Date: Sat, 24 Aug 1996 12:45:45 -0400 (EDT) From: Brian Campbell <brianc@ottawa.net> To: avalon@coombs.anu.edu.au (Darren Reed), freebsd-hackers@FreeBSD.org (FreeBSD Hackers) Subject: Re: Triton II chipsets Message-ID: <199608241645.MAA00175@ottawa.net> In-Reply-To: <199608240902.FAA07902@poboxer.pobox.com> from Darren Reed at "Aug 24, 96 07:02:44 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> In some mail from Brian Campbell, sie said: > > I suspect the HX chipset will support the same configuration registers > > as the VX, and FX, but I haven't got one to try it with. > > FX is Triton, HX is Triton II, isn't VX Triton III ? In the early Intel literature, I believe both were referred to as Triton II family chipsets. I believe Intel is no longer referring to them as triton chipsets anymore (some trademark restriction?) > and is HX faster then VX, if so, why does VX exist ? Well, I've heard the rumor before, but have seen nothing to substantiate it yet ;-) In terms of memory bandwidth, which may not be important to some, my machine is leading the pack in the stream benchmark. It's worth mentioning that that "pack" includes a PentiumPro and Alpha system as well. There was also guy from sandia labs that posted to some of the pc hardware newsgroups looking for people with various systems, notably sdram systems, to run a "lattice gas flow simulations" benchmark. I ran it and sent him the results. Again, my system beat all the others he tested including one or more P6's. I also get 4.5-5M/s on my IDE drive, and 3M/s on an old SCSI drive. I know the FIFO depths on the HX are deeper than the VX, but I haven't seen anything, yet, to illustrate that that's a problem. The VX is also limited to 128M, but as I've only got 32M, 128M seems a long way away ;-) I've heard nasty rumors about how UMA, which only the VX supports will impair performance, but haven't actually heard of anyone using that feature. I'll also note the VX doesn't support SMP, which the HX does, but I've got only one processor so that's not a problem for me ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199608241645.MAA00175>