Michael Kraemer wrote:
>
> I have a PWS 500au running with 640MB (2x64,2x128,2x128) right now.
> I'm currently trying to upgrade memory to the max 1.5GB possible,
> without luck so far.
> I've got a pair of 2x256MB pieces (ECC) with exactly the same
> specs as one of the 2x128 pair (except twice the capacity).
> When built in, console recognizes the extra memory,
> however, when booting Tru64, it crashes miserably with memory failures.
> I tried another pair of 2x256MBs, changed slots etc, same result.

I think this is just the behaviour you get when you try
installing registered DIMMs.
Also, the PWS in fact needs some weird memory speed like
83MHz. (not sure about the actual number.) PC100 DIMMs will
usually work, but PC133 might not.

gruss,
Dennis

--
Don't suffer from insanity...
Enjoy every minute of it.

In article , Dennis Grevenstein
writes:
>
> I think this is just the behaviour you get when you try
> installing registered DIMMs.
> Also, the PWS in fact needs some weird memory speed like
> 83MHz. (not sure about the actual number.) PC100 DIMMs will
> usually work, but PC133 might not.

that's the conventional wisdom, sure. However, I have
2x128MB ECC SDRAM CL3 133MHz
which seem to work (i.e. I ran Tru64 for several weeks continously),
but from the very same vendor the 2x256MB parts with otherwise same specs
crash the OS during boot.
Anyway, I"ll probably have to look for the 100MHz parts now.

On Aug 22, 7:25 am, m.krae...@gsi.de (Michael Kraemer) wrote:
> In article , Dennis Grevenstein
>
> writes:
>
> > I think this is just the behaviour you get when you try
> > installing registered DIMMs.
> > Also, the PWS in fact needs some weird memory speed like
> > 83MHz. (not sure about the actual number.) PC100 DIMMs will
> > usually work, but PC133 might not.
>
> that's the conventional wisdom, sure. However, I have
> 2x128MB ECC SDRAM CL3 133MHz
> which seem to work (i.e. I ran Tru64 for several weeks continously),
> but from the very same vendor the 2x256MB parts with otherwise same specs
> crash the OS during boot.
> Anyway, I"ll probably have to look for the 100MHz parts now.

When PC100 was current and PC133 was the new thing, supposedly the
chip manufacturers made chips that could run at either speed (or the
chips were built and then tested; those that could run reliably at 133
were sold as such, those failing but still working at 100 were sold as
that); that meant that the early 133 chips could generally run at the
slower speeds.

Later when PC100 was moribund, the manufacturers pushing for yield on
133s modified the chips so that they would no longer run reliably at
the slower speeds but had a higher acceptance rate at 133. That is
supposedly why the early stories about PC133 memory running OK on
PC100 systems is around, but why it is not true for most PC133 memory
floating around today.

Or so I heard...

My PWS600au has 2x256 DEC DIMMs, presumably 83MHz models, plus 2x256
Micron/Crucial DIMMs that claimed to be 100MHz; they work just fine
together.

Rich

The other reason why newer PC133s would not work in older motherboards is chip
density. The newer DIMMs were manufactured with a smaller number of high
density memory chips, and this causes timing issues when they are installed on
older motherboards... Ben Myers

On Wed, 22 Aug 2007 11:10:46 -0700, Rich Jordan wrote:

>On Aug 22, 7:25 am, m.krae...@gsi.de (Michael Kraemer) wrote:
>> In article , Dennis Grevenstein
>>
>> writes:
>>
>> > I think this is just the behaviour you get when you try
>> > installing registered DIMMs.
>> > Also, the PWS in fact needs some weird memory speed like
>> > 83MHz. (not sure about the actual number.) PC100 DIMMs will
>> > usually work, but PC133 might not.
>>
>> that's the conventional wisdom, sure. However, I have
>> 2x128MB ECC SDRAM CL3 133MHz
>> which seem to work (i.e. I ran Tru64 for several weeks continously),
>> but from the very same vendor the 2x256MB parts with otherwise same specs
>> crash the OS during boot.
>> Anyway, I"ll probably have to look for the 100MHz parts now.
>
>When PC100 was current and PC133 was the new thing, supposedly the
>chip manufacturers made chips that could run at either speed (or the
>chips were built and then tested; those that could run reliably at 133
>were sold as such, those failing but still working at 100 were sold as
>that); that meant that the early 133 chips could generally run at the
>slower speeds.
>
>Later when PC100 was moribund, the manufacturers pushing for yield on
>133s modified the chips so that they would no longer run reliably at
>the slower speeds but had a higher acceptance rate at 133. That is
>supposedly why the early stories about PC133 memory running OK on
>PC100 systems is around, but why it is not true for most PC133 memory
>floating around today.
>
>Or so I heard...
>
>My PWS600au has 2x256 DEC DIMMs, presumably 83MHz models, plus 2x256
>Micron/Crucial DIMMs that claimed to be 100MHz; they work just fine
>together.
>
>Rich