Annika, is the slowdown dramatic for the Clawhammer? Or just a bit
Hi, I'd really say between the two. The first two WUs indicate a slowdown of roughly 10% (10.8 to be exact), which is not really THAT dramatic (at least not compared to what we've seen with some other hosts with the unpatched Windows app) but might be noticeable in the course of a few days of crunching...
Annika, is the slowdown dramatic for the Clawhammer? Or just a bit
Hi, I'd really say between the two. The first two WUs indicate a slowdown of roughly 10% (10.8 to be exact), which is not really THAT dramatic (at least not compared to what we've seen with some other hosts with the unpatched Windows app) but might be noticeable in the course of a few days of crunching...
CU Annika
Hi!
(And I thought I was the only one up that late, or should I say early :-) )
I guess 10% is more than what could be explained by different workunits (I've seen maybe 5% differences in CCH for different WUs on the same machine). Hopefully there will be a Win version soon that is at least as fast as the Linux version on all platforms (the Linux app uses only one codepath for all CPUs and links the math library dynamically).
(I spent almost a day figuring out what was going wrong there before I noticed that the information at http://www.sandpile.org/ia32/cpuid.htm is plain wrong: in fact all K8 CPUs report the same family and extended family code (15 and 0, resp.)).
P.S.: In case you ask yourself : "What is that guy talking about, the information is correct!" : The page on sandpile.org was fixed after I dropped the author an e-mail.
Yeah, I occasionally tend to live in the universal "geek time zone" too :-D As for different WU sizes contributing to different runtimes, I'd like to add that both WUs were of an identical size and frequency, so the difference should have been minimal (at least that's what I would think).
RE: Annika, is the slowdown
)
Hi, I'd really say between the two. The first two WUs indicate a slowdown of roughly 10% (10.8 to be exact), which is not really THAT dramatic (at least not compared to what we've seen with some other hosts with the unpatched Windows app) but might be noticeable in the course of a few days of crunching...
CU Annika
RE: RE: Annika, is the
)
Hi!
(And I thought I was the only one up that late, or should I say early :-) )
I guess 10% is more than what could be explained by different workunits (I've seen maybe 5% differences in CCH for different WUs on the same machine). Hopefully there will be a Win version soon that is at least as fast as the Linux version on all platforms (the Linux app uses only one codepath for all CPUs and links the math library dynamically).
Good night/morning
BRM
RE: (I spent almost a day
)
P.S.: In case you ask yourself : "What is that guy talking about, the information is correct!" : The page on sandpile.org was fixed after I dropped the author an e-mail.
Yeah, I occasionally tend to
)
Yeah, I occasionally tend to live in the universal "geek time zone" too :-D As for different WU sizes contributing to different runtimes, I'd like to add that both WUs were of an identical size and frequency, so the difference should have been minimal (at least that's what I would think).
well, you should NOT try to
)
well, you should NOT try to change the new app (4.33) 'cause its finishing the WU with error and downloading itself again : (