. My RTX cards were choking on them. What I found out: these are lower atom count models. Apparently, projects like Covid-Moonshot have many of these. Checking into GPU utilization & power they were both down to match the low PPD yield. Turns out based on a very recent GitHub thread discussion... there is no WU allocation moderation to match the slot device capability. Based on Chodera`s comment "Just a heads-up that we're working on extensive benchmarks of a variety of systems to be able to gather data to be able to predict performance on individual GPUs based on some easily-computed features of any project." So, they`re working on methods to improve the process. Anyway, I felt better realizing lowerPPD was lower hardware burn and my whining about toxic WUs catalyzed a GitHub enhancement request to improve WU allocation.

Awesome info thanks, i started going through and updating drivers on my GPU and Mobo chipset after the windows updates thinking it might be something related but did not see improvements in performance. also ran into an issue with my rx5700xt, seemed like it wouldn't wake up but after rebooting i wasnt getting any graphics, and i had a 02 qcode and white diagnostic LED, had to re seat my GPU, which then led to my tempered glass side panel exploding when i put it back in place.. not a good night last night.. 

cine_chris

@0legravy_leg 
that`s a mess.  Your gear is much fancier than mine.  I did have one blue LED fan, but I replaced it with a plain black 62 cfm Sickleflow 

MightyMayfield

Holy moly, how tight was that tolerance that a side panel exploded!?

kmiller922

Ouch!

ryau

0legravy_leg said:

Awesome info thanks, i started going through and updating drivers on my GPU and Mobo chipset after the windows updates thinking it might be something related but did not see improvements in performance. also ran into an issue with my rx5700xt, seemed like it wouldn't wake up but after rebooting i wasnt getting any graphics, and i had a 02 qcode and white diagnostic LED, had to re seat my GPU, which then led to my tempered glass side panel exploding when i put it back in place.. not a good night last night.. 

What a pain! Looks like a sweet build though!

cine_chris

MCO is now #58 overall with a +2 jump for the week, also a +2 for 24Hr-avg to 28th spot!
Some of the `new faces` & LoneWolves are helping on the team`s +2 jumps.

kmiller922

Has anyone else been getting a lot of bad work units seems to be the reason for my ppd dropage?

MightyMayfield

kmiller922 said:

Has anyone else been getting a lot of bad work units seems to be the reason for my ppd dropage?

One particular machine has been getting nothing but bad WUs on graphics card slots. Dropped from an output of around 1.2mppd to about 12000. Sure others have been as well and I haven't seen them.

cine_chris

kmiller922 said:

Has anyone else been getting a lot of bad work units seems to be the reason for my ppd dropage?

@kmiller922 did you see my comment on `toxic WUs` above?  I seldom run CPUs for F@H and have a mix of GTX & RTX GPUs.  The recent mix of lower atom count WUs has some cards running at 50-60% capacity.  Surprisingly, my little GTX1050ti is at 150% excursions!!  I`ve seen my current mix of gear running between 5-9M PPD, but the recent mix of WU has been a challenge.  

kmiller922

The bad wu's i've been getting have all been on rtx 2080ti just thought it was a bit strange.

kmiller922

Congrats to Microcenter045 top 1000

ryau

Nice work everyone! Nice to see we're continuing the fight.

My machine has seen some odd issues, not with bad WUs, but not being able to submit some of the CPU WUs, so they're just hanging. A couple of power outages this weekend and a failing UPS didn't help either. Now, if I only knew a place I could get a new UPS...

ChrisHarshman

kmiller922 said:

Has anyone else been getting a lot of bad work units seems to be the reason for my ppd dropage?

I had a problem on 2 of my GPUs yeterday, They weren't even trying to download new work units, next attempt was unknown even though they were sitting there idle in Ready status.  I had to completely reboot the computers for F@H to download a new work unit once I noticed it, as other things I tried didn't work.

ChrisHarshman

0legravy_leg said:

...ran into an issue with my rx5700xt, seemed like it wouldn't wake up but after rebooting i wasnt getting any graphics, and i had a 02 qcode and white diagnostic LED, had to re seat my GPU, which then led to my tempered glass side panel exploding when i put it back in place.. not a good night last night...

Wow!  hope everything's alright after that.  Looks like you're still killing it points-wise though.

0legravy_leg

ChrisHarshman said:

0legravy_leg said:

...ran into an issue with my rx5700xt, seemed like it wouldn't wake up but after rebooting i wasnt getting any graphics, and i had a 02 qcode and white diagnostic LED, had to re seat my GPU, which then led to my tempered glass side panel exploding when i put it back in place.. not a good night last night...

Wow!  hope everything's alright after that.  Looks like you're still killing it points-wise though.

yes seems to be back on track and knocking off work units like a champ!!

cine_chris

kmiller922 said:

The bad wu's i've been getting have all been on rtx 2080ti just thought it was a bit strange.

@kmiller922 has your RTX2080ti recovered with the current mix of WUs?  The last 2 days have been better for me.

cine_chris

ryau said:

My machine has seen some odd issues, not with bad WUs, but not being able to submit some of the CPU WUs, so they're just hanging. A couple of power outages this weekend and a failing UPS didn't help either. Now, if I only knew a place I could get a new UPS...

@ryau I gave up on trying to keep UPSes functioning. Did your WUs get ingested? I had a few GPU WUs backup several weeks ago, iirc they cleared without drastic measures. I checked on other forums and it was attributed to server issues. Don`t know if I got PPD credit or not. In general, the FAH client has always impressed me and the remote access has been very solid. Wish there was more logging features to track performance.

kmiller922

cine_chris said:

kmiller922 said:

The bad wu's i've been getting have all been on rtx 2080ti just thought it was a bit strange.

@kmiller922 has your RTX2080ti recovered with the current mix of WUs?  The last 2 days have been better me.

I had to do a restart of the system but it seems to have corrected, almost seems like there was a uncomplete cleanup that may have caused the issue. Still having a lot of bad wu usually around 43-73% through the wu, but still chewing away at it.