Subject: (fwd) Re: 128MB simms in a SS20?
To: None <port-sparc@netbsd.org>
From: Darren Reed <avalon@caligula.anu.edu.au>
List: port-sparc
Date: 02/03/2004 17:10:34
The article below appeared on aus.computers.sun and had an interesting
little tid-bit of info...:

In article <bvcsgr$l0g$1@Gaia.teknon.de>,
Volker Borchert <v_borchert@despammed.com> wrote:
>Joerg Schilling wrote:
>
>|> I don't have the number of slots in mind but I remember that it
>|> supports 512 MB max. Even if you add that amount of memory you will have
>|> no joy as the way that the MMU and the cache work when syncing the
>|> cache and when setting up desriptors makes a SS20 with 512 MB partially 
>|> slower than one with 256 MB.
>
>Does the SS10 suffer from the same problem? I'm just considering
>to fill up my file/boot server SS10/85 from 256 to 512. Do you
>estimate overall throughput will increase or decrease?

From my memory, the SS20 has a better memory controller than the SS10.
I believe the only difference to the SS10 is:

-	fixed a PCB routing bug that prevented you from using the SX
	framebuffer in the SS10

-	Added some local "intelligence" to the memory controller that 
	should allow MPEG decoding from RAM into the SX framebuffer
	and could be used by things like bcopy() if there was support
	in libc.

If you don't need to buy the RAM, I just would give it a try.
Performance in many cases depends on the exact usage scenario.

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily
-- end of forwarded message --