Thoughts On Credits

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,550
Credit: 288,905,381
RAC: 87,563

RE: My mom had a "Mind

Message 83669 in response to message 83666

Quote:

My mom had a "Mind Meld" session with the kitchen floor last month. Got 8 stitches over her eye. It's been a round of doctors ever since. Thankfully she is okay and there are no lasting effects other than one eyebrow looks weird.

Thanks for understanding during my 1/2 logical rant.


That's good to hear. What can I say - Your Mom's Your Mom! I'm a doctor, in a countryside where people often drive a 'fifty miler' to see me ... so I know 'bout that. :-)

Cheers, Mike.

( edit ) Winternight - you ought retract that 'bending the truth' sentence. We'd all think the more of you if you did.

( edit ) Brian - 'Instead of coming as close as I can to calling you a liar, I'm going to say that you are confused' is actually implying that he is, sotto voce, so likewise I would be similiarly pleased if you would also retract.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

RE: RE: My Dad has

Message 83670 in response to message 83668

Quote:
Quote:
My Dad has surgery tomorrow, so I can relate...

Brian, I hope that everything goes well with your dad's surgery and that he has a speedy recovery. Hope a cold one is waiting for you to relax when its over.

Thanks...

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

Done... As am I... Too

Message 83671 in response to message 83669

Done...

As am I... Too many people have:

...taken the blue pill

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

RE: I took Brian's point as

Message 83672 in response to message 83665

Quote:
I took Brian's point as asking if 2005 ..... 06 ..... 07 .... 08 credits are comparable.

That's a better way of stating it. As it stands, I don't see how they are. The recalibration to some mythical "cross-project equivalence" value mangles the comparison unless the credit rates are comparable between (during) the times that any two given systems are active.

This is essentially an "inside the same project" view of the same phenomenon that is the burr in some people's saddles in regards to higher amounts in external projects (to their projects), thus if this idea is "preposterous" and shunned, why is the other phenomenon accepted as an "inequity that must be rectified at all costs"?

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,550
Credit: 288,905,381
RAC: 87,563

RE: Done... ...taken the

Message 83673 in response to message 83671

Quote:
Done... ...taken the blue pill


Terrific.

Now on the upside Bernd's post explaining the credit movements is helpful. I'll admit to not twigging to that, as of course we would be starting on the longer WU's. Doh! The maxima are at sequence zero for a given frequency, and then 'drape down' sinusoidally to the minima. I guess when a few are collected I'll dust off ( some version of ) RR and start plotting.

I'll admit to waning interest in it's development when it was mooted to have more uniform WU runtimes for this run. But now the cyclical character will be with us for the best part of another year, I'll revisit the Java version ( it needed polishing anyway ).

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Winterknight
Winterknight
Joined: 4 Jun 05
Posts: 1,279
Credit: 336,545,387
RAC: 506,994

RE: Winternight - I took

Quote:
Winternight - I took Brian's point as asking if 2005 ..... 06 ..... 07 .... 08 credits are comparable


That's a difficult one.
One could say yes, if we only consider the stable longer running projects, and try and ignore the users of optimised/power applications, who are actually very few and mainly the people who visit and use these forums.
Most of the gains of using optimised apps is short term. The optimisation is usually incorporated into the next generation of default applications. That has been the case here and on Seti.

My own personal view is that the third party BOINC clients where they attempted to adjust the benchmarks was the real start of the cross project parity troubles. As some users of them either didn't RTFM, or deliberately misused them.

This caused the projects to find other ways to calculate credits and the methods used usually required a crystal ball to get anywhere near accurate. And to the best of my knowledge crystal balls are as common as hens teeth especially in the scientific community.
Plus the project leaders, scientists who hate having to deal with the public and make political decisions, had to deal with the fallout from those members who didn't or wouldn't understand when credits were reduced. And therefore the credit decrease was limited, that may happen here, see my previous. This probably did cause some credit inflation and attempts to rectify again cause strife within the ranks. Didn't some general say the troops were only happy when they were complaining.
There are also those who take a very narrow view of the situation, i.e. the performance of their own choice of system and then complain because A. N. Other made a completely different choice and gets x% greater cr/time.

There is also the subject of open and closed source applications, which I will leave there.

P.S. I hope all goes well for Brian's Dad and Arion's Mom.

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,550
Credit: 288,905,381
RAC: 87,563

RE: Plus the project

Message 83675 in response to message 83674

Quote:
Plus the project leaders, scientists who hate having to deal with the public and make political decisions, had to deal with the fallout from those members who didn't or wouldn't understand when credits were reduced.

Well, I'd say that collectively they vastly prefer their scientific work. I would if I were them. :-)

Quote:
And therefore the credit decrease was limited, that may happen here, see my previous. This probably did cause some credit inflation and attempts to rectify again cause strife within the ranks. Didn't some general say the troops were only happy when they were complaining.

I know that during wargame exercises George S. Patton snuck up on a camp sentry at night, and asked him from what direction he thought the greatest threat would come. Not recognising Patton, the private indicated Patton's tent and said 'There!'. This pleased General Patton no end. :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

RE: RE: Winternight - I

Message 83676 in response to message 83674

Quote:
Quote:
Winternight - I took Brian's point as asking if 2005 ..... 06 ..... 07 .... 08 credits are comparable

That's a difficult one.
One could say yes, if we only consider the stable longer running projects, and try and ignore the users of optimised/power applications

Exit out of the cross-project mode and enter into intra-project mode.

Focus only on this project and the credit granted, not the credit/hour. I'm not talking about cpcs, cpch, cpcd, cpcm, cpcy, cpcD (decade), I'm talking about Credit per Task at Einstein, and no place other than Einstein. No SETI, no Rosetta, no CPDN, no RieselSieve, no QMC, no Cosmology...no other project at all.

To maybe make the comparison simpler, let's say you were contracted to build widgets. Let's say that you were told you'd be paid $1 USD for each widget built. Let's say you built 100 of them. Now let's say your friend is contracted by the same company a few months later to build some widgets. Let's say that they are told they'd be paid $1 USD for each widget. They take the same amount of time as you did to only build 80. Now let's say that Joe Schmoe is contacted a month later and builds 100 widgets for $0.85 USD each. Joe takes the same amount of time as you and your friend did. Next, let's say that Freddy Faster is contacted a month after that and builds 200 widgets in the same amount of time as everyone else, but is paid $0.40 USD each. Finally, let's say that Melanie Median builds 100 widgets, but takes about 20% longer to build them. She is also paid $0.40 per widget

1: Was the compensation the same for all people?
2: If viewed by amount paid, did Freddy do as much work as your friend?
3: If viewed by amount paid, did Melanie do as much work as you did?
4: Did Freddy really do the same amount of work as your friend?
5: Did Melanie really do less work than you did?

* When answering #2 and #3, neither you nor your friend know that Freddy's and Melanie's pay rate per widget was different from your own.

Quote:
There are also those who take a very narrow view of the situation, i.e. the performance of their own choice of system and then complain because A. N. Other made a completely different choice and gets x% greater cr/time.

If this is about the significant difference between the Linux and Windows apps running on the same hardware (dual boot), then how can you possibly advocate cross-project parity? If your statement is about this gripe of mine, bear in mind that it is a gripe of others as well and is something that Bernd sees as a deficiency. If Bernd corrects the issue, even if it is to only within some sort of reasonable difference (within 5%), I'll shut up about it... Tony's research some time ago showed that there was a 20% gap... That is a very substantial penalty simply based on the OS. If we're not going to care about those things and/or minimize the viewpoint of those who see it as an issue, then nobody should be worrying about "cross-project parity" at all...because then it becomes "if it is not impacting me, I don't care".

Thanks for the well wishes for my Dad...

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6,550
Credit: 288,905,381
RAC: 87,563

RE: ...... research some

Message 83677 in response to message 83676

Quote:
...... research some time ago showed that there was a 20% gap... That is a very substantial penalty simply based on the OS.


Recently Gary and I were hypothecating on the new platform registration stats ie. is it possible that surges of registrations were largely due to Windoze -> *.nix conversions on the very same hardware? [ I believe it's a lot easier nowadays to have a machine dual booting to XP/Linux, say ]. Whether this necessarily reflects much credit chasing toward the better app ( OS wise ) on the day is moot, but the tactic potentially exists. Moreover it can be readily steered back across later on, to advantage, if you've set up a dual boot. I'm sure Turing would be grinning from ear to ear ( fit a banana sideways ) over this eventuality as an astounding proof of the equivalence of computing machines.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282,700
RAC: 0

RE: RE: ...... research

Message 83678 in response to message 83677

Quote:
Quote:
...... research some time ago showed that there was a 20% gap... That is a very substantial penalty simply based on the OS.

Recently Gary and I were hypothecating on the new platform registration stats ie. is it possible that surges of registrations were largely due to Windoze -> *.nix conversions on the very same hardware? [ I believe it's a lot easier nowadays to have a machine dual booting to XP/Linux, say ].

I have no idea, but I know that my experiment to prove what is now known to be a fact here (Linux app is faster than Windows app on same hardware) failed when I tried to use a VM. The VM overhead was very significant.

As I don't have the desire to do a dual boot, since I am retaking the class that I dropped last fall (the one with the Linux Zealot for the instructor), I may try running the new apps in a VM again. Class starts in just a couple of weeks. Yes, it is a different instructor this time, although they could be just as much (or more) of a jerk as/than the other guy...

I'm off to bed...

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.