> > As far as I can tell, the formula for SS10's and SS20's is as follows:
> > 	UNIT/sen
> > 	  |  |||
> > 	  |  ||\------	number of cpus. blank = 1 cpu
> > 	  |  ||
> > 	  |  |\-------	external cache. 0 = no, 1 = yes
> > 	  |  |
> > 	  |  \--------	cpu speed. 4 = 40MHz, 5 = 50Mhz, 6 = 60MHz, 7 = 75MHz
> > 	  |			   8 = 85MHz (?), 15 = Ross 150MHz hypersparcs
> >           |
> > 	  \-----------	base type. SS10, SS20, SC1000, SC2000
> 
> 
> Further to the above...
> 
> Other S10 models were the Model 20 and 30.  Model 20 was 33MHz, Model 30
> was 36MHz.
> 
> HyperSPARC S20 models also included the Model HS11 (100MHz), HS21
> (125MHz), HS22MP (2x 125MHz), HS14MP (4x 100MHz), as well as the later
> Model 151 (150MHz) and 152 (2x 150MHz) above.
> 
> I don't recall any 85MHz SuperSPARC S20s.

Sun may not have shipped them, but Axil does (85mhz Axil 320 and 420
models 81 and 812).  Sun made the CPU boards but their marketing people
are sorta odd some times.

The designations for the HS models were singularly messed up.
Axil Engineering always used the following pattern (our marketing
people sometimes follow this and sometimes Sun's equivalent):
9H1
10H1
10H12
11H12
12H14
12H14-512
15H12
20H12

The H designates HyperSPARC; other than that, the first part is the
speed (x10 mhz, so the above are 90,100,100,110,125,125,150,200),
the second part is E$ (always present on HS) and # installed if
more than one.  The -512 is for those models which came in more than
one size E$, we'd have postfix on the larger sizes.

We always thought this scheme was more easy to read and consistent
with the SuperSPARC designators.


-george william herbert
gherbert@crl.com  gherbert@axil.com

[note: I've gotten some comments asking where my project of benchmarking
NetBSD Sparcs has gone.  It's gotten pushed down the queue some due to
Axil's expansion into the NT marketplace, but I do intend to get to it.]