Hi all, On 6/23/05, Dmitri Pogosyan <[EMAIL PROTECTED]> wrote: > I would be very interested to check if somebody could benchmark if this > does not hurt the memory access performance. >
Well, the application I work with uses some memory (3.1Gigs of real in use memory, 3.7 of effective "allocated memory) and I run two of those on my dual-opteron with 8gigs of ram. This app runs for 3 days, and each simulation lasts the same time has before I changed the bios settings to make that software mem hole.. So.. at least on my case.. I don't see performance issues with this enable.. I can do some "stream" runs to check how it does... in this config.. Best regards, > Celestica is quite adamant that it will (but it is out of opteron business, > so its opinion may be old) > > > > Miguel Filipe posted <[EMAIL PROTECTED]>, > > excerpted below, on Wed, 22 Jun 2005 20:50:55 +0100: > > > > > Just for the record, I'm not having the problem anymore. > > > The BIOS had a option to go around the memory hole, by software, or > > > by hardware (only for rev ? cpus).. I chose by software.. and worked.. > > > now I have: > > > > > > [EMAIL PROTECTED] ~> free -m > > > total used free shared buffers cached > > > Mem: 8037 8000 36 0 0 3625 > > > -/+ buffers/cache: 4375 3661 > > > Swap: 11452 0 11452 > > > > Very good to read, not only for you, but because I'm looking to upgrade, > > but haven't yet, and now I know it'll work properly when I do. > > > > -- > > Duncan - List replies preferred. No HTML msgs. > > "Every nonfree program has a lord, a master -- > > and if you use the program, he is your master." Richard Stallman in > > http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html > > > > -- > > [email protected] mailing list > > > > -- > Dmitri Pogosyan Department of Physics > Associate Professor University of Alberta > tel 1-780-492-2150 412 Avadh Bhatia Physics Labs > fax 1-780-492-0714 Edmonton, AB, T6G 2J1, CANADA > > > -- > [email protected] mailing list > > -- Miguel Sousa Filipe -- [email protected] mailing list
