Change in BA requirements

[Bill1024]

You're welcome

With all due respect, I beg to differ.
Saying please is a courtesy when you ask a favor. Then you say thank you when it's done.
Several people stated they want communication, and all they have to do is ask.
Well we got what we asked for. Some respect and communication.
They could have just ended the bigadv saying the science is all up to date and will be for a month or two.
Or they can make it 32 core needed to get bigadv and you have 24 hours to do it.
The point system has always been a problem and never made everyone happy.
Lets just offer a hand to get this done and show that we can step it up when needed.


They raise the points for SMP the GPU folders will complain.
Then they raise the points on the GPUs, then the bigadv folders will complain.
So they raise the points for bigadv, then no one wants to fold SMP.

We came full circle

[P5-133XL]

The real solution is so obvious - PG created too much demand for bigadv by making the reward high. All they need to do is create more demand for SMP by making its reward appropriately high - whether on a temporary or permanent basis. Begging seems like an acknowledgement that they really don't know how to manage their own points system.

There is already a severe problem with point-inflation. Adding points to SMP just aggravates the issue. Doing so temporarily just causes alienation when you stop giving them. The SMP people will react no less aggrieved to a decrease in SMP points than the bigadv group does for the exact same reasons.

[ChristianVirtual]

They raise the points for SMP the GPU folders will complain.
Then they raise the points on the GPUs, then the bigadv folders will complain.

not so sure on those two points; GPU and CPU are different categories with no overlap ... I don't think a GPU folder would care too much about CPU points (and a moderate raise); actually with Core17 and QRB the GPU have not much to complain about.

[P5-133XL]

They raise the points for SMP the GPU folders will complain.
Then they raise the points on the GPUs, then the bigadv folders will complain.
So they raise the points for bigadv, then no one wants to fold SMP.

We came full circle

Theoretically, Core_17 projects are benchmarked against SMP to make them equivalent. If you raise SMP points, GPU points should follow. Do note PG historically has not rebenchmarked previously released projects but only applied changes to just new projects. It is why Core_11/15/16 do not give out QRB.

[Bill1024]

They raise the points for SMP the GPU folders will complain.
Then they raise the points on the GPUs, then the bigadv folders will complain.

not so sure on those two points; GPU and CPU are different categories with no overlap ... I don't think a GPU folder would care too much about CPU points (and a moderate raise); actually with Core17 and QRB the GPU have not much to complain about.

Trouble is the gpu points are based off the smp points as in "equal work for equal pay." (oh the gpu folders did complain ) Ever hear equal work for equal pay?
So if they raise the smp points they have to raise the gpu points again.
Then when those get too high the guys who spend thousands on a 4P system want their points raised.
Hence we went full circle.
We can go round and round and round.

[kerryd]

OK.

Please aim your resources at SMP rather than BA.

(I've got to admit I feel pretty stupid doing this.)

And heres my post at EVGA
I will put crunchers to no more work and move over my 4p's and all but 24/7 cruncher.BUT if no word on not upping the core count by mid Jan. I will be done with stanford for good.I will not play a game with them they asked for help so I will give it but its a 2 way street.

[Grandpa_01]

You say Please will not work. How do you know? Has PG put out the call and ask everyone to knock off the back log?

He says it, because he knows it. We all know it. PG has asked before and it didn't work. Some of us have been doing this for 10+ years. We know what works and what doesn't.

For example, when I suggested earlier that all the BA people run SMP WUs for the next 30 days so as to delay the next BA core hike they all laughed and joked. It's a serious suggestion but no one seems interested in serious solutions to serious problems. No 48 core people willing to "lose a few points" for their 24 core BA brothers and sisters.

Bill, you were so willing and cooperative to being asked to change BA settings in a few months that you turned off your systems. Ya, asking really worked good there. It's hard to take that seriously.

We're victims of our own human nature. As a result, Stick works, and Carrot works. Nothing else moves the masses. So be it.

That is not true I ran smp the last time they asked us to, I also said at that time they needed to adjust the points for smp projects or they would just have the same problem again and several of you said that the people that would no longer be able to run bigadv would run the smp, I believe the reasoning was that there would be nothing else for them to run and they would not just shut there machines down. Guess what they did or they switched to GPU or built another bigadv rig or just said the heck with it and moved onto another project.

Thinking people are going to run smp was foolish then and it is foolish now people will just shut them down.

The only thing that is going to fix the problem at this time is an incentive for people to run smp.

When I bring my machines back online I will be running smp for a 2 week period if after that 2 week period there is no improvement in there pay I will go back to bigadv. There is a simple cure make it worth the while, if that does not happen then I would say it really is not that big of a problem.

[Bill1024]

Tim wrote
Bill, you were so willing and cooperative to being asked to change BA settings in a few months that you turned off your systems.

Well Tim
There is a big difference being asked and being told to to something.
Now one asked up until now, no one even said there was a backlog.
Oh Tim Tim Tim why why why.......................................

[Adak]

To address the points problem, I believe we need to increase SMP points by 5% and decrease bigadv projects by 5%, to help balance out the large disparity between those projects, on any PC. Just the opinion of one bigadv folder. Yes, there would be gripes from some bigadv folders, but it would be small, overall. We know we've had a tremendous points windfall since the bigadv project was initiated.

I also believe we've lost where our focus should be, in this thread. I agree that the FAH points system has always been a bit "wild", and many of the comments made in this thread, are completely correct and justified.

HOWEVER

Points are just for a bit of fun and/or competition. They are inherently useless, otherwise. It is the SCIENCE that we need, not the points. I'd prefer a better points system. It's quite upsetting when your rig may lose a large portion of it's ppd, due to an administrative edict -- noted. Let's try to remember however, why we started folding, and what our real goals are - and not focus too much on our points.

[Xavier Zepherious]

Ok

the solution is obvious


first
to reduce people building BA server systems - reduce Bigadv points - they are way over-inflated
when I went from bigadv to smp - I got 1/3 of the PPD I was producing before - in no way is a point drop of that much is fair on the same hardware (just based on the switch from BA to SMP)
or raise both smp(&gpu) and lower bigadv -and make smaller systems more attractive again

look SMP points do not have to equal bigadv , but the points shouldn't be 1/3 of bigadv
SMP values have been shrinking over time - you wonder why SMP folders are declining and why everyone is trying to go GPU or BA

I would be happy with 70% - giving a 30% bonus to bigadv rigs

2nd

priority queuing
you should have programmers that can set the software - assignment server to see backlogs of a project - increasing there assignment - even to bigadv
you can also do it based on outgoing projects and the % of whats going out - if BA goes over 4% it starts assigning SMP until its back to 2%

3rd
benchmarking systems

test systems every few days(or week) with a test unit (5 or 10 min WU)- allowing machines to do BA that can meet the Prefered - not final deadline
easily to embed a hash(in a encrypted reference file - with time marks) and encode it so it can't be tampered with
a 5 min test unit doesn't bite too much into anyone's pocket - and it's fair across the board

By doing this you would not need to change CPU guidelines
because machines will be tested and will be barred from doing BA that don't meet it
it will also allow machines once barred to able to do BIGADV again - if they meet the deadlines

sure You may have a few more going back to BA - but if you institute Priority assigning then the BA systems get SMP wu's when there is a backlog
They may whine about doing a bunch of SMP WU's for awhile - but that's no different than GPU systems getting core 15 instead of core 17 Wu's (we take our lumps they can take some too)
And yes that should be there risk on building a BA system that they may get SMP WU's
just like anyone with kepler cards get core15 WU's - there are no guarantees on what WU's you get

[Nathan_P]

When I bring my machines back online I will be running smp for a 2 week period if after that 2 week period there is no improvement in there pay I will go back to bigadv. There is a simple cure make it worth the while, if that does not happen then I would say it really is not that big of a problem.

That's a fair compromise - at least it will have some data to prove to PG what the issues are - I'll switch over my "slowest" rig as well

[tear]

OK.

Please aim your resources at SMP rather than BA.

(I've got to admit I feel pretty stupid doing this.)

Credibility is essential prerequisite for making such requests.

You are in no position to make such requests, bruce.

[orion]

OK.

Please aim your resources at SMP rather than BA.

(I've got to admit I feel pretty stupid doing this.)

Credibility is essential prerequisite for making such requests.

You are in no position to make such requests, bruce.

True.

This is something that Dr. Pande needed to ask of us very much.

[Punchy]

There is already a severe problem with point-inflation. Adding points to SMP just aggravates the issue. Doing so temporarily just causes alienation when you stop giving them. The SMP people will react no less aggrieved to a decrease in SMP points than the bigadv group does for the exact same reasons.

It's easy to criticize; how about you provide a better solution?

And I have to agree with tear; if there is a need for help with SMP, let's hear it from PG themselves as a direct request. I don't think they will do so, since it's basically an admission that they broke the points system.

[Skripka]

To address the points problem, I believe we need to increase SMP points by 5% and decrease bigadv projects by 5%, to help balance out the large disparity between those projects, on any PC. Just the opinion of one bigadv folder. Yes, there would be gripes from some bigadv folders, but it would be small, overall. We know we've had a tremendous points windfall since the bigadv project was initiated.

I also believe we've lost where our focus should be, in this thread. I agree that the FAH points system has always been a bit "wild", and many of the comments made in this thread, are completely correct and justified.

HOWEVER

Points are just for a bit of fun and/or competition. They are inherently useless, otherwise. It is the SCIENCE that we need, not the points. I'd prefer a better points system. It's quite upsetting when your rig may lose a large portion of it's ppd, due to an administrative edict -- noted. Let's try to remember however, why we started folding, and what our real goals are - and not focus too much on our points.

5% adjustments in no way make a difference. It is like trying to hold the ocean back with a broom. Increasing the PPD of say a i7-2600 at stock from 8-10K PPD to...8400-10.5K PPD is not an "incentive" to fold SMP. It is a joke. We're focusing on the power and points because at the end of it THAT is entirely why people quit doing SMP. Because for the electrical bill, the points are a terribly unfunny joke.

Asking donors to drop their PPD by 1/2 at a minimum if not more and keep the electrical bill constant is just silly. No one should forget, this little project costs donors money every month of course they're going to turn on clients in the manner most advantageous to them. I.e. with the most PPD.

Hell given the ludicrous TPFs for smp on 1p, they should call it BigAdv-Long. As the TPFs on 1p SMP are twice what they are doing BigAdv on 4p. That fundamentally is the problem and reason why we have a "backlog" of SMP units. For 1p folders the TPFs for SMP are too massive for the WUs to be completed in a reasonable amount of time, and the returns are crap even if you have a 100% folding and nothing else 1p desktop SMP system. And guess what? Donors put their money and effort into systems that get better yield and for the electrical bill-a FAR FAR better yield per Watt than SMP.

They need BigAdv clients to do SMP because Stanford has gone and made the SMP units ludicrously big and long (larger in TPF than bigadv for a 1p at stock)....and surprise surprise....the SMP clients online cannot finish them fast enough, and when they do the points are a joke. Only the 2p and 4p and higher server blades normally doing bigadv have enough computational power to calculate SMP WUs in a reasonable amount of time.