SUMMARY: SM71 vs. Ross 200MHz...

From: Paul H. Yoshimune (
Date: Wed Jul 28 1999 - 14:02:30 CDT

Thanks to the following for the reponses...

David L. Markowitz
Rick Fincher
Scott R. Kulp
Michael J. Connolly
Scott McDermott

Conclusions were:

1) If you can afford to go UltraSPARC, do so
2) For the tasks mentioned, the dual SM71 setup would likely be more beneficial
   than a single Ross 200 (servers and compiles), mostly due to cache -
   although one person felt the clock speed of the Ross would probably negate
   the lack of cache
3) For CAD/Photoshop-type work, the single Ross is probably the way to go
4) Ross is no longer in business, so hanging on to one of the old processors
   might be cheap insurance in case of failure

Since posting my query, a visit to the SunStore ( has shown
that they're offloading SM81's at $499 each, brand new. This is roughly twice
what the SM71's cost used, and I don't think the differential in performance
between the two is that high. Just something to keep in mind if you're looking
for an upgrade, though.

Original query follows:


I'm looking to upgrade an SS20 with dual SM61's to either one or two SM71's, or
a single HyperSparc 200. This box is being used as a list server, mail hub,
web server, DNS server, and code compiler primarily. So, it's running as a
server as well as having to deal with code builds and interactive use.

I understand the the generalization is that the Ross is better for floating
point work, and that the SuperSparc is better for server work because of the
larger cache. But, given that the HS200 specs out so much faster, does this
still apply?

My specific questions are:

-Given the above usage, which would you choose?
-Are there any issues with running an HS200 in a SS20? The installation manual
 I downloaded from only covers through the HS150.
-The SM61's I use currently generate a fair bit of heat; how much worse is an
 SM71 or HS200 setup going to be?
-I know with the Ross I need OBP 2.22 or later; any considerations for the

Will summarize, of course...thanks.

Paul H. Yoshimune

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:24 CDT