[comp.sys.amiga.hardware] A3000/16 vs. A3000/25 non-CPU speed

jwwalden@miavx1.acs.muohio.edu (Darc Tangent) (03/05/91)

Would there be a noticeable speed difference between an A3000/16 with a 68040
accelerator and an A3000/25 with the same accelerator?  I know that the new
custom chips in the A3000 are clocked at the same rate as the main processor
so there should be some difference, but I was wondering if anyone had a good
idea of how much difference there would be and what operations in particular
the difference would affect.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Darc Tangent			   "She's a brain trapped in the body of a
jwwalden@miavx1.acs.muohio.edu	    game show hostess" - _Say_Anything_

holgerl@amiux.UUCP (Holger Lubitz) (03/06/91)

In article <4101.27d2e586@miavx1.acs.muohio.edu> jwwalden@miavx1.acs.muohio.edu (Darc Tangent) writes:

>I know that the new
>custom chips in the A3000 are clocked at the same rate as the main processor

This is the first time I read about custom chips running at a higher rate
than 14.32 MHz (and even slightly slower in PAL Amigas)
Processor Clock Rate on the chip bus ? You are not talking about a
chip bus running at 25 MHz bus speed, are you ?

Could some hardware guru out there (hello dave !?) confirm this if it
is true ?  Or is it simply some kind of misunderstanding on my part ?

Eagerly waiting for an answer,
Holger

--
Holger Lubitz            | holgerl@amiux.uucp
Kl. Drakenburger Str. 24 | holgerl@amiux.han.de
D-W-3070 Nienburg        | cbmvax.commodore.com!cbmehq!cbmger!amiux!holgerl

krauss@mansw1.enet.dec.com (Michael Krauss) (03/06/91)

In article <holgerl.0243@amiux.UUCP>, holgerl@amiux.UUCP (Holger Lubitz) writes:

>
>In article <4101.27d2e586@miavx1.acs.muohio.edu> jwwalden@miavx1.acs.muohio.edu (Darc Tangent) writes:
>
>>I know that the new
>>custom chips in the A3000 are clocked at the same rate as the main processor
>
>This is the first time I read about custom chips running at a higher rate
>than 14.32 MHz (and even slightly slower in PAL Amigas)
>Processor Clock Rate on the chip bus ? You are not talking about a
>chip bus running at 25 MHz bus speed, are you ?
>                        [...]

I don't no at which rate the custom chips are running, but as far as I remember
there is a jumper on the motherboard of the A3000 that changes some timings of
the customchips.(It has two settings: one for 16M and one for 25M.) 

>
>--
>Holger Lubitz            | holgerl@amiux.uucp
>Kl. Drakenburger Str. 24 | holgerl@amiux.han.de
>D-W-3070 Nienburg        | cbmvax.commodore.com!cbmehq!cbmger!amiux!holgerl
>
--
==============================================================================
Michael Krauss			#include <std_disclaimer.h>
DEC Mannheim, Germany		"AMIGA or VAX, nothing else please !"
"krauss@mansw1.enet.dec.com" or "...!decwrl!mansw1.enet!krauss"

milamber@caen.engin.umich.edu (Daryl Scott Cantrell) (03/06/91)

In article <1991Mar6.142824.12241@hollie.rdg.dec.com> decwrl!mansw1.enet!krauss writes:
>>
>>In article <4101.27d2e586@miavx1.acs.muohio.edu> jwwalden@miavx1.acs.muohio.edu (Darc Tangent) writes:
>>
>>>I know that the new
>>>custom chips in the A3000 are clocked at the same rate as the main processor
>>
>>This is the first time I read about custom chips running at a higher rate
>>than 14.32 MHz (and even slightly slower in PAL Amigas)
>>Processor Clock Rate on the chip bus ? You are not talking about a
>>chip bus running at 25 MHz bus speed, are you ?
>>                        [...]
>
>I don't no at which rate the custom chips are running, but as far as I remember
>there is a jumper on the motherboard of the A3000 that changes some timings of
>the customchips.(It has two settings: one for 16M and one for 25M.) 
>

  As far as I know, the custom chips in all Amigas run at 7.14 MHz (for now,
hopefully).  The 16/25 MHz clock in an A3000 drives a totally different bus.
I think you might be thinking of the two pairs of shorting posts in the 3000
that configure ROMs based on the type of chip and the speed of the bus.  I
think these are for determining wait states and such..

  As to the original question, the only differences between a 3000/25 and a
3000/16 are:

  oscillator
  68030 rating
  FPU type/rating

In other words, the entire 030 bus of a 3000/16 is built to go 25 MHz, but
doesn't (unless you were to plop something into the 200-pin CPU slot).. So
when you get that 040 accelorator, it will go 25 MHz in either.


>>Holger Lubitz            | holgerl@amiux.uucp
>Michael Krauss			#include <std_disclaimer.h>
>"krauss@mansw1.enet.dec.com" or "...!decwrl!mansw1.enet!krauss"

--
+---------------------------------------+----------------------------+
|   // Daryl S. Cantrell                |   These opinions are       |
| |\\\ milamber@caen.engin.umich.edu    |    shared by all of    //  |
| |//  Evolution's over.  We won.       |        Humanity.     \X/   |
+---------------------------------------+----------------------------+

jwwalden@miavx1.acs.muohio.edu (Darc Tangent) (03/07/91)

Thank you for your answer to my question.  What I meant by new custom chips
(which seems to have confused everyone) was the RAMSey and other new chips that
are specific to the A3000.  Sorry for the confusion about that.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Darc Tangent			   "Giants are just like people, only bigger.
jwwalden@miavx1.acs.muohio.edu	    Much, much bigger!"
jwwalden@miavx1.bitnet			- the witch in _Into_the_Woods_

daveh@cbmvax.commodore.com (Dave Haynie) (03/09/91)

In article <holgerl.0243@amiux.UUCP> holgerl@amiux.UUCP (Holger Lubitz) writes:
>In article <4101.27d2e586@miavx1.acs.muohio.edu> jwwalden@miavx1.acs.muohio.edu (Darc Tangent) writes:

>>I know that the new
>>custom chips in the A3000 are clocked at the same rate as the main processor

>This is the first time I read about custom chips running at a higher rate
>than 14.32 MHz (and even slightly slower in PAL Amigas)

OK.  The original "Amiga" chips in the A3000, Agnus, Denise, and Paula, work
from a two phase clock, nominally 7MHz -- 7.16MHz for NTSC, 7.09MHz for PAL.
Since the two clocks, called C7M and CDAC, are 90 degrees apart, the chips
are effectively clock at 14MHz, and the basic chip memory cycle is equivalent
to a 14MHz 68000 cycle.

The new A3000 system chips, Buster, DMAC, Gary, and Ramsey, are clocked by
the CPU clock (16MHz or 25MHz).  At least Buster and Ramsey are actually clock
by this clock, called CPUCLK, and another 25MHz clock delayed by 90 degrees,
called CLK90.  Buster actually uses both edges of both clocks at various points
in the logic.

There are a few reasons for using a two phase clock rather than a single clock
running at double speed.  First of all, the 68030 runs from the 25MHz clock,
so it's much easier to feed that same clock to the custom chips that need it
than it is to build a 50MHz clock that's exactly synchronized to that 25MHz
clock.  Secondly, real fast clocks are harder to keep clean on the PCB, and
noisier when it comes time for FCC testing.  Finally, the gate array technology
used for these parts is much happier with a couple 25MHz clocks on input than 
a 50MHz clock.  From the point of view of the actual circuit design, you really
don't care.
-- 
Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: hazy     BIX: hazy
	"What works for me might work for you"	-Jimmy Buffett