credit per unit dropped a second time

Trog Dog
Trog Dog
Joined: 25 Nov 05
Posts: 191
Credit: 541562
RAC: 0

RE: First we crunched Time

Message 44176 in response to message 44175

Quote:
First we crunched Time X f�U and get Y Credits. This was the default credit level. Then the optimized (beta) client speed up the chrunching up to 20% and the credit per hour went up with this.
Now we are crunching more data in the same time due to this optimized apps. This means more mips/flops per time?
If yes this increase in Credit is justified against all other projects.
If not why the use of optimized apps is helpfull for the project?
The optimized clients are a step ahead in science, but why we are punished with the decrease in credit?
Then could I use the old client just as well, but what sense make the optimized client at all?

G'day Rockhount

Optimised apps help the project because the same work gets done faster. If the project the decreases credits so that you get the same credits/hour as before the optimisation how are you disadvantaged?

The reason an app becomes optimised is that it uses "shortcuts", so instead of a+a+a+a (4 operations) it does a*4 (1 operation). When you look at it like this your computer is actually doing less operations per wu, so it is entitled to less credits per wu.

Yes you are getting less credits per wu, but your credit/hour should still compare to the credit/hour before the optimisation.

The intention is also to make sure that Einstein is not awarding radically more credits/hour than the other BOINC projects. Why? Because if it doesn't it's likely that a credit upping war will erupt between the projects.

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

Thanks a lot Trog Dog, that

Thanks a lot Trog Dog, that was informative. I've been wondering about the nature of "optimized apps" for a while now...
As for the credits, I'm not overly fussed. I mean, come on... everyone else here in Einstein gets just the same credits per WU you do, and besides, it's just a game... I do like a bit of friendly competition but reading some of the message boards I must say that some of you seem to take all this a bit too seriously. I mean, where's the point flaming at someone simply because you suspect he might have gotten more credits than you did, or flaming at the project staff because you think they're not giving you as much credit as they should? To me, that seems downright immature. The people working here are imho doing an awesome job, so why do some people make it more difficult by whining about credits?
I mean, why should one even be here if it was only about "being the best" according to credits? I, for one, certainly wouldn't, because I know full well that with my little Celeron laptop I'll never have a chance to get really good. And even when my new workstation is finally ready to go (and believe me, I'm looking forward to this, I like my gadgets like the next person) I won't be able to compete against people who maybe have 2 or 3 computers for crunching only... And those of you who are playing in that league should remember Bruce Allen letting a whole farm of dual opterons join in the fun ;-) so, in short, there will always be someone who's better than you... if you can't deal with that, why play?
Luckily, there are other things to be enjoyed here... and they are for people like me as well, regardless of how much credit we make... I simply like taking part in interesting science (being something like an astro-nut since I was five), getting to know interesting people from all over the world, the feeling of contributing to something wortwhile... and of course I'm also interested in the organisation and all the technology behind a project like this and especially the programming. And I do like seeing my credits rise because it tells me that I've achieved something. But that's only a small part of it... so just chill out, guys...

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4267
Credit: 244929956
RAC: 16515

Note that adjusting the

Note that adjusting the credit level as a whole (i.e. of all Workunits) doesn't affect the competition within a project at all. It simply doesn't matter if hosts running faster Apps do get more credit (as they do e.g. in the Beta Test phase) or hosts that doesn't (or can't) run them get less credit. It's just getting used to new numbers, and a bit of transition time with some turbulences.

The "overall credit level" is just important to avoid credit inflation between different projects.

We are still watching the "credit level" and will continue to make adjustments as needed.

BM

BM

tahanko
tahanko
Joined: 28 Feb 06
Posts: 15
Credit: 242321
RAC: 0

RE: Thanks a lot Trog Dog,

Message 44179 in response to message 44177

Quote:
Thanks a lot Trog Dog, that was informative. I've been wondering about the nature of "optimized apps" for a while now...
As for the credits, I'm not overly fussed. I mean, come on... everyone else here in Einstein gets just the same credits per WU you do, and besides, it's just a game... I do like a bit of friendly competition but reading some of the message boards I must say that some of you seem to take all this a bit too seriously. I mean, where's the point flaming at someone simply because you suspect he might have gotten more credits than you did, or flaming at the project staff because you think they're not giving you as much credit as they should? To me, that seems downright immature. The people working here are imho doing an awesome job, so why do some people make it more difficult by whining about credits?
I mean, why should one even be here if it was only about "being the best" according to credits? I, for one, certainly wouldn't, because I know full well that with my little Celeron laptop I'll never have a chance to get really good. And even when my new workstation is finally ready to go (and believe me, I'm looking forward to this, I like my gadgets like the next person) I won't be able to compete against people who maybe have 2 or 3 computers for crunching only... And those of you who are playing in that league should remember Bruce Allen letting a whole farm of dual opterons join in the fun ;-) so, in short, there will always be someone who's better than you... if you can't deal with that, why play?
Luckily, there are other things to be enjoyed here... and they are for people like me as well, regardless of how much credit we make... I simply like taking part in interesting science (being something like an astro-nut since I was five), getting to know interesting people from all over the world, the feeling of contributing to something wortwhile... and of course I'm also interested in the organisation and all the technology behind a project like this and especially the programming. And I do like seeing my credits rise because it tells me that I've achieved something. But that's only a small part of it... so just chill out, guys...

this is the same i wanted to say - u have a good attitude

Pooh Bear 27
Pooh Bear 27
Joined: 20 Mar 05
Posts: 1376
Credit: 20312671
RAC: 0

Of all the project boards I

Of all the project boards I read, I feel Einstein does have a better attitude on their boards. Discussions usually turn out well. People have their differing opinions, and some can not be totally changed, but they come to fruition with the best things for the project.

I am glad that the people here are civil. It's an awesome thing to see. There have been some heated arguments on the boards, but usually they are still within reason. I've only seen a couple of messages in the past that were uncalled for.

As a community I think the people on Einstein can go far. Dr. Bruce Allen, Bernd Machenschalk, et. al are doing a good job keeping us pretty well informed. Keeping people on a playing field that is as level as they can make it, and working to improve that all the time. They want to know their outcome as much as we do, and as fast as they can conceivibly make it, yet allow as many machines to crunch as they do.

I applaud all the work done by the team. I applaud the community as a whole. I believe this could be one of the most successful project around, with the people involved.

Scott Brown
Scott Brown
Joined: 9 Feb 05
Posts: 38
Credit: 215235
RAC: 0

RE: Note that adjusting the

Message 44181 in response to message 44178

Quote:
Note that adjusting the credit level as a whole (i.e. of all Workunits) doesn't affect the competition within a project at all.

Ideally, yes. However, this assumes that all hosts within that project are equally benefited by use of an optimized application. This has been historically not the case at E@H with various architectures and OS'es not being able to use optimized applications (including the current distribution of clients--I believe that the Macs are still odd man out???). In reality, users whose credit should not be adjusted because the optimized applications are not available to them are penalized resulting in within project credit issues.

I would also add that such users would have a cross-project deficit. Indeed, to properly correct for cross-project equality, one would need to weight the credit by the balance of users for whom optimized applications are and are not available. Simply reducing the optimized client granted credit proportionally will result in some degree of understated credit by the project as a whole.

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

Yes Scott, I believe this can

Yes Scott, I believe this can be frustrating. But I'm confident the people in charge are aware of the problem and trying to find a solution for it. Some things just take some time and people constantly asking or complaining (and I don't mean you at all, I'm just speaking generally here) don't speed up the process one little bit. As I'm an IRC administrator, server admin and involved in a gameserver firm and a support community myself, I know what it's like to work nights and still get flamed at because users think you're not being fast enough or they don't like your solutions. Just give these guys a rest, I'm quite sure they know what they're doing...
@tahanko: Thanks a lot for saying that, I appreciate it.
@Pooh bear: That's what I think too. I'm running SETI as well and have had a look at some other projects but Einstein sure had the community that appealed to me most- and also interesting science, of course... I'm not that much into biology and I must admit that I was always rather hopeless at math ;-)

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4267
Credit: 244929956
RAC: 16515

RE: RE: Note that

Message 44183 in response to message 44181

Quote:
Quote:
Note that adjusting the credit level as a whole (i.e. of all Workunits) doesn't affect the competition within a project at all.

Ideally, yes. However, this assumes that all hosts within that project are equally benefited by use of an optimized application.

Well, optimized applications indeed do affect the competition, but the credit level doesn't. Optimization means squeezing the last out of a particular architecture(*), and thus it's only this architecture that benefits from it. But if the credit granted for the same work is levelled for all work (and thus all architectures), the competition between them isn't affected.

So you might complain about optimizing the Apps, but not about levelling the credit ;-)

Seriously: For competition within a project it doesn't matter if hosts running faster Apps will get more credit than before or hosts running slower Apps will get less. For people that decide which project to join based on the credit they get, this points in the direction where every project attracts the machines where the best Application exists for, i.e. (ideally) the project that fits them best (also see my post over here).

Edit: (*) That's true only for our present application. Structural and algorithmical optimizations have been incorporated into the analysis code even before the public launch of Einstein@Home. A single nice idea from Akos was general enough that it sped up all code by a factor of two, but that was also about half a year ago. Since then (sloppily speaking), everything is about tweaking and twisting the (assembler-) code so that it runs faster on more CPUs than it runs slower on.

BM

BM

Ziran
Ziran
Joined: 26 Nov 04
Posts: 194
Credit: 356403
RAC: 1626

Bruce wrote here:

Bruce wrote here: http://einsteinathome.org/node/191705&nowrap=true#46005

Quote:

My intention is a simple one: ON THE AVERAGE a host machine running Einstein@Home should get the same number of credits/cpu-hour as a host machine running the other BOINC projects that grant credit.

Here ON THE AVERAGE means averaged across all the hosts that are attached to multiple projects, and averaged across all the projects (suitably weighed by the number of cross-project hosts).

Rationale: this way, people will chose projects based on their scientific and other merits, and likelihood of success and impact, NOT for other reasons such as credit granted.

Corollary: assuming that other BOINC projects do the same, this will tend to make hosts move to the projects that they are best suited for.

Before Akos released his S4 app, the Mac’s were the top dogs on this project. So many farmers increased the Einstein resource share on there Mac’s and decreased them on there other hosts. Then Akos released his S4 app, that changed and they used there windows machines on this project and moved there Mac’s elsewhere. Now all x86 hosts are equally fast, but the old Mac’s didn’t get as much speedup as the x86 platform this time.

Regardless if you are running BOINC because of the science, credits or something else, in the end, for the projects its all about science. One of the goals with BOINC is that it should be easy to share your resources between the projects you wish to participate in. If we take this one step further, one could say that each project only are using there share of the total BOINC resources, then they need it. All BOINC projects are booth partners and competitors. They compete with each other for there share of the BOINC resources, but there also in some way interested in that the total BOINC resources are as large as possible and that these resources are used as efficient as possible.

If your only interested in one project, you will be getting credit in proportion to the usefulness of your contribution. Its in the projects interest to have the best possible application available for each platform. From a BOINC standpoint, its also an interest to use the best suited hosts on each project. Granted credit can therefore be used as a tool to move hosts to a different project that is better suited for the task in hand. By doing this the BOINC system produces more science.

My main host is a Sempron 3000+. I also got an old AMD 1200 (no SSE) that I am considering to start crunching on. I guess the AMD 1200 would take over 3 times as long as my Sempron 3000+ to crunch an Einstein result. On the other hand, it would do a SIMAP result in less then twice the time of my Sempron. Regardless if I am in this for the credit or the science, having the Sempron do Einstein and the AMD1200 do SIMAP would maximize booth.

Then you're really interested in a subject, there is no way to avoid it. You have to read the Manual.

Scott Brown
Scott Brown
Joined: 9 Feb 05
Posts: 38
Credit: 215235
RAC: 0

@Bernd I think you are

@Bernd

I think you are missing my point...credit level adjustments are in response to the speed-up (and resulting credit spike) caused by the introduction of the optimized application. Since this is not a uniform introduction across platforms, an adjustment that brings optimized application capable systems completely back in-line credit-wise will skew the overall project credit award in the opposite direction (the degree of which depends on the proportion of project hosts not able to use the optimized app.).

For example...(assuming the original app. credit has cross-project comparability)

System A Original App Granted Credit = 100
System B Original App Granted Credit = 100

Optimised app. is introduced for system A that is 33% faster resulting in inflation of credit. The project responds by lowering all credit claims to 66% of the original level to compensate.

System A Optimized App Granted Credit = 66 (equals 100% of original)
System B Original App Granted Credit = 66 (equals 66% of original)

Thus, the overall project credit totals are articially low. Furthermore, all system B's that may have been originally on par with system A's now operate at a predefined (and project-specified) disadvantage. Therefore, adjusting the credit does affect within-project competition.

Mind you that this is NOT a complaint. I am simply trying to point out that your original statement is not as simply put as you have done.

@Annika

I am not sure why you interpreted my comment as a complaint, since I didn't actually complain anywhere in my post? I am also uncertain why your occupation has anything to do with my comment unless you mean to imply that I work less (in time or effort) than either you or Bernd (just FYI, a 60 hour week for me means that I have taken a couple of days off to be with the kids)? I also don't recall questioning the knowledge of Bernd or anyone else at E@H?

What I did do was question a statement that Bernd made because I think he was in error when he said it. Oh, and also FYI, I do not now nor have I ever owned a Mac (my systems are windows-based Intels and AMDs).

Comment viewing options

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