LTT Official Folding Month 2020!!!

[Egg-Roll]

1 hour ago, chaozbandit said:

Mind you this is only for my hardware and isn't representative of what others may experience, just some food for thought.

Sounds about right, since I'm the only one crazy enough to run a 640 lol... It's currently working on another 14908 and every time I've checked it was always that one, so no idea where the 13000 ones came from, just like idk how the Cancer one sneaked in, but it did.

https://folding.lar.systems/folding_data/gpu_ppd_datasheet?gpu_model=gk107_geforce_gt_640_693

[porina]

3 hours ago, yaboistar said:

They already are on the used market for less than a 3070. I've seen one go for under 400 quid here

If used 2080Ti's were going much below where a new 3070 would be, it could bring them into consideration for me. Still, in a quick look on ebay they're mostly finishing 600 upwards. Might as well see what 3070 availability is like in 2 days...

[NetoriusNick]

SR-2 can do PCIe 1-6 x8 PCI 7 x16 ooppa  next question is x8 too slow

[Egg-Roll]

2 hours ago, porina said:

If used 2080Ti's were going much below where a new 3070 would be, it could bring them into consideration for me. Still, in a quick look on ebay they're mostly finishing 600 upwards. Might as well see what 3070 availability is like in 2 days...

I'd wait till around/after black friday/cyber monday, hopefully stores get some stock in and keep it. Until they do expect the 2080 Ti's to float around the price of a 3070 maybe even higher, not like anyone should buy it for more than a 3070 but for half the original price it's a good deal esp if only used for light gaming.

[Dr.Roboknopp]

16 hours ago, cbigfoot said:

Reported for stalking me  Now back down PLEASE!!!!!

 

 @Dr.Roboknopp We all cordially request you back down that is a little extreme for this Neighbor Hood....We have a Max speed of 10M PPD.  and faster and you will be ejected  

@TVwazhere Well you are gone so no more picking on you for now

 @LazyDev you had a good run today lets see what tomorrow brings

 @Airdragonz Could you have folded any closer to my number...and im watching you.

 @Yabdat I have my eye on you!!

 

Don't you worry. I will not be able to keep this up. You can probably see me slide past you on my way down again. 

[DailyProcrastinator]

Anyone else having WU issues? Completion time has (seemingly) randomly gone up significantly. My 1080 is taking over 4hrs now, vs 2-2.5hrs it would take previously. 

[NetoriusNick]

2 minutes ago, TheDailyProcrastinator said:

Anyone else having WU issues? Completion time has (seemingly) randomly gone up significantly. My 1080 is taking over 4hrs now, vs 2-2.5hrs it would take previously. 

yes

[Egg-Roll]

1 hour ago, TheDailyProcrastinator said:

Anyone else having WU issues? Completion time has (seemingly) randomly gone up significantly. My 1080 is taking over 4hrs now, vs 2-2.5hrs it would take previously. 

Completion times seem ok for me, maybe we are doing more complex work?

 

However I'm getting this:

20:48:43:WU01:FS01:Requesting new work unit for slot 01: READY gpu:1:GP104 [GeForce GTX 1070] 6463 from 129.213.157.105
20:48:43:WU01:FS01:Connecting to 129.213.157.105:8080
20:49:01:WU00:FS01:Upload 1.11%
20:49:04:WARNING:WU01:FS01:WorkServer connection failed on port 8080 trying 80

Port 80 worked however, and the new unit is now saying 4 hours. (1070, WU 14907)

[Spotty]

49 minutes ago, TheDailyProcrastinator said:

Anyone else having WU issues? Completion time has (seemingly) randomly gone up significantly. My 1080 is taking over 4hrs now, vs 2-2.5hrs it would take previously. 

It might be a more complex work unit that takes longer to complete.

 

If you're using LARS_System's F@H Chrome plugin it will tell you how your card is comparing to other cards on the same work unit.

[Favebook]

4 hours ago, LAR_Systems said:

I image when the event is over there will be a slight drop in PPD as people set their power profiles back to "don't burn the house down" so their electric bill is not terrible all winter.

For 2080S you shouldn't worry that much. I am running all 4 of mine at 70% PL with small OC which I will keep running even after event at same settings

3 hours ago, NetoriusNick said:

I also take it that mining cards can be used if flashed to stock?

Of course. They might even work w/o flashing. Depends on card and bios that you are running

3 hours ago, Shlouski said:

I think this site is interesting and offers useful information, I mean no disrespect but its basically userbenchmark.com for folding, there are vast numbers of variables that can't be accounted for, so any info I see on there I take with a gain of salt.

Its an impossible task for this site to eliminate the vast number of variables to achieve accurate numbers, but it does a good job under the circumstances, I'm impressed.

I will repeat what has been said many times in this thread already. That database is up to date which is good reference for buying GPUs right now, and PPD is average of all OCs, PLs, BIOSes and any other settings that you could change. There is no database better and more up to date than this one IMHO. You cannot EVER eliminate all the variables UNLESS you make a database which would take EVERY variable in consideration. It would take much more time, much more energy, much more space and it wouldn't be worth it.

What @LAR_Systems made of that database is damn perfect, and soon I am going to ask him for donation address so I can give something in return as his database as well as extension has helped me much more than anything else folding related.

50 minutes ago, TheDailyProcrastinator said:

Anyone else having WU issues? Completion time has (seemingly) randomly gone up significantly. My 1080 is taking over 4hrs now, vs 2-2.5hrs it would take previously. 

That is not an issue. It's just WU difference. Some can take 20min, some can take FEW hours. Get used to it, it's luck of the draw .

[DailyProcrastinator]

26 minutes ago, Spotty said:

It might be a more complex work unit that takes longer to complete.

 

If you're using LARS_System's F@H Chrome plugin it will tell you how your card is comparing to other cards on the same work unit.

 

19 minutes ago, Favebook said:

That is not an issue. It's just WU difference. Some can take 20min, some can take FEW hours. Get used to it, it's luck of the draw .

 

Probably just a more complex WU, I just have never seen that long of a completion time. I'll pay attention to that, thank!

[Gorgon]

2 hours ago, NetoriusNick said:

SR-2 can do PCIe 1-6 x8 PCI 7 x16 ooppa  next question is x8 too slow

No, x8 would be just fine for up to and including 3080.

 

SR-2 is way overkill for folding. So unless you need an HEDT system for something other than folding a x570 and a PCIe x4/x4/x4/x4 is way less. I have run 6 High end GPUs (2070S) off an x570 with a 2700 and 8GB DDR4 but I'd recommend 16GB these days.

 

If you are need a HEDT system going x399 and ThreadRipper 1 or 2 would likely be more bang for your buck.

 

If it is a dedicated F@H build save some $ on the system so you can put it towards GPUs.

[NetoriusNick]

4 minutes ago, Gorgon said:

No, x8 would be just fine for up to and including 3080.

 

SR-2 is way overkill for folding. So unless you need an HEDT system for something other than folding a x570 and a PCIe x4/x4/x4/x4 is way less. I have run 6 High end GPUs (2070S) off an x570 with a 2700 and 8GB DDR4 but I'd recommend 16GB these days.

 

If you are need a HEDT system going x399 and ThreadRipper 1 or 2 would likely be more bang for your buck.

 

If it is a dedicated F@H build save some $ on the system so you can put it towards GPUs.

already have sr-x and sr-2

[Egg-Roll]

1 hour ago, Favebook said:

You cannot EVER eliminate all the variables UNLESS you make a database which would take EVERY variable in consideration. It would take much more time, much more energy, much more space and it wouldn't be worth it.

Mathematics can help you there, however I do agree would be a huge time sink. I've bumped into individual data about WUs completed, I've just not found any dumps. It would be interesting to look into to see if I can crack the code

[NetoriusNick]

4 minutes ago, Egg-Roll said:

Mathematics can help you there, however I do agree would be a huge time sink. I've bumped into individual data about WUs completed, I've just not found any dumps. It would be interesting to look into to see if I can crack the code

would more indepth monitoring help?

[Egg-Roll]

1 hour ago, NetoriusNick said:

would more indepth monitoring help?

Truth be told, no. However the data is there, I just don't have access to it. Being able to make a formula to separate those who are limited or OC'd their cards from the average user would give a far better detail about which card preforms the best instead of the averages we have currently.

 

Plus in it's current state there is nothing stopping some rich troll from buying all the 3090s possible and under clocking them to the point they reach my 640 cards level in completing tasks, which is currently at the bottom of the list. However it is very useless for the average person, I just fell like trying to do it

 

I have a feeling they are using the API as it does have that kind of information, but I rather a daily dump txt file over using the API.

[NetoriusNick]

6 minutes ago, Egg-Roll said:

Truth be told, no. However the data is there, I just don't have access to it. Being able to make a formula to separate those who are limited or OC'd their cards from the average user would give a far better detail about which card preforms the best instead of the averages we have currently.

 

Plus in it's current state there is nothing stopping some rich troll from buying all the 3090s possible and under clocking them to the point they reach my 640 cards level in completing tasks, which is currently at the bottom of the list. However it is very useless for the average person, I just fell like trying to do it

 

I have a feeling they are using the API as it does have that kind of information, but I rather a daily dump txt file over using the API.

i was thinking in conjunction with lar and temp /volt freq monitoring it could output something  like this 

[Egg-Roll]

1 hour ago, NetoriusNick said:

i was thinking in conjunction with lar and temp /volt freq monitoring it could output something  like this 

Not even they would have that information. Here is an example of what they would have:
https://apps.foldingathome.org/wu?p=14905&r=203&c=12&g=5#project=14905&run=5&clone=6&gen=2

I'm sure where they get the additional information from likely tells GPU with WU completion time, more than likely in one request, w/o the user data.

[NetoriusNick]

10 minutes ago, Egg-Roll said:

Not even they would have that information. Here is an example of what they would have:
https://apps.foldingathome.org/wu?p=14905&r=203&c=12&g=5#project=14905&run=5&clone=6&gen=2

I'm sure where they get the additional information from likely tells GPU with WU completion time, more than likely in one request, w/o the user data.

was thinking with the additional infomation about the freqs would help with PPD evaluations

[Spotty]

1 hour ago, Gorgon said:

SR-2 is way overkill for folding.

What is an SR-2?

All I think of is the SR-2 Normandy from Mass Effect.

[Egg-Roll]

1 hour ago, NetoriusNick said:

was thinking with the additional infomation about the freqs would help with PPD evaluations

That would give it away, also that is information the F@H doesn't collect. In order for anyone to collect that it would have to be voluntary as it's meaningless information for the devs at F@H

 

1 hour ago, Spotty said:

What is an SR-2?

All I think of is the SR-2 Normandy from Mass Effect.

https://www.evga.com/articles/00537/

It's classified.

[NetoriusNick]

3 minutes ago, Egg-Roll said:

That would give it away, also that is information the F@H doesn't collect. In order for anyone to collect that it would have to be voluntary as it's meaningless information for the devs at F@H

 

https://www.evga.com/articles/00537/

It's classified.

its classified lol

 

i was thinking for lar_systems DB of cards and work units

[LAR_Systems]

22 minutes ago, NetoriusNick said:

was thinking with the additional infomation about the freqs would help with PPD evaluations

You need to keep in mind:   PPD is just fake internet points for us geeks to have a little fun trying to dial in our gear to get some more of them.   

They provide no value to the community of the scientists using the F@H project to process their data.

 

This means F@Hs primary goals are keeping the client / work servers working well enough to get the data out to us, us to crunch it in a reliable way, and get that data back to them in a time window which is days... not the hours we all strive to get them back to them in to max out PPD.

 

You don't even really get the PPD unless you get a F@H API key... which has no bearing on the science... you can do work just fine without one, but the numbers will look low, just something fun they added so Den-Fi you could get more points and show off / have fun.

 

So with this in mind... the software does not access or keep track of the voltages, over/under clocks... particular GPU models and editions by manufacturer etc. 

F@H would not want to spend their resources or time to integrate and expand the scope of the platform to do any of this... cause it could only introduce problems or barriers to installation (admin access, driver / os patch changes etc.) more so then they already have to deal with the current base service, and all is off the mission of research, remember for most users it's really intended to be on your computer and just never be seen using spare time on your hardware.

 

This is 100% the right attitude for them to have as it's about the science first to them which I respect.

 

So you get the log analyzers, services like I'm making, forums full of people with suggestions to help squeeze a little more PPD out and have some fun with the gamification of doing some charity processing for a good cause.   

 

But in the end the only thing that matters is that the work units get completed before they expire, and that your system is running stable 100% of the time so you're not wasting F@H researcher time creating bad / incomplete work units from overclocks etc.

 

This is why a person with a room full of 3090s is as important as the people folding on their laptops, each WU counts and the points don't matter. 
(unless your Den-Fi)

 

[NetoriusNick]

1 minute ago, LAR_Systems said:

You need to keep in mind:   PPD is just fake internet points for us geeks to have a little fun trying to dial in our gear to get some more of them.   

They provide no value to the community of the scientists using the F@H project to process their data.

 

This means F@Hs primary goals are keeping the client / work servers working well enough to get the data out to us, us to crunch it in a reliable way, and get that data back to them in a time window which is days... not the hours we all strive to get them back to them in to max out PPD.

 

You don't even really get the PPD unless you get a F@H API key... which has no bearing on the science... you can do work just fine without one, but the numbers will look low, just something fun they added so Den-Fi you could get more points and show off / have fun.

 

So with this in mind... the software does not access or keep track of the voltages, over/under clocks... particular GPU models and editions by manufacturer etc. 

F@H would not want to spend their resources or time to integrate and expand the scope of the platform to do any of this... cause it could only introduce problems or barriers to installation (admin access, driver / os patch changes etc.) more so then they already have to deal with the current base service, and all is off the mission of research, remember for most users it's really intended to be on your computer and just never be seen using spare time on your hardware.

 

This is 100% the right attitude for them to have as it's about the science first to them which I respect.

 

So you get the log analyzers, services like I'm making, forums full of people with suggestions to help squeeze a little more PPD out and have some fun with the gamification of doing some charity processing for a good cause.   

 

But in the end the only thing that matters is that the work units get completed before they expire, and that your system is running stable 100% of the time so you're not wasting F@H researcher time creating bad / incomplete work units from overclocks etc.

 

This is why a person with a room full of 3090s is as important as the people folding on their laptops, each WU counts and the points don't matter. 
(unless your Den-Fi)

 

oh yeah not wrong It was more of a thought (likely too far down the rabbit hole) as a way to expand on your db not F@Hs

[NetoriusNick]

13 minutes ago, Spotty said:

What is an SR-2?

All I think of is the SR-2 Normandy from Mass Effect.

yeah I got a bug when I first saw it prior to its release so it ended up on my hardware wish list which is getting seen too in dripps and drabs