From: VMACS Test Account
Subject: 75 MHz / 1M Supersparc vs 125 MHz / .5M Ross Hypersparc for lisp especially CXL
Date: Wed, 11 Jun 1997 00:00:00 +0000
Message-ID: <FRD.97Jun11145234@sunrise.Stanford.EDU>

Folks --

Any body have an concrete experience and/or benches
for these two?

I have to make a non-returnable choice between the two
processor boards. My machine is SS20, 64M, SunOS4.1.4,
lcl4.1 . The application of concern runs under CLX.

I am hoping for empirical data, like from any Garnet
people (under cmucl is fine).

I know some theory -- like more cache is good, and
beyond that it is very difficult to predict performance.

I have two data points so far: a Ross 125 / .25M
was only 30% faster than an SM41 40MHz / 1M -- on both
display intensive CLX tasks as well as pure symbol
manipulation.

Any ideas appreciated.

Thanks,
f