Date: Wed, 23 Feb 2005 18:39:42 -0800 From: "David O'Brien" <obrien@freebsd.org> To: Scott Long <scottl@samsco.org> Cc: freebsd-amd64@freebsd.org Subject: Re: More than 4 GBytes RAM on Socket939 system? Message-ID: <20050224023942.GA39443@dragon.nuxi.com> In-Reply-To: <421CD345.2060305@samsco.org> References: <421C4B1B.5070102@mail.uni-mainz.de> <20050223180142.GB26395@dragon.nuxi.com> <421CC86E.2050509@mail.uni-mainz.de> <421CD345.2060305@samsco.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Feb 23, 2005 at 12:02:29PM -0700, Scott Long wrote: > O. Hartmann wrote: > > > >AMD64 processores only address 4 memory slots and can not address > > > >128MBit memory chips, > > > > > >What is a 128 MBit chip? > > > >I found this info on ASUS homepage for the A8N-SLI and it means (I think) > >a single IC, 128MBit x 8 organized. > > I guess you're implying that the memory controller on the 939 only has > enough address decode and refresh logic to drive DRAM chips of a certain > maximum size? David, can you comment? I don't know enough about the ?weirder? DIMM organizations to comment. I can say that a 939-pin Athlon64 has a dual-channel memory controller. If only one of the dual-channels is populated, one does 64-bit accesses. If both channels are [evenly] populated, one does 128-bit accesses. One can use 2GB non-registered DIMM's in a 939-pin Athlon64 system. 4GB DIMM's don't exist in the desktop(workstation) space yet, so I won't comment on them. From an address-line POV, they will work fine. From an electrical POV maybe there could be issues if manufacturers push the JEDEC spec beyond its limits. -- -- David (obrien@FreeBSD.org)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050224023942.GA39443>