Further optimisation of the app is possible. Credit is now determined server side based on the properties of the WU. If Akos or others do produce a more optimised app, using it will give a credit advantage, even without a calibrating client. Because some optimisations have already been included in the official app, the advantage will never be what it was for S4.
Akos and the others have already started Alpha/Beta testing here :]
Further optimisation of the app is possible. Credit is now determined server side based on the properties of the WU. If Akos or others do produce a more optimised app, using it will give a credit advantage, even without a calibrating client. Because some optimisations have already been included in the official app, the advantage will never be what it was for S4.
Akos and the others have already started Alpha/Beta testing here :]
Cool, I knew it !
Quote:
Quote:
If not then it's the same effect as the Enhanced had on the SETI project and
I'm quiting this until until shorter WUs are available.
Its also nice to know that credits now are a recognition of our own individual performance, and are no longer enhanced by (inadvertently) picking the pockets of other members achievements.
Alan, what exactly does that sentence mean? If it means what I think it does, then I strongly disagree with you.
Further optimisation of the app is possible. Credit is now determined server side based on the properties of the WU. If Akos or others do produce a more optimised app, using it will give a credit advantage, even without a calibrating client. Because some optimisations have already been included in the official app, the advantage will never be what it was for S4.
Akos and the others have already started Alpha/Beta testing here :]
Please be aware that akosf has asked us to stop using the optimizes app, due to that it´s not an official app and can have an negative impact on the project please read post "New S5Txxxx.dat patch files - READ ONLY - Only Akos to post"
On my P4 1.7Ghz a long S5 takes >74K seconds for a credit of 175!
Short S5 took >8k for 20 credit.
You get more credit for much less time.
Algy: Not really... your >74K for 175 credits is 8.51 credits per hour, and your >8K for 20 credits is 9.0 credits per hour. If the >8k is really 8400 secs, then the credits per hour are essentially the same. Seems close enough for government work...
On my P4 1.7Ghz a long S5 takes >74K seconds for a credit of 175!
Short S5 took >8k for 20 credit.
You get more credit for much less time.
Algy: Not really... your >74K for 175 credits is 8.51 credits per hour, and your >8K for 20 credits is 9.0 credits per hour. If the >8k is really 8400 secs, then the credits per hour are essentially the same. Seems close enough for government work...
C
C, average for long is 74,800s over six results all gaining 175.78 credits - 8.42/hr.
average for short is 8,315s over twelve results all gaining 20.02 credits - 8.67/hr.
The point I was making was that both are a long way from 13.62/hr that Idefix is getting - AND - that he was getting 177.36, not 175.78.
Maybe I misunderstood the point of changing the credit system.
I thought that it was going to be determined by the server on a consistent rate per hour computed regardless of machine type, - or -
If not that then it would be - like CPDN - a fixed value per 'trickle' - in Einstein terms a fixed value per WU regardless of machine type or processing performed.
I am not overly concerned about it - just confused about where the consistency lies.
Maybe I misunderstood the point of changing the credit system.
I thought that it was going to be determined by the server on a consistent rate per hour computed regardless of machine type, - or -
If not that then it would be - like CPDN - a fixed value per 'trickle' - in Einstein terms a fixed value per WU regardless of machine type or processing performed.
...the credit system was changed that you get the same credits for the same WUs - independent of the application version used.
BUT: if you have a faster CPU, your system will need less seconds per WU and therefore you get more credits/h
My workunits were a little bit bigger than yours. Therefore they got a little bit more credit than yours. At the moment I'm crunching smaller workunits which will get 170.83 credits.
As long as you are crunching the same workunits you will get the same credit.
...the credit system was changed that you get the same credits for the same WUs - independent of the application version used.
BUT: if you have a faster CPU, your system will need less seconds per WU and therefore you get more credits/h
Udo
And AMD's seem to be doing these a lot faster, all of my systems are Intel (CRY).
RE: Further optimisation of
)
Akos and the others have already started Alpha/Beta testing here :]
RE: RE: Further
)
Cool, I knew it !
RE: Hi,RE: The question
)
Nothing like consistency, is there?
On my P4 1.7Ghz a long S5 takes >74K seconds for a credit of 175!
Short S5 took >8k for 20 credit.
You get more credit for much less time.
RE: Its also nice to know
)
Alan, what exactly does that sentence mean? If it means what I think it does, then I strongly disagree with you.
FD.
RE: RE: Further
)
Please be aware that akosf has asked us to stop using the optimizes app, due to that it´s not an official app and can have an negative impact on the project please read post "New S5Txxxx.dat patch files - READ ONLY - Only Akos to post"
RE: Nothing like
)
Algy: Not really... your >74K for 175 credits is 8.51 credits per hour, and your >8K for 20 credits is 9.0 credits per hour. If the >8k is really 8400 secs, then the credits per hour are essentially the same. Seems close enough for government work...
C
Join Team MacNN
RE: RE: Nothing like
)
C, average for long is 74,800s over six results all gaining 175.78 credits - 8.42/hr.
average for short is 8,315s over twelve results all gaining 20.02 credits - 8.67/hr.
The point I was making was that both are a long way from 13.62/hr that Idefix is getting - AND - that he was getting 177.36, not 175.78.
Maybe I misunderstood the point of changing the credit system.
I thought that it was going to be determined by the server on a consistent rate per hour computed regardless of machine type, - or -
If not that then it would be - like CPDN - a fixed value per 'trickle' - in Einstein terms a fixed value per WU regardless of machine type or processing performed.
I am not overly concerned about it - just confused about where the consistency lies.
Cheers.
RE: Maybe I misunderstood
)
...the credit system was changed that you get the same credits for the same WUs - independent of the application version used.
BUT: if you have a faster CPU, your system will need less seconds per WU and therefore you get more credits/h
Udo
Udo
Hi, RE: he was getting
)
Hi,
My workunits were a little bit bigger than yours. Therefore they got a little bit more credit than yours. At the moment I'm crunching smaller workunits which will get 170.83 credits.
As long as you are crunching the same workunits you will get the same credit.
Regards,
Carsten
RE: ...the credit system
)
And AMD's seem to be doing these a lot faster, all of my systems are Intel (CRY).
Try the Pizza@Home project, good crunching.