FAH Version 5 soon
Missileman
Orlando, Florida Icrontian
Just got this from one of the other forums and thought I would pass it on to the TEAM .
Quote from Bruce @folding community forums.
You should avoid -forceSSE (it's being removed in the next version) and anything you don't understand.
There will be no v4.1.
Version 5.0 is currently in beta test and has made good progress. Any firm date is pure speculation, but it might be as soon as next week or it might be several weeks if a serious bug or two are discovered. (We all want it released - - and we all want all the bugs found and fixed before it is released. If those two are in conflict, it really should be delayed.
Additionally, I'd like to correct some erroneous earlier posts in these forums.
-advmethods has no effect whatsoever any longer as to whether a tinker or a gromac is next downloaded. It does alert Stanford that you are prepared to test beta WU's and beta cores, when they are on offer. Its use will most likely do nothing. Some time ago when gromacs and cores 78,79 were in beta, it did affect the chance of getting a gromac or a tinker, but no longer.
-forcesse has also largely become redundant. There was an issue which caused amd cpus to use 3DNow! instead of the preferred SSE, but that was resolved between Pande Group and AMD some time ago.
The FAH4 cores all automatically enable SSE for both AMD and P4 now and apparently have done so for some time.
-forcesse does have one limited use. If you have an improper termination of a WU in progress, the next time you restart that WU standard loops rather than
optimised (SSE) processing will be used. In that circumstance, -forcesse will be beneficial on a one off basis, presuming you have removed or corrected tha cause of the improper termination.
FYI
Quote from Bruce @folding community forums.
You should avoid -forceSSE (it's being removed in the next version) and anything you don't understand.
There will be no v4.1.
Version 5.0 is currently in beta test and has made good progress. Any firm date is pure speculation, but it might be as soon as next week or it might be several weeks if a serious bug or two are discovered. (We all want it released - - and we all want all the bugs found and fixed before it is released. If those two are in conflict, it really should be delayed.
Additionally, I'd like to correct some erroneous earlier posts in these forums.
-advmethods has no effect whatsoever any longer as to whether a tinker or a gromac is next downloaded. It does alert Stanford that you are prepared to test beta WU's and beta cores, when they are on offer. Its use will most likely do nothing. Some time ago when gromacs and cores 78,79 were in beta, it did affect the chance of getting a gromac or a tinker, but no longer.
-forcesse has also largely become redundant. There was an issue which caused amd cpus to use 3DNow! instead of the preferred SSE, but that was resolved between Pande Group and AMD some time ago.
The FAH4 cores all automatically enable SSE for both AMD and P4 now and apparently have done so for some time.
-forcesse does have one limited use. If you have an improper termination of a WU in progress, the next time you restart that WU standard loops rather than
optimised (SSE) processing will be used. In that circumstance, -forcesse will be beneficial on a one off basis, presuming you have removed or corrected tha cause of the improper termination.
FYI
0
Comments