[Top] [All Lists]

Re: ISA audio in Magnum 4000pc,possibly?

To: Eric Jorgensen <>,
Subject: Re: ISA audio in Magnum 4000pc,possibly?
Date: Wed, 12 Aug 1998 16:06:25 +0200
In-reply-to: <>; from Eric Jorgensen on Tue, Aug 11, 1998 at 12:44:19PM -0600
References: <> <> <> <> <>
On Tue, Aug 11, 1998 at 12:44:19PM -0600, Eric Jorgensen wrote:

> > >     On a completely unrelated note - out of curiosity I pulled the 
> > > heatsink
> > > off the cpu in my system - Never seen a mips 4k processor before. Anyway, 
> > > I
> > > was surprised to find the word "PROTOTYPE" emblazoned acrost the chip. Is
> > > this common? Or do I have something really cool/funky here? if anybody
> > > wants to see, I have a digital camera here.
> >
> > I've never seen a processor like that before.  The only thing I bet on is
> > that this processor is buggy like shit.
>     Well, it has what I believe is a Fujitsu logo (has that curly S thing on 
> it
> that was used as an F centuries ago, if you look at old documents like the
> US constitution) and it identifies itself as (unless I'm remembering wrong) a
> 76R4000PC-50G - there is also a letter G hand-written in pencil on the corner 
> of
> the ceramic, and two other places on the ceramic where it looks like things 
> have
> been scribbled over.

No idea what the G means.  The 50 btw is the external clock feed into the CPU
rsp. the maximum bus clock.  The internal clock rate is twice as fast, so this
is effectivly a 100Mhz machine.  For later CPU versions MIPS started measuring
the CPU just as others do in order to get a 100% bloat factor ...

>     Possibly this is an early low-power version?
>     The thing is, there's no way this was a pre-release prototype of the
> original magnum, because the original magnum was Microsoft-designed box
> code-named Jazz. I actually have the whole story from the horses mouth 
> (Atleast,
> a guy who was at mips during the fiasco) of how Mips was falling behind on 
> their
> own design, and ended up buying the design that Microsoft had built to Dave
> Cutler's specs. So, since the board has Mips logos all over it, it's clearly 
> not
> first-run, tho it could be something from the transitional period after MIPS
> bought the design from microsoft.

I've got one of the first machines, the inside of the Magnum's case has over
dozen signatures of the Magnum's development team on it.  I don't recall this
board to be any different from later versions I've seen.

>     On the other hand, only the processor says "prototoype" on it.
>     I don't know much about where this box came from. I bought it as a surplus
> item from a local university. Word has it that a professor retired and nobody
> wanted his computers, so they stripped them for memory and storage devices, 
> kept
> the monitors, and surplused the rest. I bought my Magnum for $5 USD, including
> the high-rez frame buffer and sound card. I also bought his old fileserver -
> DELL 4066/XE - a 486 but it's the size of a doghouse. Cost me $10.

Boy, you mu$t be rich :-)

>     I've actually never successfully booted it, still building a video cable 
> to
> let it hook up to a Sun monitor that should support it's rez/refresh but has
> different wiring and possibly composite sync.
>     So maybe after everything else stablizes it won't be too hard to adjust a
> linux/i386 driver won't be too hard to adjust.

Correctly written drivers are easy to fix.  The required change is just adding
a couple of subroutine calls.  The problem is that most MIPS boxes don't
deal with cache coherence in hardware.  So that has to be done in software.
So one has to writeback / invalidate the caches before a DMA operation can
start.  See <asm/io.h> in the MIPS kernel sources at the bottom.

>     One question - is there an ECU disk to configure EISA devices or is it in
> firmware?

It's probably a floppy disk as that software isn't in the firmware.  On the
other side I've never seen such a floppy disk.


<Prev in Thread] Current Thread [Next in Thread>