[rescue] e4k / 250/1mb cache vs 250/4mb cache

[Sandwich Maker]

" From: Mike F <lists at ibrew.net>
" 
" On Sat, Oct 09, 2004 at 07:43:52PM -0700, Skeezics Boondoggle wrote:
" > 
" > I was under the impression that until the Ultra-III-based systems came
" > out, mixing speeds was officially unsupported, and even if it does happen
" > to work, the system simply finds the least common denominator - clocks
" > down the faster CPUs to match the slowest ones, and ignores any larger 
" > caches than the smallest.
" > 
" > I'm almost positive that was true of, say, the sun4d and sun4m boxes; put 
" > an SM51 and an SM61 in a Sparc 20 and it'll run them both at 50Mhz.  Put 
" > some SM81's in an SC2000 with some SM81-2's, and it'll just ignore the 2nd 
" > meg of cache on the larger modules.
" [snip]
" 
" For no other reason than the fact that I'm cheap, I'm running a 
" SPARCStation 20 with an SM71 and an SM51. prtdiag doesn't work on SUN4M 
" systems, but here's what `psrinfo -v` reports:
" 
" Status of virtual processor 0 as of: 10/12/2004 10:32:41
"   on-line since 06/05/2004 02:10:43.
"   The sparc processor operates at 75 MHz,
"         and has a sparc floating point processor.
" Status of virtual processor 2 as of: 10/12/2004 10:32:41
"   on-line since 06/05/2004 02:10:47.
"   The sparc processor operates at 50 MHz,
"         and has a sparc floating point processor.
" 
" These modules both have 1MB of cache, so I can't say what would happen 
" with differing caches, but it does seem to maintain differing processor 
" speeds.

yeah, but iirc it still schedules them as if they were both 50s.
________________________________________________________________________
Andrew Hay                                  the genius nature
internet rambler                            is to see what all have seen
adh at an.bradford.ma.us                       and think what none thought