Milkway has knuckled under to David A and his team of Kredit Kops, lowering their credit by half.

Joined the flock of other admin sheep.

That's great news!

The credit thieves have lost once more, and the credit system is hopefully getting fairer

The only disappointment I still have is that SETI is taken as a standard, who itself is unwilling to include good code into it's application and thus inflates the credits artificially with the optimized apps. But you can't have everything.

I hope Travis will include the optimized code soon in his app, so that it really get's down to normal @Milkyway.

yes - another bogus-project decided to measure itself @ the mother of wasted cycles.

I totally agree with that.

Now call me a Kredit Kop or whatever you want; although it's obvious that it is impossible to have a total cross project credit parity, all the combined stats sites would be totally worthless (and I wonder if anyone would still care whether a project is granting more credit than another, if there were no combined stats).

At least MilkyWay is on the right way now... and not closing down like the two other projects in that situation earlier this year... btw. good job by all those "don't care what they do, I just want a million credits for each millisecond cpu time an will leave if they lower credits to 999999 per millisecond" guys, who have destroyed those project...

OK. Youre a Kredit Kop.



No. DA destroyed them. Maybe he should go after all the projects that grossley underpay. If he did I might be a bit less sceptical about his motives.

It'time the "holier than thou, all about the science" Types realised that to some of us it's all about the credits, personally I see the whole thing as nothing more than a game which after 9 years is now fast lossing intrest.

@Saenger


What on earth are you rambling about!

Sure, Milkway divided the granted credits by half, but hundreds of users are using the optimized app and destroy the whole statistic. I would ask any Stats-Site to exclude Milkyway until they get their Sh** together...

Milkyway admins cant get anything right. They even suggest users to use the optimized app - instead of making the optimized app the standard app and change the credit per hours accordingly - I dont care if they grant as much (or less) than SETI. As long as it isnt 20 times the amount like it is now...

My Phenom Quad did 20,000 credits per day - 1200-1500 are normal...

You *can't* make an optimized app the standard app. It won't run on all processors.

You can include them in an app-package and run the right one after asking the puter about itself. Methinks that's the way Einstein did it.
I'm no programmer, but I would guess an IF-THEN_ELSE or something alomng the lines will do the trick.
Perhaps it's even possible to include the processor information known of the puter additional to the OS in the app-distribution on the scheduler to reduce DL-size. That's one standard app (i.e. the algorithm) compiled OS/CPU-specifically.

great argument - then you will stand up and demand to exclude SETI from the stats because optimized app isn't default app!

go, make some noise - so everyone will hear you...

That's a great idea. I support that.
I really don't know why they didn't do it.
But as I don't crunch there any more I can't post anything in their fora

Hmm, whats the difference between standard and opt. app @ SETI? 50% maybe - or a bit more? Who cares?

At Milkyway, its 2000%

Edit: I fully support what Milksop is doing!

it was your argument..



you better kill that guy who taught you maths!

Before on a Q9550 is was getting 260 credits for a 2 hr task . Now I get 7.5 credits for a 4 minute task....2000%????

Please get your facts straight before you post such garbage in BOLD

You appear to have an agenda against Milkyway.

Before I got 260 for a 2.5h WU, it was 102 C/h
Then I got 17 for a 4.5min WU it's 216 c/h.

I should get about 25 - 30 c/h, that's the average for BOINC, the benchmark is 25.

25 c/h to 216 c/h is not 2000%, only a bit above 850%
As I think I hit the ceiling and would get even more per benchmark on an outdated machine 2000% for an old P4 seems not to be unreasonable.

From what I understand from the MW forums, you currently get a maximum of 0.03 credits/second, or a maximum of 108 credits per hour per core. This should be the same for both the optimized app and the standard app, so the only one benefiting from the optimized app is the project, as the work is being done at a very accelerated pace.

From reading Milksop's posts and explanations, the optimized app is compiled without the usual MMX/SSEx optimizations, so it should run on all (normal) windows and linux systems. Heck, from what I understand the project should be able to up the versionnumber and distribute the optimized app as standard.
The next app-version from the project has some changes to the science, but as I understand from their forums, will also include a number of Milksop's improvements.

Sorry, "only" 1600% - cause I was referring to the 20000 I did at Milkyway right after they made the opt. app. publicly available - compared to the 1200 credits that are normal on my computer.

I stopped MW after getting 15,000 credits with the normal app which took a bit less then 18 hours, for the first 10000 credits it took a week...

I the meantime, they divided the granted credits by half, then of course its "only" 800-something percent...

But if you still dont think thats too much... Well, it was just my opinion... Never mind...

Read my post again, and again, and again...



There is some (mental) scope, dont you think?

Granted, its "only" 8times more now...

BOINC-SG,

I agree it is still slightly over the top, but I suspect they will enforce another credit-cut when they release their own new (somewhat optimized) app.
On the other hand, it does balance out my work on "underpaying" projects

just a question, don't take it the wrong way, but are you also this vocal about projects that are rewarding credits well below the arbitrary norm?

With all the drama coming from Cosmo, Riesel Sieve, Ramsey, FreeHal and MilkyWay. . .

. . . I remember why I crunch QMC, Cels, WCG, Rosetta and Einstein.

Recent servers have a way to do that. Not quite easy to use, but possible.

The optimization done by Milksop does not use any extensions like MMX or SSE. It just does some calculations only once, which are done more than 8 billion times by the standard app. So it definitely runs on all processors.

Good question: I'm not THAT vocal, but for example: On the 3rd of April 2007, I asked @ WCG Forums to increase BOINC Credits by decreasing the divider of their own credits.

WCG-BOINC-Credits = WCG-Credits divided by 7, I suggested to decrease divider to 4...

Obviously, its still the same @ WCG. And obviously I'm still crunching WCG - its my #1 project!

Well then that's not an optimized app. It's just written better. "Optimized" means it is designed to use something not common to all processors, such as being optimized to use SSE or 64 bit or whatever.

"Optimised" means "better than the stock stuff", no matter by what means.
Simply compiling it with the right flags for SSE or MMX is optimising as well, but so is real coding.

sorry, but optimized to me does not mean "only for special processors". every code has potential for improvemts without using special features of certain platforms. that's exactly what MILKSOP did. there mostly allways is extra potential when coding (or compiling) for special features - but this is not the case here. project could make it the standard app for win and linux, MAC should just be a cause of someone to take the time and build it.