BOINC Community Board

[Gorgon]

WCG June 3rd Update:

"By our best estimate, this issue and others related to user experience on the WCG website should be fixed by early next week."

[Ithanul]

On 5/4/2022 at 7:31 AM, GOTSpectrum said:

I would have done it but didn't wanna stop on anyones toes, i'm the folding team leader, I'm happy to do it next year if no one else does. 

 

if you can scrape together a way to monitor points i'd be happy to host an internal event so we can better utilise our resources and have prizes 

 

 

Sorry, slip my mind to get the team register for the Pent plus darn date wind up right when I got thrown on orders for a bit.  This whole year been a busy one for me.

 

If folks want, maybe can hash out a mini BOINC event to do later this year.  

[rkv_2401]

On 6/6/2022 at 1:57 AM, Gorgon said:

WCG June 3rd Update:

"By our best estimate, this issue and others related to user experience on the WCG website should be fixed by early next week."

It's still not back online...and they've removed this post. LOL!

[Gorgon]

5 hours ago, rkv_2401 said:

It's still not back online...and they've removed this post. LOL!

The website and Forums are back on-line but they're still having issues with the work assignment and collection so tasks aren't being sent out yet. It's progress though.

 

The timing is unfortunate as those of us in the Northern Hemisphere are at the start of the summer meaning many of us are scaling back production to manage the heat.

[dogwitch]

fyi. if you see any lt admin not log on in a very long time. for  the team. let me know. am taking over none active adim of ltt teams on boinc

[kiska3]

Since this is related to BOINC, I present to you updated munin(if you ever used https://munin.kiska.pw/munin/)

 

New feature: Grafana
https://grafana.kiska.pw/d/boinc/boinc?orgId=1
https://grafana.kiska.pw/d/boinc-log/boinc-log?orgId=1

2 new dashboards, one linear and the other log base 10.
These dashboards feature numerous graphs some the same from munin and new ones.
Graphs include:

• Results(ready to send and in progress)
• Workunits(Assimilation, Validation and Deletions)
• Transitioner backlog
• NEW: Compute Performance
• NEW: Per app results(ready to send and in progress)
• NEW: Per app runtime(min, avg and max)
• NEW: Per app user count
• NEW: Status(offline or online)
• NEW: (HTTP)Response Time
• NEW: Uptime percentage gauge

Project selector on top so no more endless scrolling to find the project of choice also on top is date picker to change time range of data

[danwat1234]

It looks like World Community grid is back up and running! Most of my computers have received a full queue of mapping cancer marker work units. It may be stealing some computational power from Sidock@home because i see global power went from about 50 to 44Tflops

[Gorgon]

2 hours ago, danwat1234 said:

It looks like World Community grid is back up and running! Most of my computers have received a full queue of mapping cancer marker work units. It may be stealing some computational power from Sidock@home because i see global power went from about 50 to 44Tflops

It appears they are still having networking issues at their co-lo facility so its more of a dribble than the faucet being open full but still some great progress.

 

The project names have changed a bit though for app_config.xml files:

mcm -> mcm1

opn -> opn1

 

but opng is the same but though I'm getting a lot of mcm and a few opn tasks the opng are still only a couple a day across 4 systems with GPUs on BOINC.

 

I just finished putting 7/9 rigs back on WCG after noticing the traffic yesterday and having to shut it down (30C in the basement 8-).

[rkv_2401]

I didn't even know WCG was active again despite checking out their page a couple of times. Just tried to accepting new tasks on a hunch, got a whole bunch of MCM tasks but the data for them isn't downloading properly. And the website, which used to be snappy before, has slowed to a crawl. Ugh.

[Gorgon]

18 hours ago, rkv_2401 said:

I didn't even know WCG was active again despite checking out their page a couple of times. Just tried to accepting new tasks on a hunch, got a whole bunch of MCM tasks but the data for them isn't downloading properly. And the website, which used to be snappy before, has slowed to a crawl. Ugh.

They're still having issues with their networking infrastructure. They don't have nearly as many servers as they need on-line yet so Downloads get stuck. Often you need to go into the Transfers tab in BOINC Manager and select all the stuck downloads then mash retry a bunch. The good news is once you fill the queue it's a lot less often this needs to be done

[kiska3]

Well it seems those participating in milkyway@home something exploded about 10 days ago, and it has been limping...

Spoiler

 

[Gorgon]

On 9/25/2022 at 9:20 PM, rkv_2401 said:

I didn't even know WCG was active again despite checking out their page a couple of times. Just tried to accepting new tasks on a hunch, got a whole bunch of MCM tasks but the data for them isn't downloading properly. And the website, which used to be snappy before, has slowed to a crawl. Ugh.

As of the last couple of days it appears that the task download issue on WCG has finally been resolved.

[GOTSpectrum]

On 9/29/2022 at 4:22 PM, Gorgon said:

As of the last couple of days it appears that the task download issue on WCG has finally been resolved.

I have seen reports of the issue yesterday over on discord btw 

[Gorgon]

2 minutes ago, GOTSpectrum said:

I have seen reports of the issue yesterday over on discord btw 

Yeah, it worked great for a few days then broke again

[Gorgon]

18 hours ago, antonymagnus said:

Yeah, just recently got my toes wet getting things set up and noticed WCG is totally screwed. Nothing new right?

Correct - Nothing new there. IBM dumped WCG and Krembil Research took the reins and moved the platform to their Data Center starting back in February, a move that was supposed to take 2 months. Now we are in November and things are improving but they still have network/load-balancer issues resulting in stalled downloads and their Device Manage on the website is having trouble with new devices not showing up in a users profile.

 

I was quite active there but have pulled my CPUs and GPUs off WCG and Einstein during the LTT Folding Months.

 

What I ended up doing to keep my BOINC slots well-fed was to run WCG to get OPNG GPU tasks and OPN & MCM CPU tasks and I also ran Einstein@Home with a project weight or Priority of 0 for GPU tasks as well. That way whey there were no OPNG GPU tasks E@H tasks would keep my BOINC GPUs busy.

[Gorgon]

Yea! - Hit 1 Billion points in BOINC today using mostly just Einstein@Home and WCG. No Collatz or other questionable projects that give ridiculous points compared to others.

 

Now if OpenPandemic GPU tasks would start again I might actually be able to take the crown back from @danwat1234

[Gorgon]

On 10/5/2022 at 8:59 PM, GOTSpectrum said:

I have seen reports of the issue yesterday over on discord btw 

Near as I can figure out reading the Forums at WCG the issue is Storage Bandwidth and they are waiting on a bunch of Enterprise SSDs for an upgrade which they hope will fix the issue. It's been OK for the last couple of days but it seems every time they launch a batch of ARP WUs it brings the download issue back.

[Gorgon]

On 1/28/2023 at 9:43 PM, Gorgon said:

Yea! - Hit 1 Billion points in BOINC today using mostly just Einstein@Home and WCG. No Collatz or other questionable projects that give ridiculous points compared to others.

 

Now if OpenPandemic GPU tasks would start again I might actually be able to take the crown back from @danwat1234

@danwat1234 OPNG is Back - I'm coming for you

[Gorgon]

World Community Grid (WCG) has had a bad week. Late in the weekend and moving into this week they had a problem with their Task scheduler where they ended up running most systems dry of tasks. They got that resolved then yesterday evening the RAID Controller on their Storage Server died resulting in no work being sent out nor being able to be returned and as an added bonus their web site being down.

 

As usual the best source for their status is their Twitter account:

https://twitter.com/WCGrid/status/1631001765431705611?s=20

 

Luckily I had enough time to refill my 2-day work buffer so as long as they get it resolved by the weekend I should have enough tasks to crunch.

[Gorgon]

On 3/1/2023 at 8:35 PM, Gorgon said:

World Community Grid (WCG) has had a bad week. Late in the weekend and moving into this week they had a problem with their Task scheduler where they ended up running most systems dry of tasks. They got that resolved then yesterday evening the RAID Controller on their Storage Server died resulting in no work being sent out nor being able to be returned and as an added bonus their web site being down.

 

As usual the best source for their status is their Twitter account:

https://twitter.com/WCGrid/status/1631001765431705611?s=20

 

Luckily I had enough time to refill my 2-day work buffer so as long as they get it resolved by the weekend I should have enough tasks to crunch.

@danwat1234 The latest updates from today is that the Fielsystems have all finished the rebuild with the new RAID Controller and as of a couple of hours ago they were just running a file system consistency check before going live again.

 

I suspect it will be late this evening or early tomorrow before they're back in Production. The sooner the better as I'm starting to see Completed work that couldn't be Uploaded being aborted as it is past the deadline.

[danwat1234]

2 hours ago, Gorgon said:

@danwat1234 The latest updates from today is that the Fielsystems have all finished the rebuild with the new RAID Controller and as of a couple of hours ago they were just running a file system consistency check before going live again.

 

I suspect it will be late this evening or early tomorrow before they're back in Production. The sooner the better as I'm starting to see Completed work that couldn't be Uploaded being aborted as it is past the deadline.

 

[Gorgon]

19 hours ago, danwat1234 said:

As of this morning 2 out of my 9 systems running WCG have run dry with a third down to 5/22 threads in operation. How are you faring with your giant pile of SFF systems?

 

Krembil doesn't seem to recognize that while their staff's time is more expensive on weekends the electricity their volunteers use is cheaper an this is when more work will get processed if it is available.

[danwat1234]

4 hours ago, Gorgon said:

As of this morning 2 out of my 9 systems running WCG have run dry with a third down to 5/22 threads in operation. How are you faring with your giant pile of SFF systems?

 

Krembil doesn't seem to recognize that while their staff's time is more expensive on weekends the electricity their volunteers use is cheaper an this is when more work will get processed if it is available.

After looking at a few of them it looks like most have completed all WCG work and are now working on Rosetta or Sidock. Have them set for 10 days queue. My house is as warm as ever so really I judge whether they have work to do by heat. Gridcoin magnitude has gone way down from 1600 to 1000 now but I believe that is because Sidock's work units take multiple days now so it should bounce back https://www.gridcoinstats.eu/cpid/33c55b16c9b0fcc5446bc4e21803fec4/staking#subMenu

[Gorgon]

5 hours ago, danwat1234 said:

After looking at a few of them it looks like most have completed all WCG work and are now working on Rosetta or Sidock. Have them set for 10 days queue. My house is as warm as ever so really I judge whether they have work to do by heat. Gridcoin magnitude has gone way down from 1600 to 1000 now but I believe that is because Sidock's work units take multiple days now so it should bounce back https://www.gridcoinstats.eu/cpid/33c55b16c9b0fcc5446bc4e21803fec4/staking#subMenu

Ahh, OK so your doing GridCoin. I only use my CPUs for WCG and 4 of my older GPUs for WCG and Einstein@Home when there's no WCG OpenPandemics to do.

 

My basement's still at a toasty 25C as my CPUs have only gone from 480 to 290W so far but my GPUs are running at a constant 1.7kW.

[danwat1234]

----Latest electric bill shows 304KWh used per day (/24 12.66KW average). House members use little power ourselves it's mostly the crunchers. That includes a ~1PB Chia farm maybe 1.3KW. House heat is always off, and it is gas. SiDock #1 Rosetta #10 because of WCG down. Usually WCG gets most of the work because SiDock does not allow much Work Unit caching, and Rosetta usually has no work but coincidentally does now.
Note; i have undervolted all the boxes i can as far as possible to increase efficiency. Skylake-class typically undervolt by 140mV Haswell 70mV