I'm not sure how is this 25% decrease. True I get 25% less credit per unit, but it takes me twice as long to complete a work unit than it did for S5R3 units with optimized client.
The same for me here. The credits per S5R4 unit are only slightly reduced; however, each unit takes twice the time to process. I suppose the (Windows/x86) 6.04 application has not been properly compiled - I am comparing the run times to the power application I used for S5R3. Thus, at the moment, I am earning less than half credits per processor hour for S5R4 on 6.04 as compared to S5R3 on power application.
I must confess that I don't really understand the need for "cross project parity" which can never be achieved anyway. I also don't understand the need for intra-project parity either for that matter.
Honestly, the only intra-project parity thing I am really concerned about is the application performance. My reason for bringing the cause-and-effect examples up is as a direct counter-point to a key component of the cross-project quest. One of the justifications for the cross-project effort is that they want to be able to have faith that 1 million credits in SETI reflects 1 Million credits in Rosetta, which in turn reflects 1 million credits in WCG, which in turn reflects 1 million credits in BURP, etc, etc, etc... My hope is that in showing them how distortions are made with each adjustment that one day they will see that the method they are using to get to the "parity" that they so strongly believe in actually is generating disparity in its' wake...
There has to be a better way. This is my main point. I'm not "against them", I'm just against their current methodology.
Quote:
Once you get rid of the myth of inter-project parity and each project has a "floating currency" that people put time and effort into judging the value of, isn't this more likely to pique the interest of new converts and restore the sense of fun to those who are getting turned off by the artificial, forced adjustments that seem to be ever more frequently hitting us?
People use the argument that if project X pays twice the "going rate", a whole bunch of crunchers will flock to project X to the detriment of other projects. There might be some truth in that if you can get enough people to believe in the "uniform credit value" myth. If you stop promoting the myth and actually promote the reverse, wont people gravitate towards projects that interest them or which have a "quality currency" rather than an inflated one? Won't projects be more able to attract participants on the basis of quality rather than quantity?
All actual evidence indicates that the number of people selecting projects based on the amount of credit awarded vs. another project are in the minority of the total BOINC population. This "monster" does exist, but it is made out to be much more/larger than what it really is. There is a high level of FUD that comes along with this idea that people are selecting based on credit.
Someone pointed out to me that they think that when people like Trux came out with their "calibrating" BOINC client that adjusted things having to do with the benchmark, that triggered the reaction from David Anderson. I think that's probably correct. Unfortunately, BOINC is open-source. He picked a credit scheme that relied on values that could be adjusted in the software that the participating public would have access to.
David, and the pushers of the cross-project parity schemes of today, need to realize that the mechanism is of poor base design and needs to be abandoned and a new plan invented...
I hope it is realised that the credit concept has already morphed vastly beyond original intention/definition - not so much by a change in design, but by progressive and ever widening usage. Whom-so-ever and what-so-ever was the very first awarded instance - the first claimed/granted credits for WU number 1 on host X - has now been replicated ( as an 'instance of awarded credits' ) a ~ Gazillion times. This seems subtle, as it is a meta-concept - a concept about a concept.
I'm not looking at the adjustment(s) to credit values/rates directly as an issue. I'm examining the utility of the idea of credit awards. Each reward instance does not go to some bland/identical/uniform bit-bucket. It goes to a thinking & feeling person who will respond internally in some fashion to the 'gift'. As time passes some representation of that response may be divulged or returned. Like a rain shower, drop by drop, an accumulated version of what credit 'means' emerges. With variants. As it snowed here last nite, that suggests a hopefully useful analogy :
How many flakes did you get? Are they the same as mine? Gee, that's a lovely pile! Keep at it, your pile's getting bigger. Look, I have a one million flake stack right here! I heard there's a bloke over yonder who has a snow machine! Aren't they white? What are you doing out in the cold scooping up bloody snow flakes? Get inside you'll catch pneumonia. Stop throwing them at your sister! Ah, snow these days just isn't like the quality stuff we used to get! The boffins reckon that the first snow arrived on a comet. Come off it, you get your snow from who?! Is that car bogged in that drift? I hope he has Snow Insurance. A bloke in the pub told my mate that his brother's budgies' boss at work said that The Government has rigged the weather. My snowman looks better than yours. Should it have three ears? Here, I've sketched the hillsides covered in snow. What a beautiful view. Do you watch the Snow Channel? I hear there's more snow than last year, but next year there'll be less. I have painted my snow pile. Perhaps there'll be less hereabouts and more over yonder. Will we move away? I get my snow chains from Bunnings. Dad, what does Global Snowing mean? ...... :-)
So in the end snow really means what people mean/think/feel/use/mis-use/perceive it to be. And now is a long time after the very first flake fell.
We are actually talking about people and how credit, originally ( probably ) a proxy measure of thanks for contributing, now means a blizzard of other things. :-)
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
As it snowed here last nite, that suggests a hopefully useful analogy :
I love the way you put it, so as someone who has lived in a warm climate that got to see snow for a very short time all I can say is I LOVE SNOW and I want MORE!!!!!!
As it snowed here last nite, that suggests a hopefully useful analogy :
I love the way you put it, so as someone who has lived in a warm climate that got to see snow for a very short time all I can say is I LOVE SNOW and I want MORE!!!!!!
Did you say that so and so has a snow blower? :-)
Come this December and January, you're welcome to stop by and collect as much as you want off my driveway and sidewalks. :)
as someone who has lived in a warm climate that got to see snow for a very short time all I can say is I LOVE SNOW and I want MORE!!!!!!
I know how you feel. Apart from snow on our local mountains tops, I have to drive about 3 hours to actually get to snow! Actually it would be about 2 hours but the increased traffic caused by other 'snow hunters' makes it longer!
All actual evidence indicates that the number of people selecting projects based on the amount of credit awarded vs. another project are in the minority of the total BOINC population. This "monster" does exist, but it is made out to be much more/larger than what it really is. There is a high level of FUD that comes along with this idea that people are selecting based on credit.
There are actually not enough people that select what projects there hosts should work on based on credits. If you have a PS3 then you would do more science and get more credit if you attach it to a project that uses all the special processors. Since we got an SSE application hear (and an SSE2 application for Linux) one could argue that more science would be done (in the BOINC community) if all none SSE capable hosts would move to a project that doesn’t have an application that uses SSE. This also raises the question if a project should try to convince participants to move hosts to an other project that would make better use of them, or if they only should look after there own project and try to get as many hosts as possible, no matter how little science some of those hosts contributes.
Then you're really interested in a subject, there is no way to avoid it. You have to read the Manual.
This also raises the question if a project should try to convince participants to move hosts to an other project that would make better use of them, or if they only should look after there own project and try to get as many hosts as possible, no matter how little science some of those hosts contributes.
Not trying to contradict here - just a thought to ponder:
This is basically what Boinc does. Lots of hosts doing little science, all adding up to big science. Somewhere in all the credit discussions in various projects the idea seems to be creeping in that only the better hosts should crunch and not the weaker ones. If a weaker one can return a result in time, then surely its work is of value to the science?
And to repeat myself somewhat, Boinc is made up of lots of weaker hosts adding up to a very powerful computer. Even the best desktop is classified as a 'weaker' one in that last sentence.
This also raises the question if a project should try to convince participants to move hosts to an other project that would make better use of them, or if they only should look after there own project and try to get as many hosts as possible, no matter how little science some of those hosts contributes.
Not trying to contradict here - just a thought to ponder:
This is basically what Boinc does. Lots of hosts doing little science, all adding up to big science. Somewhere in all the credit discussions in various projects the idea seems to be creeping in that only the better hosts should crunch and not the weaker ones. If a weaker one can return a result in time, then surely its work is of value to the science?
And to repeat myself somewhat, Boinc is made up of lots of weaker hosts adding up to a very powerful computer. Even the best desktop is classified as a 'weaker' one in that last sentence.
Now to go check on some of my weaker hosts...
Rod
Define 'weaker host', if you compare RAC of a core2 quad to a P4, you might say a P4 is a weak host, until you look at the overall figures and realise there is still ten times as many P4's as there are core2's. BOINC cpu breakdown
Define 'weaker host', if you compare RAC of a core2 quad to a P4, you might say a P4 is a weak host, until you look at the overall figures and realise there is still ten times as many P4's as there are core2's.
I hope I didn't give the impression that I'm against the 'weaker hosts'. Define 'weaker host' is exactly what bothers me. I've seen postings where the poster apologizes for a P4 3GHz, and others where it's suggested that one shouldn't bother with anything less. But as your link so nicely shows, they are certainly the majority!
Just to show that I really do want slower hosts running:
How about a Pentium II 400MHz?
Too fast? Ok, I also have a Pentium 133MHz crawling on Boinc - unfortunately 24MB RAM is not enough for Einstein.
Ok, let me go blush quietly in a corner somewhere!
Rod
RE: I'm not sure how is
)
The same for me here. The credits per S5R4 unit are only slightly reduced; however, each unit takes twice the time to process. I suppose the (Windows/x86) 6.04 application has not been properly compiled - I am comparing the run times to the power application I used for S5R3. Thus, at the moment, I am earning less than half credits per processor hour for S5R4 on 6.04 as compared to S5R3 on power application.
Best, Igor
RE: I must confess that I
)
Honestly, the only intra-project parity thing I am really concerned about is the application performance. My reason for bringing the cause-and-effect examples up is as a direct counter-point to a key component of the cross-project quest. One of the justifications for the cross-project effort is that they want to be able to have faith that 1 million credits in SETI reflects 1 Million credits in Rosetta, which in turn reflects 1 million credits in WCG, which in turn reflects 1 million credits in BURP, etc, etc, etc... My hope is that in showing them how distortions are made with each adjustment that one day they will see that the method they are using to get to the "parity" that they so strongly believe in actually is generating disparity in its' wake...
There has to be a better way. This is my main point. I'm not "against them", I'm just against their current methodology.
All actual evidence indicates that the number of people selecting projects based on the amount of credit awarded vs. another project are in the minority of the total BOINC population. This "monster" does exist, but it is made out to be much more/larger than what it really is. There is a high level of FUD that comes along with this idea that people are selecting based on credit.
Someone pointed out to me that they think that when people like Trux came out with their "calibrating" BOINC client that adjusted things having to do with the benchmark, that triggered the reaction from David Anderson. I think that's probably correct. Unfortunately, BOINC is open-source. He picked a credit scheme that relied on values that could be adjusted in the software that the participating public would have access to.
David, and the pushers of the cross-project parity schemes of today, need to realize that the mechanism is of poor base design and needs to be abandoned and a new plan invented...
I hope it is realised that
)
I hope it is realised that the credit concept has already morphed vastly beyond original intention/definition - not so much by a change in design, but by progressive and ever widening usage. Whom-so-ever and what-so-ever was the very first awarded instance - the first claimed/granted credits for WU number 1 on host X - has now been replicated ( as an 'instance of awarded credits' ) a ~ Gazillion times. This seems subtle, as it is a meta-concept - a concept about a concept.
I'm not looking at the adjustment(s) to credit values/rates directly as an issue. I'm examining the utility of the idea of credit awards. Each reward instance does not go to some bland/identical/uniform bit-bucket. It goes to a thinking & feeling person who will respond internally in some fashion to the 'gift'. As time passes some representation of that response may be divulged or returned. Like a rain shower, drop by drop, an accumulated version of what credit 'means' emerges. With variants. As it snowed here last nite, that suggests a hopefully useful analogy :
How many flakes did you get? Are they the same as mine? Gee, that's a lovely pile! Keep at it, your pile's getting bigger. Look, I have a one million flake stack right here! I heard there's a bloke over yonder who has a snow machine! Aren't they white? What are you doing out in the cold scooping up bloody snow flakes? Get inside you'll catch pneumonia. Stop throwing them at your sister! Ah, snow these days just isn't like the quality stuff we used to get! The boffins reckon that the first snow arrived on a comet. Come off it, you get your snow from who?! Is that car bogged in that drift? I hope he has Snow Insurance. A bloke in the pub told my mate that his brother's budgies' boss at work said that The Government has rigged the weather. My snowman looks better than yours. Should it have three ears? Here, I've sketched the hillsides covered in snow. What a beautiful view. Do you watch the Snow Channel? I hear there's more snow than last year, but next year there'll be less. I have painted my snow pile. Perhaps there'll be less hereabouts and more over yonder. Will we move away? I get my snow chains from Bunnings. Dad, what does Global Snowing mean? ...... :-)
So in the end snow really means what people mean/think/feel/use/mis-use/perceive it to be. And now is a long time after the very first flake fell.
We are actually talking about people and how credit, originally ( probably ) a proxy measure of thanks for contributing, now means a blizzard of other things. :-)
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
RE: As it snowed here last
)
I love the way you put it, so as someone who has lived in a warm climate that got to see snow for a very short time all I can say is I LOVE SNOW and I want MORE!!!!!!
Did you say that so and so has a snow blower? :-)
RE: RE: As it snowed
)
Come this December and January, you're welcome to stop by and collect as much as you want off my driveway and sidewalks. :)
Seti Classic Final Total: 11446 WU.
RE: as someone who has
)
I know how you feel. Apart from snow on our local mountains tops, I have to drive about 3 hours to actually get to snow! Actually it would be about 2 hours but the increased traffic caused by other 'snow hunters' makes it longer!
Cheers
Rod
RE: All actual evidence
)
There are actually not enough people that select what projects there hosts should work on based on credits. If you have a PS3 then you would do more science and get more credit if you attach it to a project that uses all the special processors. Since we got an SSE application hear (and an SSE2 application for Linux) one could argue that more science would be done (in the BOINC community) if all none SSE capable hosts would move to a project that doesn’t have an application that uses SSE. This also raises the question if a project should try to convince participants to move hosts to an other project that would make better use of them, or if they only should look after there own project and try to get as many hosts as possible, no matter how little science some of those hosts contributes.
Then you're really interested in a subject, there is no way to avoid it. You have to read the Manual.
RE: This also raises the
)
Not trying to contradict here - just a thought to ponder:
This is basically what Boinc does. Lots of hosts doing little science, all adding up to big science. Somewhere in all the credit discussions in various projects the idea seems to be creeping in that only the better hosts should crunch and not the weaker ones. If a weaker one can return a result in time, then surely its work is of value to the science?
And to repeat myself somewhat, Boinc is made up of lots of weaker hosts adding up to a very powerful computer. Even the best desktop is classified as a 'weaker' one in that last sentence.
Now to go check on some of my weaker hosts...
Rod
RE: RE: This also raises
)
Define 'weaker host', if you compare RAC of a core2 quad to a P4, you might say a P4 is a weak host, until you look at the overall figures and realise there is still ten times as many P4's as there are core2's. BOINC cpu breakdown
RE: Define 'weaker host',
)
I hope I didn't give the impression that I'm against the 'weaker hosts'. Define 'weaker host' is exactly what bothers me. I've seen postings where the poster apologizes for a P4 3GHz, and others where it's suggested that one shouldn't bother with anything less. But as your link so nicely shows, they are certainly the majority!
Just to show that I really do want slower hosts running:
How about a Pentium II 400MHz?
Too fast? Ok, I also have a Pentium 133MHz crawling on Boinc - unfortunately 24MB RAM is not enough for Einstein.
Ok, let me go blush quietly in a corner somewhere!
Rod