It does disqualify machines from work - the ones that just take too long
(several hours to several days) to do the benchmarks every week.
This is the very first that I have heard of an opt-out option. Who gets to
opt out, the user or the project?
jm7
"Paul D. Buck"
<p.d.b...@comcast
.net> To
Sent by: [email protected]
<boinc_dev-bounce cc
[email protected] BOINC Developers Mailing List
u> <[email protected]>
Subject
Re: [boinc_dev] [boinc_alpha] Card
09/25/2009 11:03 Gflops in BOINC 6.10
AM
On Sep 25, 2009, at 5:13 AM, [email protected] wrote:
> Not all projects are appropriate for all machines. The change you are
> proposing would mean that a large number of machines would be cut
> off from
> ALL projects, not just some projects. I have never said that slow
> machines
> would be able to do all projects.
If you would read what I am writing, not what you think I am writing,
what I am proposing would not do anything of the sort. Yes I would be
using real work, and yes it would take more time than the current
benchmarks. But in no way would it disqualify machines from
participating. Unless you mean that people would opt out of BOINC
because you would champion the idea that testing machines and
calibrating the system is somehow a communist plot ....
But the frequency of the test can be adjusted and, though I would not
do this because I think you are throwing up another straw-man issue
that does not really exist today, an opt-out option could be
included. I have said these things repeatedly ... and you have
ignored it repeatedly.
Like I said, I do find it funny that you are so concerned about
efficiencies and "waste" here with this proposal (which could solve
some real problems) and not at all in other places in the BOINC
architecture.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.
