First, congratulations and thanks to the Einstein team for quickly responding to the on-going problem with validating results from Unix and Windows systems. I had considered switching projects, since it is clear that both results cannot be valid and some of us were wasting our computer time.
Based on Bernd's postings last night, this is recognized by the Einstein team and a solution is being worked on, including the work necessary to purge the results data base of incorrect calculations. Again, congratulations and thanks for your professionalism.
I will temporarily switch to another project until this has been resolved, not in anger or vindictiveness, but simply to quit wasting time on potentially useless calculations. I will be checking the Einstein web site and forums for updated software or clarification that the Windows/AMD results are scientifically correct.
Please let me suggest that the project team prominently post notes on its progress on the two issues of importance to volumteers -- what computer/OS combination produces valid results with the current software and when a new release is available.
Sincere thanks,
Ron Subler
Copyright © 2024 Einstein@Home. All rights reserved.
Policy suggestion
)
I suggest to grant credits for failed calculations too.
They should be marked as beta-credits in opposition to succredits.
This way no CPU-time gets wasted.
Unasked beta-testers stay.
The developer of applications get a reply by the used total of beta-credits.
They can even start a contest among developers.
There is little efford to start the distinction between beta-credits and succredits.
Both count to the credits.
Just install a field in database for beta-credits along with a field for failed application and let it be filled out by the granting mechanism.
Good luck!
hans_lion