[comp.sys.amiga.programmer] 68030

pjohnson@jabba.uucp (Palmer Johnson) (04/17/91)

Well I just put in my 2630 baord....

What I need is a few of the 030 utilities that are out there if YOU people
could email them to me!!!

PLUS I have one problem....

    when I hold both mouse buttons down and get the menu to PICK between
    68000 and 68030... if pick 68000 the system does not BOOT..just
    sit there with a white screen??

    I have the 4.4 version mother board....
    ANY idea what is up??


     ///  Palmer Johnson
    ///   UUCP: jabba!pjohson@hybrid
\\\///	  another TeleBit 2500 site
 \XX/	  Jabba the Hutt's Palace (519) 258-0056

daveh@cbmvax.commodore.com (Dave Haynie) (04/23/91)

In article <pjohnson.3388@jabba.uucp> pjohnson@jabba.uucp (Palmer Johnson) writes:
>PLUS I have one problem....

>    when I hold both mouse buttons down and get the menu to PICK between
>    68000 and 68030... if pick 68000 the system does not BOOT..just
>    sit there with a white screen??

Are you doing this from a cold boot, or after booting?  If the latter, you 
might expect a white screen, or some other failure mode.  When you first boot
up, Exec will get built for an '030, and a reboot to the 68000 won't normally
wipe that out.  So Exec lunches as soon as the 68000 executes some '030-only
code.  The other thing to look out for is something 68030 specific happening
early in the Startup-Sequence.  I don't know of any other problems.

>     ///  Palmer Johnson

-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
      "That's me in the corner, that's me in the spotlight" -R.E.M.

sschaem@starnet.uucp (Stephan Schaem) (04/24/91)

 Another thing I found on a A2630.
 My SCSI controler is not reconize when running in 68000 mode via the
 A2630 boot menu (But work perfectly without the A2630).
 So I cant use my HD's in 68000 modes...
 Also, something about the 68030 but in the A3000.Why burst is turned
 off for the data cache? or why does it cause a slow down?

daveh@cbmvax.commodore.com (Dave Haynie) (04/26/91)

In article <1991Apr24.075729.6741@starnet.uucp> sschaem@starnet.uucp (Stephan Schaem) writes:
>
> Another thing I found on a A2630.
> My SCSI controler is not reconize when running in 68000 mode via the
> A2630 boot menu (But work perfectly without the A2630).

Don't know why that should be, the hard disk controller is out on the
expansion bus.  The most likely cause of such a problem would be when you have 
your hard disk partitions explicitly set up for FAST RAM (not just any RAM),
and all your Fast RAM goes away when the A2630 memory disappears.  This is an
unlikely condition with any Amiga hard disk controller, though third parties
may set things up differently as a default.  One of the reasonably common
expansion board listing tools should be able to tell you if the system has
actually recognized the hard disk controller board.

> Also, something about the 68030 but in the A3000.Why burst is turned
> off for the data cache? or why does it cause a slow down?

Under 1.3, only the I-Cache is turned on, because the OS only knows about
the 68020, and it didn't have D-Cache or burst.  Based on their benchmarks, the 
software people have found that data bursting probably slows things down as 
much as it helps, at least on the A3000.

Burst mode is essentially an extended cache prefetch mode.  When the 68030 
reads a longword location, and cache bursting is enabled, it can get the next
three longwords, modulo 4, very quickly.  However, they don't get fetched in
zero time, just "very quickly" (on the A3000, each bursted word is fetched
2.5x faster than a work fetched the normal random access memory way).  Since
this is a cache prefetch, these extra longwords are just put into the I or D
cache, as appropriate.  There's no guarantee that they'll ever be used before
overwritten by the cache.  Apparently, the testing has shown that data so
prefetched isn't used enough to pay for the prefetch time.  In other words, 
D-Cache burst slows things down.  Now, it certainly won't slow down everything,
and in fact, very simplistic articifial benchmarks probably see this effect
less than real world software.  Via the SetCPU or 2.0's CPU command, you're
free to turn on D-Cache burst if you want to.


-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
      "That's me in the corner, that's me in the spotlight" -R.E.M.

dltaylor@cns.SanDiego.NCR.COM (Dan Taylor) (04/26/91)

In <1991Apr24.075729.6741@starnet.uucp> sschaem@starnet.uucp (Stephan Schaem) writes:

> My SCSI controler is not reconize when running in 68000 mode via the
> A2630 boot menu (But work perfectly without the A2630).

Which disk controller?  My A2091s both work fine whether I run the 68000
or the 68030 (selected at boot time from the option screen).  I do lose
the 4 Meg of fast RAM, naturally.

>Why burst is turned off for the data cache? or why does it cause a slow down?

That's controllable.  It's normally off because the small capacity RAM
does not work in burst mode.  If you have the page mode RAM, turn it on.
The directions should be in your manual.

Dan Taylor

chucks@pnet51.orb.mn.org (Erik Funkenbusch) (04/27/91)

sschaem@starnet.uucp (Stephan Schaem) writes:
>
> Another thing I found on a A2630.
> My SCSI controler is not reconize when running in 68000 mode via the
> A2630 boot menu (But work perfectly without the A2630).
> So I cant use my HD's in 68000 modes...
> Also, something about the 68030 but in the A3000.Why burst is turned
> off for the data cache? or why does it cause a slow down?


You have probably specified fast ram for you controller.  in which case, if
you have a 1 meg chip ram, and no other memory besides what's on the 30 board,
then there is no fast mem, so the controller isn't recognized.  burst mode is
not implemented on the 2630, so it should make no difference if it's off or
on.

/s
.--------------------------------------------------------------------------.
| UUCP: {amdahl!tcnet, crash}!orbit!pnet51!chucks | "I know he's come back |
| ARPA: crash!orbit!pnet51!chucks@nosc.mil        | from the dead, but do  |
| INET: chucks@pnet51.orb.mn.org                  | you really think he's  |
|-------------------------------------------------| moved back in?"        |
| Amiga programmer at large, employment options   | Lou Diamond Philips in |
| welcome, inquire within.                        | "The First Power".     |
`--------------------------------------------------------------------------'

sschaem@starnet.uucp (Stephan Schaem) (04/28/91)

 Tank you for the clarification on the Cache subject.
 I never even thought of the fast ram problem, I still use the mount
 command since its an old scsi card and dont offer autoboot.
 I have check timing with the data burst on, but didn't know why it was
 slower.I'm not into 680x0 programing yet, but its something else worth
 considering...
 I have done only 1 program running on 68020/30, and special desing for
 the coresponding processor speed things in a major way.
 Since this is a IFF show programe I will be able to test if data burst
 is efective and how mutch.
 Thank's about for the reply.

							Stephan.

sschaem@starnet.uucp (Stephan Schaem) (04/28/91)

 Its an Overdrive from Pacific Peripherial, and dont have the autoboot
 feature.But I will check the mountlist to check the buffer memtype...

 I know its user controlable, but didn't knew how the burst mode affect
 memory fetching.I will try it out and see how efetive it is.

							Stephan.