What is 390Z55 processor on SS20?

[WORLOK]

Hi,

I hate when they list CPU's like this in ads. I tried looking this up
on TI's website, couldn't find it.

Anyone know anything about the performance of this CPU: 390Z55 in a
Sparc20 ??

Thanks,

Tom

--
================================
Viva Linux!! Viva La Revolutión!
================================

Sent via Deja.com
http://www.deja.com/

[WORLOK]

Follow up: Here is another SS20 cpu that tells no info - if anyone knows?

cpu model # 5012617002824

In article <92frk8$i5j$1...@nnrp1.deja.com>,

[Steve Sheldon]

I believe it's like this:
The 390Z55 is a sparc processor with cache.
The 390Z50 is a sparc processor without cache.

These designations don't tell you much because it could be a 50Mhz or a
70Mhz, but you can't know from that designation.

Now if you have the model #, then you can piece out some information.

This one happens to be a 501-2617. You get this from the first seven
digits.

And that happens to be a SM51 processor... i.e. a 50Mhz sparc with a 1Meg
cache.

Here's a list of them:
http://www.sunstuff.org/hardware/partnumbers/501-xxxx.shtml

"WORLOK" <the_w...@excite.com> wrote in message
news:92fs1t$ice$1...@nnrp1.deja.com...

[WORLOK]

OK, I would think that 50mhz would be pretty slow, but 1mb cache is
alot. Is this full speed cache? Anyone know about this CPU ?

Thanks,

Tom

In article <92ftef$lfi$1...@node17.cwnet.frontiernet.net>,

[Martin Slaney]

With later versions of the PROM (certainly 2.25) you can use

OK>module-info

This reports :-

sbus clock speed
mbus clock speed
CPU core speed
cache size/type

OK ?

[Barbie LeVile]

WORLOK wrote:
>
> OK, I would think that 50mhz would be pretty slow, but 1mb cache is
> alot. Is this full speed cache? Anyone know about this CPU ?
>

Its quiet a fast beasty for that little Mhz, i have such in my SS10 and i
can
play back mp3 at full quality with about 33% cpu load, using xaudio it.

--
Barbie - Prayers are like junkmail for Jesus

I have seen things you lusers would not believe.
I've seen Sun monitors on fire off the side of the multimedia lab.
I've seen NTU lights glitter in the dark near the Mail Gate.
All these things will be lost in time, like the root partition last week.
Time to die.

bar...@toppoint.de http://www.amigaworld.com/barbie/index.html

[Steve Sheldon]

"WORLOK" <the_w...@excite.com> wrote in message

news:92g425$pg3$1...@nnrp1.deja.com...

>
>
> OK, I would think that 50mhz would be pretty slow, but 1mb cache is
> alot. Is this full speed cache? Anyone know about this CPU ?
>

I found a really old table of benchmarks out on the net:
ftp://ftp.cs.toronto.edu/cs/ftp/pub/jdd/spectable

Way down at the bottom it shows an SM51 in a SS20 to be faster than an SS10.
Ok, not sure what that means, or how accurate that is.

But, the SM51 is like 3-4 times the speed of a Sparcstation 2. It's also
roughly the same speed as a SS5 with a 85-110Mhz processor.

> --
> ================================
> Viva Linux!! Viva La Revolutión!
> ================================

It's too bad Sun raised the price on Solaris for personal use. :(

[Stephen Dowdy]

)Way down at the bottom it shows an SM51 in a SS20 to be faster than an SS10.
)Ok, not sure what that means, or how accurate that is.

Hmm, the SM51 has to run its MBus at 40MHz, so even though the max MBus
on the SS10 is 40MHz, but 50 on the SS20, it is still running the SS20's
MBus at 40MHz. The reason is that the MBus must run slower than the CPU
/Cache speed. Sun clocked the SM41 at 40.N MHz (forget what N was, but it
was marginally faster than the 40MHz MBus speed). They didn't do that
on the SM51, thus requiring the backoff on the MBus down to 40MHz.

So, given that the SPEC value is higher for the SS20, one has to figure
that the 70ns memory access cycle on the SS20 (vs 80ns on the SS10) makes
the difference. (it's also unclear about the overall system configuration
used to generate said SPEC values, that SS20 may have had more RAM)

There may have been other bandwidth improvements on the SS20 buses over
the SS10 (IIRC there were some design flaws in the SS10's architecture
that required work-arounds in the SW)

)But, the SM51 is like 3-4 times the speed of a Sparcstation 2. It's also
)roughly the same speed as a SS5 with a 85-110Mhz processor.

However, the microSPARC has an itty-bitty cache and far fewer context
registers, so under load that SS5 will perform much poorer than an SS[12]0

--stephen
--
--
Stephen Dowdy - Systems Administrator - CS Dept - Univ of Colorado at Boulder
do...@cs.colorado.edu -- http://www.cs.colorado.edu/~dowdy/signature.html

[Dolf de Waal]

WORLOK heeft geschreven in bericht <92fs1t$ice$1...@nnrp1.deja.com>...

>Follow up: Here is another SS20 cpu that tells no info - if anyone knows?
>
>cpu model # 5012617002824

This is a SM51 50 Mhz SuperSparc module for SS10 , SS10SX , SS20 and SS1000

[mike.s...@hil.hftrust.com]

In article <3a4bcba9$1...@cs.colorado.edu>,

But don't forget that the SS5 (microSPARC-II) didn't suffer *too*
badly from it's small (16Kb+8Kb) L1 cache - because it has the
lowest latency access to main memory of *ANY* Sun workstation,
possibly excepting the Blade.

> do...@cs.colorado.edu --
http://www.cs.colorado.edu/~dowdy/signature.html

[Joga Dhaliwal]

This looks to me to be a 50mhz cpu.