Conference hydra::amiga_v1

    re: .0
    
    This is the second ad I've seen for a 68040 card for an Amiga 2000.
    The previous one also didn't specify a price or an availability date.
    I think it's vapor.
        John Sauter

re        <<< Note 4652.0 by STAR::GUINEAU "but what was the question?" >>>
>                            -< 68040 cards are in >-

	Only two questions:

	'ow much ?

	when ?

> It's for the Amiga 2000 only so far, (sorry Steve!).

	Oh, thats just  FINE  with me (-:

	Reg

	This a request more than a question;

	Have you posted this in the Mac conference ?

    Just a tiny question...How does a machine that runs at 25 MHZ get a 27
    MIPS performance (multiple instruction per clock cycle.  Now that's
    interesting!)
    
    :^)
    
    Tony

    A machine can get 27 MIPS with a 25 Mhz clock by issuing multiple
    instructions per cycle.
        John Sauter

               <<< Note 4652.5 by SAUTER::SAUTER "John Sauter" >>>
                          -< possible, but unlikely >-
    
    Not unlikely at all.  This is exactly what Motorola is doing -- it's
    called a "superscalar" design.  Some of the techniques used in RISC
    processor design have fed back to the CISC (68xxx,80xxx) world.  I
    think _Byte_ had an article on the 68040, as did several magazines in
    the electronics trade press.
    
    Wes

>               <<< Note 4652.5 by SAUTER::SAUTER "John Sauter" >>>
>                          -< possible, but unlikely >-
    
>    Not unlikely at all.  This is exactly what Motorola is doing -- it's

	There was an article around Oct 89 describing the application of
RISC design styles to CISC.   "Cray on a chip"  or some such headline, I
can probably find it, described specific details of the 68040, 80486 and 
????, I forgot.  It was VERY optimistic abpout the future of CISC as these
design styles are applied and the design tools continue to expand to handle
the more complex designs.  Dec has it soon, etc.

	Reg

A 25mhz 68040 is about 15 mips.  2 to 2.5 mflops.

Article 10488 of comp.sys.amiga.hardware
Path: nntpd.lkg.dec.com!news.crl.dec.com!deccrl!bloom-beacon!mintaka!think.com!zaphod.mps.ohio-state.edu!wuarchive!uunet!mcsun!hp4nl!alchemy!ptavoly
From: [email protected] (Peter Tavoly)
Newsgroups: comp.sys.amiga.hardware
Subject: 68040 board (Fusion Forty) available (LONG!)
Keywords: speed!!!!
Message-ID: <[email protected]>
Date: 30 May 91 16:00:29 GMT
Organization: Utrecht University, Dept. of Computer Science
Lines: 202

This is a review of the information I got, not of the board itself.

Last friday I finally got the info I requested some months ago (!) via the
Amigaworld reader service card. Perhaps this was not their fault, but they
did manage to spell my name as 'Thoams' (weird :^)

My own comments are between square brackets [] and are to be disregarded,
best way is to go through 'sed' deleting all in between []'s :^)

The letter consisted of four pages, [1] with a picture and specs of the
board, [2] with some test results with the Toaster and Imagine 3D, [3] with
a comparison benchmark (linpacks and dhrystones) and [4] an order form.

Anyway, [1] said:

"Let those who have understanding reckon the power of"

[hmm.. one of the developers (or ad people) must have been an Iron Maiden
 fan :^) ]

[large augmented capitals] FUSION FORTY

[a small rastered picture of the card on the left, showing the MC68040 with
 heat sink and what seems to be 16MB of SCRAM ZIPs. The board is pretty
 tightly packed (full length) and looks OK. I can see about 36 rectangular
 chips (of which four are probably EPROMs), two square chips (of which one
 is the 68040), one quartz and some miscellaneous components like resistors,
 condensors and other stuff I can't make out. No SMD, but everything seems
 to be socketed (very hard to see). The CPU is obscured by an enormous heat
 sink; is the 68040 really that large? There also seems to be a connector
 near the RAM space that looks like a floppy/parallel/serial/whatever
 connector (?) ]

[to the middle right:]

Mega-steroid acceleration for the Amiga 2000

Performance:
 o Motorola MC68040 microprocessor at 25 Mhz.
 o 18 - 25 MIPS, 3.5 - 8.0 MFLOPS       [seems a bit too optimistic..]
 o 32-bit RAM

Quality and availability:
 o 6-layer board with separate ground and power planes for uniform voltage
    stability.
 o extensive use of decoupling devices minimizes electrical noise.
 o high-performance clock for the tight electrical specifications of the
   MC68040.
 o heat sink to dissipate heat generated by the MC68040.

Expandability:
 o memory available in 4 MB, 16 MB and 32 MB configurations.

Features:
 o hardware select switch to disable accelerator board and run original
   processor.
 o asynchronous design for Genlock compatibility.
 o user object code compatibility with all earlier Motorola 68000 series
   microprocessors
 o 9 watt power usage.
 o compatible with Videotoaster, Imagine and other software.
 o one year warranty.

[to the bottom-middle a box with comparisons:]

Board:  Fusion40       C= 2630        A3001 GVP     IBM i486

CPU:    MC68040        MC68030        MC68030       i80486
Clock:  25 Mhz         25 Mhz         25 Mhz        25 Mhz
MIPS:   20+            5.8+           6.4+          15

[remember: MIPS = Meaningless Information about Processor Speed (or Million
 Instructions Per Second). No SPECMARKs are given]

MFLOPS: 3.5+           <1             <1            1. [yes, like that: 1.]
Cache:  4KB x2         256bytes x2    256 bytes x2  8 KB
Burst:  Yes            No   [?]       Yes           Yes
Memory: 4 MB standard  Max. 4 MB      needs         [empty space]
                                      daughterboard
on board: Max. 32 MB   Max. 4 MB      [empty space] [empty space]

[address:]

RCS Management Inc.
120 Mc Gill, Montreal, Quebec,      Tel.:(514) 871-4924
Canada, H2Y 2E5                     Fax.:(514) 871-4926

[2 lines of copyright stuff]

[to the bottom right the name and phone # of the ad agency. I might as well
 assert here that I do not consider this as 'professional looking': bad
 choice of letter styles, graphics not very good, bad picture quality, wrong
 kind of fancying (ad printed slanted) etc. Either this was a minute job, or
 they are like Commodore :^) -> IMHO, IMHO]

[2]

[centered] Fusion-Forty
           16 MB 25 Mhz

Videotoaster test:
[all kinds of selection numbers probably intended to dazzle the novice]

objects: space: spaceships
Rendered with high resolution: 2 minutes 54 seconds.
with alias:                    6 minutes 23 seconds.

[no comparison to 68030, so numbers are pretty useless unless you know what
 that spaceship thing really is..]

Imagine 3D:

We are 2.5 - 3 times faster than 68030 25 Mhz.

[that is at least a clear claim, however, it is not supported by any
 numbers, apart from some wording like: "The F40 generated the image twice
 while the 68030 reached only 97%"]

[3]

Comparison benchmarks using linpack and dhrystones

[no mention of what version..]

[I left out the column processor type, because it didn't fit, you can derive
 the type from the name anyway]

           Int.  Ext.  D(Kflops) S(Kflops)        Ronin     Clock MIPS Max.
           Cache Cache Linpack   Linpack  Drhyst. CPU-speed (Mhz)      memory

NCR
 S486/MC33 8K    128K            882.2    26929   N/A       33    15.39 64MB
EVEREX STEP
 486/33    8K    128K            895.9    26912   N/A       33    15.38  8MB
CLUB HAWK III
 486/33    8K    256K            926.3    27472   N/A       33    15.70  8MB
AST PREMIUM
 486/33    8K    NONE            894.7    25849   N/A       33    14.77  4MB
Commodore
 A2620  256bytes x2 NONE    50            4725     3.97     14     2.70  4MB
GVP 68030 256bytes x2 NONE 102            10330    9.13     25     5.90 20MB
GVP 68030 256bytes x2 NONE 204            20660   18.26     50    11.81 20MB
Fusion-Forty
           4K x2 NONE     1920            31645   27.40     25    18.08 32MB

[there are also two notes:]
* figures shown for the 50 Mhz version of the GVP board are double the ones
  of the 25 Mhz board as it was not actually tested
** The figures used for the Fusion-Forty does not at present represent
  maximum throughput, as the programs used were not optimized for the MC68040

[4]

PURCHASE ORDER FORM

       VERSION    PRICE   QUANTITY     DISCOUNT   TOTAL

F40     4 MB      2995.US                $100 
F40    16 MB      3995.US                $125
F40    32 MB      5595.US                $175
AmigaNet 2000      500.US                $ 20

[various data to be entered + payment terms]

- certified check / money order / company check
- PAYMENT IN ADVANCE
- For orders greater than 1 unit, a 50% deposit will be accepted.

delivery:

6-8 weeks from date of purchase. delivery time from Motorola is 4-5 weeks
from date of order.

[does this mean that you will have to wait at most 8+5=13 weeks ??]

[*large* letters:]

THIS OFFER ENDS IN 60 DAYS

[date on the postage stamp is the verification date. Since the stamp says
 16-05-1991 for me, uhh.. you figure it out :) ] 

--------

<Sigh> I am in no way affiliated with RCS Management or their dog, except
that I *yearn* to have one of those boards.. In fact, why did I type this
up anyway? If you read through it and fell asleep, send me a mail, just so
that I may think twice next time :^) (Or just a subject: I saw it)

Now I have to go and soak my fingertips,

 -Thomas.            ->new sig!

---------------------------------------------------------------------- ____
Thomas Tavoly    |The meaning of life, the universe & everything:     / / /
Commercial CS    | Speed Metal, Waterpolo, Hungarian food,         AMIGA /
HEAO Utrecht, NL.| SF/Fantasy, c.s.a.*, UN*X, s*x & of course:____  / / /
------------------------------------------------------------- \ \ \/ / /
Don't call us, we will not call you either. -TT   .sig v3.2    \_\_\/_/ 
God made one mistake when he created man: He wrote self-modifying code.. -E'n
-----------------------------------------------------------------------------
[email protected] = Peter Tavoly != ME! (Just my account :^)

Article        10498
From: [email protected] (Michael Sinz)
Newsgroups: comp.sys.amiga.programmer,comp.sys.amiga.hardware,adsp.sw,adsp.hw
Subject: 68040 Compatibility Warning
Summary: Issues on dealing with copyback caches in the 68040 and AmigaOS 2.0
Date: 30 May 91 18:48:15 GMT
Organization: Commodore, West Chester, PA
Lines: 119
 
 
           *** 68040 CPU Compatibility Warning ***
 
 
Now that the 68040 CPU is available, it will not be long before
it is available on the Amiga. In fact, in many Amiga magazines
there are already ads for 68040 CPU cards.
 
With the 68040, you get a much faster and more powerful CPU.  It
has 4K of cache memory for instructions and another 4K for data.
The reason these are two separate caches is so that the CPU core
can access data while it is accessing instructions.  (That is, it
can do both *at the same time*)
 
Just the fact that the caches are so much larger can give
software that loads and then runs code some problems.  However,
this is not the worst case.  The 68040 data cache has a mode that
makes the machine run *much* faster in most cases.  It is called
CopyBack. CopyBack means that when a program writes data to
memory, it goes into the cache but *not* into the physical RAM.
That means that if something was to read that RAM and it did not
go through the data cache on the 68040, it will see old data.
The two main examples of this are DMA devices and the instruction
reading of the CPU itself.  This means that even if the
instruction cache is cleared, if you write to memory and then try
to execute that code it may not be in physical RAM yet and the
instruction read from RAM will be the stale data.
 
Simply put:  If you have the CPU store instructions into memory
that you wish to execute later, you *must* clear the caches if
you wish to work with 68040 caching modes.
 
AmigaOS 2.0 correctly clears the caches as needed after it does
the LoadSeg() of a program.  Applications need to do the same if
they create code at run-time.  One such example was the article
on multiple processes in AmigaMail.  What is needed is that just
before the call to CreateProc() that a call to the EXEC V37
CacheClearU() function is executed.  In C that would be:
 
 
extern struct ExecBase *SysBase;
:
:
:
/* We have just created a fake seglist and wish to start it */
/* If we are in 2.0, call CacheClearU() before CreateProc() */
if (SysBase->LibNode.lib_Version >= 37) CacheClearU();
 
/* Now do the CreateProc() call... */
proc=CreateProc(... /* whatever your call is like */ ...);
:
:
:
 
 
 
For those of you programming in assembly:
*
*******************************************************************************
*
* Check to see if we are running in V37 ROM or better.  If so,
* we want to call CacheClearU() to make sure we are safe on future
* hardware such as the 68040.  This section of code assumes that
* a6 points at ExecBase.  a0/a1/d0/d1 are trashed in CacheClearU()
*
		cmpi.w	#37,LIB_VERSION(a6)	; Check if exec is >= V37
		bcs.s	TooOld			; If less than V37, too old...
		jsr	_LVOCacheClearU(a6)	; Clear the cache...
TooOld:
*
*******************************************************************************
*
 
 
 
The above will keep the code working pre-2.0 but will do the
correct operations in 2.0.  Note that while this would mean that
your code could not work on a 68040 without the final 2.0
release, it is not much of a concern since pre-2.0 versions of
the OS could not correctly run with the 68040 and will not work
if the 68040 caches are turned on in CopyBack mode.
 
Note that CreateProc() is not the only time this could be a
problem. Whenever you create code (or load code) that is not 100%
done only via LoadSeg() you will need to call CacheClearU().
Many input.device handlers have been known to allocate and copy
up the handler code and then exit back to the system.  These
programs also need to have this call in them.
 
The other major case that can cause problems is with DMA devices.
Since DMA devices read/write data to memory directly, caches need
to be flushed as needed.  For example, if a DMA device was about
to read RAM via DMA, it would need to call CachePreDMA() to make
sure that caches have written to memory and that all is safe.
 
If a DMA device is about to write to memory, it should call
CachePreDMA() *before* the write, to make sure that no cache data
is still left to be written to RAM, then do the DMA, and finally
call CachePostDMA().
 
In addition, CachePreDMA() and CachePostDMA() gives the OS the
chance to tell the DMA device that the physical addresses and
memory sizes are not the same.  While currently this can not
happen, the future may bring such things as virtual memory and
thus it would be needed.  See the autodocs for more information
on these calls.
 
 
			-- Michael Sinz
			   Operating Systems Development Group
			   Commodore-Amiga, Inc.
 
/----------------------------------------------------------------------\
|      /// Michael Sinz  -  Amiga Software Engineer                    |
|     ///                   Operating System Development Group         |
|    ///   BIX:  msinz      UUNET:  rutgers!cbmvax!mks                 |
|\\\///                                                                |
| \XX/     Quantum Physics:  The Dreams that Stuff is made of.         |
\----------------------------------------------------------------------/

Path: nntpd.lkg.dec.com!rust.zso.dec.com!pa.dec.com!decwrl!mips!swrinde!sdd.hp.com!wuarchive!uunet!mcsun!news.funet.fi!funic!nic!vinsci
From: [email protected] (Leonard Norrgard)
Newsgroups: comp.sys.amiga.misc,comp.unix.amiga
Subject: X-Pert: "50 MHz 68040 card available"
Message-ID: <[email protected]>
Date: 11 Jun 91 11:03:47 GMT
Sender: [email protected] (Leonard Norrgard)
Distribution: comp
Organization: Soft Service, Inc.
Lines: 34
Xref: nntpd.lkg.dec.com comp.sys.amiga.misc:5150 comp.unix.amiga:1147

I just got off the phone with X-Pert Computer Services in Germany, and
to my surprise they said they have a 50 MHz 68040 card available now.
The card comes in two versions, one for the A2000 and one for the
A3000.  The A2000 version comes with 4MB of on-board RAM. Prices are
5900 DM for the A300 version, 6900 DM for the A2000 version. (Address
and phone number below).

Their Visiona 24bit graphics card isn't available yet (though they
showed it at Amiga '91 in Berlin about a month ago), but scheduled for
the end of the year. The version shown at the show didn't have a
hardware cursor, which they said they would include in the next
in-house version of the card. Programming info & libraries are
available now though. (I got mine in Berlin :-)
  For some reason they think that their MEM64 64Mb nibble mode memory
card won't have a market without it, so they plan to release them at
the same time. I suppose they're thinking of image manipulation as the
main market for the Visiona and that is of course memory intesive.
However one would think that it would be good for the Unix systems as
well!
  The person I spoke with said that their TimeHopper external cache
board for the A3000 is available now as well.

	X-Pert Computer Services GmbH
	Weiherwiese 27
	D-6270 Idstein

	Tel. +49-6126-8809-3056  (they know english...,
				  the time zone is GMT/UTC+2)
	Fax. +49-6126-54922

Motorola doesn't ship any 68040's certified for >25mhz, do they?

Given the logic errors in some of the current 040's, and the lack
of Motorola's endorsement of faster clock speeds, it sounds like a guyer
beware situation to me...

    The Dave replied to this. Moto just started shipping 33MHz '040s.
    However, the processor clock on an '040 with a 25MHz bus clock is
    normally 50MHz.
    
    So, the system is 25MHz, and most people would call it a 25MHz '040
    board. But the processor is getting a 50MHz clock, which is enough for
    a marketroid to call it a 50MHz '040 card.
    
    	<mike