Re: ECC support

["J. R. Fox" jr_fox@xxxxxxxxxx]

Patricia M Godfrey wrote:

> Jordan is right about parity/ECC support being a function of the chipset,
> but I thought the BX was precisely the one that did NOT support ECC. (ECC
> SIMMS or DIMMS will run in a non-ECC board, they just won't have ECC
> turned on. Are you sure it's running in ECC mode? Check your CMOS.)

Patricia,

I'm quite sure I had ECC running on the Asus P2B-F motherboard I used for
four years. The current motherboard, an Asus P3B-F, a very direct descendant
also utilizing a 440-BX chipset, has a BIOS setting for enabling ECC. It is
turned off at the moment (the default), because I just had this box rebuilt with
mostly new components, and I never got around to finalizing every last BIOS
setting.
(The layout of _this_ Award BIOS, vs. the prior one, is the main apparent point
of difference. It seems to me that the previous BIOS had several more screens,
and quite a few more possible settings one could make.)

The ECC capability was an explicit requirement I had in my list of specifications
for the guy who built it. I'm sure that carries over with the later board, which
is
extremely similar. I wanted to stay with the BX chipset -- one of the best to
ever
come down the pike, and with some advantages over the chipsets on later ASUS
motherboards -- and to keep everything as close as possible, in order to minimize
migration issues in maintaining compatibility for OS/2.

> The thing is, Intel makes chipsets as well as CPUs; and in some cases, Intel
> CPUs would only work on boards that had Intel chipsets. And of course,
> Intel and AMD CPUs and have needed different Mboards since Pentium Pro
> days.

Wasn't the 440-BX an Intel chipset ? I think so . . . which would be consistent
with this proposition.

> Then there was the whole Rambus-DDR mess, which was pretty much MMX
> all over again.

Get used to it. From what I'm hearing lately, DDR is pretty much the new
standard. But the prices have come down for it, at least.

>     Anyway, this mad pursuit of speed is wasted effort for the things that
> most people use their PCs for. My current machine, an AMD K6/2 at 333
> Mhz, is not appreciably slower running WordPerfect or QuattroPro than the
> office Pentium IIIs at about 500 MHz; of course, I have 160 to their 128
> MB of Ram and a cleaner setup. Their online access is faster, but that's
> because my phone lines are subject to some horrible interference. And, of
> course, we've long known that XyWrite DOS runs faster on a 286 than Word
> on a Pentium 4. (Maybe that should be XyW's slogan?)

Well, it depends. What most people run (?) . . . I'm not even sure that
applies,
because Windows and Windows app.s are such turgid bloatware, very resource
intensive and wasting of resources. 512M of RAM is not that much for NT or
later, particularly if you run stuff like PhotoShop. If you run that, or some
other
graphics app.s (can't recall some of the names right now, but there is another
one that makes really excessive hardware demands), compile code, or do stuff
with big databases, I think you would notice a big difference. I just do some
moderately heavy multi-tasking, and I notice the difference. Bootups and shut-
downs of OS/2 or W2K proceed more expeditiously also. If and when I get VPC
(to run Win-32 app.s inside OS/2, instead of having to boot into W2K), I've
been told that the P-III / 850 I now have is just about entry-level: for that
emulation to work best, more horsepower would definitely improve performance.


Jordan