603 vs. 603e vs. 604: what's the difference?
speed racer
hey all....
was looking for a bit of advice/ statistics on the differences or
benefits of choosing the 603, 603e, or 604 chip. I realize the 604 is
the best, but how much of an advantage is it to buy a system with a 604
chip now over one with a 603? What's the upgrade path like? And anyone
have hard facts on comparisons?
thanks :)
tm
--
'`^`'~=-.,__,.-=~'`^`'~=-.,__,.-=~'`^`'~=-., \|/ (___) \|/ _,.-=~'`^`
Todd R. Master "God bless those pagans!" @~./'O o`\.~@
cybe...@liii.com -Homer Simpson /__( \___/ )__\ *D'oH!!!*
Web page: http://www.liii.com/~cyberace `\__`U_/'
_,.-=~'`^`'~=-.,__,.-=~'`^`'~=-.,__,.-=~'`^`'~= <____|' ^^`'~=-.,__,.-=
~`'^`'~=-.,__,.-=~'`^`'~=-.,__,.-=~'`^`'~=-.,__,.-==--^'~=-.,__,.-=~'`^`
david miller
> hey all....
>
> was looking for a bit of advice/ statistics on the differences or
> benefits of choosing the 603, 603e, or 604 chip. I realize the 604 is
> the best, but how much of an advantage is it to buy a system with a 604
> chip now over one with a 603? What's the upgrade path like? And anyone
> have hard facts on comparisons?
>
> thanks :)
>
Here is some information I put together to answer a question concerning
the 603 vs. the 603e. This data has been gathered from Motorola's web
site.
603 Performance
SPECint92: 60 at 66 MHz, 75 at 80 MHz *
SPECfp92: 70 at 66 MHz, 85 at 80 MHz *
603e Performance: 100 MHz
SPECint92* 120, SPECfp92* 105
601 Performance:
80 MHz 100 MHz
SPECint92* 85 SPECint92* 105
SPECfp92* 105 SPECfp92* 125
604 Performance
100 MHz 120 MHz 133 MHz
SPECint92* 160 180 200
SPECfp92* 165 180 200
As you can see the 603 is definitely the weak link in the Power PC
scheme. Motorola classifies the 603 as being appropriate for portable
computers vs. even the 601 which they classify as powerful enough for both
desktop and workstation use.
Specifically look at the floating point results. An 80mz 601 is as fast
as a 100mz 603e. The 100mz 604 is 57% faster than the 100mz 603e.
david miller
Steven H. Lee
dmi...@minn.net (david miller) wrote:
> 603e Performance: 100 MHz
> SPECint92* 120, SPECfp92* 105
>
> 601 Performance:
> 80 MHz 100 MHz
> SPECint92* 85 SPECint92* 105
> SPECfp92* 105 SPECfp92* 125
>
> Specifically look at the floating point results. An 80mz 601 is as fast
> as a 100mz 603e.
Not really. Floating point is not terribly important to daily computing;
integer performance is much more important. 603e-based Macs don't perform
as well as 601-based ones MHz-for-MHz because of motherboard design.
Performas and PowerBooks have 32-bit data bus between motherboard and
system memory (vs 64-bit on Power Macintosh models). The onboard video is
much slower, and PowerBook 5300 series has very sloooow hard disks (I
know, because I have one)....
--
Steven H. Lee sh...@cornell.edu
Stephen Jonke
>Specifically look at the floating point results. An 80mz 601 is as fast
>as a 100mz 603e. The 100mz 604 is 57% faster than the 100mz 603e.
True, although the 100MHz 603e has a significantly higher integer
arithmetic rating than both the 100MHz and (obviously) 80Mhz 601, so I
guess it depends on what you use the computer for. For instance, for
games the 603e may actually be better since most games only use integer
arithmetic. That's the theory, anyway, I'm not sure how it works out in
practice....
Steve
--
Seen in recent computer peripheral ad: "User-friendly dip switches!"
Boffin Boy
> hey all....
>
> was looking for a bit of advice/ statistics on the differences or
> benefits of choosing the 603, 603e, or 604 chip.
<...Snip...>
The following is from Motorola I hope it helps settle some of the speed
debate, this of course won't answer questions of bus speeds/memory
interleaving/fast SCSI/yatta,yatta,yatta.
PowerPC 603 Microprocessor
The PowerPC 603 microprocessor is a low-power implementation of the
PowerPC (TM) Reduced
Instruction Set Computer (RISC) architecture. The PowerPC 603
microprocessor offers
workstation-level performance packed into a low-power, low-cost design
ideal for desktop
computers, notebooks, and battery-powered systems.
Superscalar Microprocessor
The PowerPC 603 microprocessor design is superscalar, capable of issuing
three instructions per
clock into five independent execution units:
integer unit
floating-point unit
branch processing unit
load/store unit
system register unit
The ability to execute multiple instructions in parallel, to pipeline
instructions, and the use of simple
instructions with rapid execution times yields maximum efficiency and
throughput for PowerPC 603
systems.
Power Management
The PowerPC 603 microprocessor features a low-power 3.3V design with three
power-saving
modes--nap, doze, and sleep. These user-programmable modes progressively
reduce the power
drawn by the processor.
The PowerPC 603 microprocessor also uses dynamic power management to
selectively activate
functional units as they are needed by the executing instructions. Unused
functional units enter a
low-power state automatically without affecting performance, software
execution, or external
hardware.
Cache and MMU Support
The PowerPC 603 microprocessor has separate 8-Kbyte, physically-addressed
instruction and data
caches. Both caches are two-way set-associative.
The PowerPC 603 microprocessor also contains separate memory management
units (MMUs) for
instructions and data. The MMUs support 4 Petabytes of virtual memory and
4 Gigabytes of physical
memory. Access privileges and memory protection are controlled on block or
page granularities.
Large, 64-entry translation lookaside buffers (TLBs) provide efficient
physical address translation
and support for demand virtual-memory management on both page- and
variable-sized blocks.
PowerPC 603 Microprocessor Block Diagram
Flexible Bus Interface
The PowerPC 603 microprocessor has a selectable 32- or 64-bit data bus and
a 32-bit address bus.
Support is included for burst, split and pipelined transactions. The
interface provides snooping for
data cache coherency. The PowerPC 603 microprocessor maintains MEI
coherency protocol in
hardware, allowing access to system memory for additional caching bus
masters, such as DMA
devices.
PowerPC 603 Major Features
Technology
3.3 Volt implementation
0.5 micron static CMOS technology
83 mm2 - 1.6 million transistors
Specifications
8K instruction and 8K data cache
Superscalar - 3 instructions per clock
On-chip power management
32/64-bit data bus mode
Fully JTAG compatible
Performance
SPECint92: 60 at 66 MHz, 75 at 80 MHz *
SPECfp92: 70 at 66 MHz, 85 at 80 MHz *
Power Consumption
Full Operation - 3 watts maximum at 80 MHz
Low Power Modes - lap, doze, sleep
Packaging
240 CQFP/256 BGA
*Estimated performance
The PowerPC 603e Microprocessor
The PowerPC 603e microprocessor is a low-power implementation of the
PowerPC (TM) Reduced
Instruction Set Computer (RISC) architecture. The PowerPC 603e
microprocessor offers
workstation-level performance packed into a low-power, low-cost design
ideal for desktop
computers, notebooks, and battery-powered systems.
Superscalar Microprocessor
The PowerPC 603e microprocessor design is superscalar, capable of issuing
three instructions per
clock cycle into five independent execution units:
Integer unit
Floating-point unit
Branch processing unit
Load/store unit
System register unit
The ability to execute multiple instructions in parallel, to pipeline
instructions, and the use of simple
instructions with rapid execution times yields maximum efficiency and
throughput for PowerPC 603e
systems.
Power Management
The PowerPC 603e microprocessor features a low-power 3.3-volt design with
three power-saving
modes-doze, nap and sleep. These user-programmable modes progressively
reduce the power drawn
by the processor.
The PowerPC 603e microprocessor also uses dynamic power management to
selectively activate
functional units as they are needed by the executing instructions. Unused
functional units enter a
low-power state automatically without affecting performance, software
execution, or external
hardware.
Cache and MMU Support
The PowerPC 603e microprocessor has separate 16-Kbyte,
physically-addressed instruction and data
caches. Both caches are four-way set-associative.
The PowerPC 603e microprocessor also contains separate memory management
units (MMUs) for
instructions and data. The MMUs support 4 Petabytes (252) of virtual
memory and 4 Gigabytes
(232) of physical memory. Access privileges and memory protection are
controlled on block or page
granularities. Large, 64-entry translation lookaside buffers (TLBs)
provide efficient physical address
translation and support for demand virtual-memory management on both page-
and variable-sized
blocks.
Flexible Bus Interface
The PowerPC 603e microprocessor has a selectable 32- or 64-bit data bus
and a 32-bit address bus.
Support is included for burst, split and pipelined transactions. The
interface provides snooping for
data cache coherency. The PowerPC 603e microprocessor maintains MEI
coherency protocol in
hardware, allowing access to system memory for additional caching bus
masters, such as DMA
devices.
PowerPC 603e Microprocessor Major FeaturesSpecifications
16-Kbyte instruction and 16-Kbyte data caches
Superscalar-3 instructions per clock cycle
On-chip power management
32/64-bit data bus mode
Fully JTAG-compliant
Performance: 100 MHz
SPECint92* 120
SPECfp92* 105
Power Consumption
Full operation-3.5 watts maximum at 100 MHz
Low Power Modes-doze, nap, sleep
Technology
3.3-volt implementation
0.5-micron static CMOS technology
98 mm2 die size
2.6 million transistors
Packaging
240 CQFP/256 BGA
*Estimated performance.
PowerPC 604 Microprocessor
The PowerPC 604 microprocessor is a 32-bit implementation of the
PowerPC(TM) Reduced
Instruction Set Computer (RISC) architecture. The PowerPC 604
microprocessor provides high
levels of performance for desktop, workstation, and symmetric
multiprocessing computer systems
and is software and bus compatible with the PowerPC 601(TM) and PowerPC 603(TM)
microprocessor family.
Superscalar Microprocessor
The PowerPC 604 microprocessor is a superscalar design capable of issuing
four instructions per
clock cycle to six independent execution units, including:
two single-cycle integer units
one multi-cycle integer unit
branch processing unit
load/store unit
floating-point unit
The PowerPC 604 microprocessor uses dynamic branch prediction to improve
the accuracy of
instruction prefetching. Dynamic branch prediction combined with the
ability to speculatively execute
through two unresolved branches, minimizes pipeline stalls and allows the
multiple execution units to
provide a high level of efficiency and throughput. While the PowerPC 604
microprocessor supports
out-of-order execution, in-order instruction completion assures precise
exceptions.
Cache and MMU Support
The PowerPC 604 microprocessor has separate 16-Kbyte, physically addressed
instruction and data
caches. Both caches are four-way set associative, and provide parity
checking at the byte level.
The PowerPC 604 microprocessor also contains separate memory management
units (MMUs) for
instructions and data. The MMU's support up to 4 Petabytes (252) of
virtual memory and 4
Gigabytes (232) of physical memory. Access privileges and memory
protection are controlled on
block or page granularities. Large, 128-entry translation lookaside
buffers (TLBs) provide efficient
physical address translation and support for demand virtual-memory
management for both page- and
variable-sized blocks.
PowerPC 604 Microprocessor Block Diagram
Flexible Bus Support
The PowerPC 604 microprocessor has a high performance 64-bit data bus and
a separate 32-bit
address bus. The interface protocol allows multiple masters to access
system resources through a
central arbiter. Additionally, on-chip snooping logic maintains cache
coherency in multiprocessor
applications.
PowerPC 604 Major Features
Specifications
16-Kbyte instruction cache and 16-Kbyte data cache
Superscalar--four instructions per clock cycle
Dynamic branch prediction
Multiple integer units
64-bit data bus
Fully JTAG compatible
Performance
100 MHz 120 MHz 133 MHz
SPECint92* 160 180 200
SPECfp92* 165 180 200
Power Consumption
Full operation; 100 MHz 120 MHz 133 MHz
watts typical: 14 16.5 17.5
Technology
3.3 volt implementation
0.5 micron static CMOS technology
197 mm2 die size
3.6 million transistors
Packaging
304 CQFP/256 BGA
*Estimated performance.
--
" Do not mistake composure for ease. "
Tuvok.
Ian Russell Ollmann
On Mon, 22 Apr 1996, Stephen Jonke wrote:
> In article <dmiller-2104...@dialup-211.minn.net>,
> dmi...@minn.net (david miller) wrote:
>
> >Specifically look at the floating point results. An 80mz 601 is as fast
> >as a 100mz 603e. The 100mz 604 is 57% faster than the 100mz 603e.
>
> True, although the 100MHz 603e has a significantly higher integer
> arithmetic rating than both the 100MHz and (obviously) 80Mhz 601, so I
> guess it depends on what you use the computer for. For instance, for
> games the 603e may actually be better since most games only use integer
> arithmetic. That's the theory, anyway, I'm not sure how it works out in
> practice....
In practice, the 603e systems run with a 32 bit memroy bus and are always
slower than their 601 counterparts with a 64 bit memory bus for pretty
much everything. I am sure that one day, Apple will get around to
releasing 603e based machines with decent hardware supporting the CPU, but
until they do, don't expect them to perform up to their SpecInt's.
Ian Ollmann
Benjamin Smith
used on the Performas. The 601 on the PowerMacs is quite fast, and the
604 gives superbly rapid performance on most programs (except Word).
Ben S.
Adam Toner
>
> In article <dmiller-2104...@dialup-211.minn.net>,
> dmi...@minn.net (david miller) wrote:
>
> >Specifically look at the floating point results. An 80mz 601 is as fast
> >as a 100mz 603e. The 100mz 604 is 57% faster than the 100mz 603e.
>
> True, although the 100MHz 603e has a significantly higher integer
> arithmetic rating than both the 100MHz and (obviously) 80Mhz 601, so I
> guess it depends on what you use the computer for. For instance, for
> games the 603e may actually be better since most games only use integer
> arithmetic. That's the theory, anyway, I'm not sure how it works out in
> practice....
>
>
> --
> Seen in recent computer peripheral ad: "User-friendly dip switches!"
What would be faster day-to-day, a 75 MHz 603 (a la the Performa 5215cd)
or a 66 MHz 601 (a la the Performa 6116cd)? My girlfriend is considering
these. Actually, I would appreciate comments on both systems in general.
--
Adam Toner
ato...@aug.com
Have a day.
Ian Russell Ollmann
On Wed, 24 Apr 1996, Adam Toner wrote:
>
> What would be faster day-to-day, a 75 MHz 603 (a la the Performa 5215cd)
> or a 66 MHz 601 (a la the Performa 6116cd)? My girlfriend is considering
> these. Actually, I would appreciate comments on both systems in general.
The 75 MHz 603 is about equivalent to a 55 MHz 601 after placing it into a
Performa motherboard with a 32 bit memory bus. The 61xx should be faster.
However, that is not to say that she should buy either. A $900 PowerMac
7200/75 (Compu-America 800-533-9005) will be much faster (especially after
you add an L2 cache), has much better video, is much more expandable, has
built in ethernet, can be upgraded to a better machine, etc., etc., etc.,
Simply put,i t is a much better machine in just about every single
important way except the hard drive may be a wee bit smaller than the
6116. This can be remedied at your leisure if it turns out to be a
problem. You also get to pick out a decent 15" multisynch monitor and a
decent keyboard to go with it. Any of these three machines woudl really
benefit from a 16 MB DIMM/SIMM.
Ian Ollmann
Mike McHugh
> Here is some information I put together to answer a question concerning
> the 603 vs. the 603e. This data has been gathered from Motorola's web
> site.
>
> 604 Performance
> 100 MHz 120 MHz 133 MHz
> SPECint92* 160 180 200
> SPECfp92* 165 180 200
Does anyone know how the 150MHz 604 rates?
Mike
---------------------------------------------------------------------
"Faith and Trust can win the day, in spite of all your losing"
-Led Zeppelin
Check out my web page:
http://www.ici.net/cust_pages/mchugh/
Anthony Siino
I just bought a 6214 (75mhz 603) because it was a deal ($1335 (bundle)).
I was also looking at the 6116 performa. I bought the 6214 bundle because
it came with a 4X CDROM, you don't have to install ram in pairs so the
4meg simm I already had fit nicely, it had a 1.0G hard drive, and came
with the 15" monitor instead of the 14" monitor the 6116 comes with. The
only advantage the 6116 had was that it can be clock chipped for more
speed and a L2 cache can be added for even more speed. I don't need the
extra speed and I thought the 4XCDROM, 15" monitor, & larger hard drive
were well worth the trade off in speed.
When I bought my machine though, the price cuts in the 7200s had not taken
effect and at the time there were Performa printer rebates so I also got a
Stylewriter for $50 which sweetened the deal. If I was doing it now I'd
look at the 7200/75 or 90. A 7200/75 with the same monitor and keyboard
as my 6214 bundle would be $1450. It's probably worth the extra $115.
Anthony
In article <317DAC...@aug.com>, ato...@aug.com wrote:
*
*What would be faster day-to-day, a 75 MHz 603 (a la the Performa 5215cd)
*or a 66 MHz 601 (a la the Performa 6116cd)? My girlfriend is considering
*these. Actually, I would appreciate comments on both systems in general.
*--
*Adam Toner
*ato...@aug.com
*Have a day.
David Smith
In article <Pine.SGI.3.93.960423193940.5389E-100000@wong>,
Seeing as how (I belive) that the 603 and 603e have 32 bit INTERFACES
while the 601 and 604 have 64 bit interfaces it would be very hard, not
to mention useless, to put a 64 bit bus on a 603. The 603 series is
designed to low power and/or low cost uses, it would kinda of defeat use
of the 603 in desktop use to put a "decent" and therefor expensive box
around it when you could put a 604 in it.
--
David A. Smith | Baylor College of Medicine
Systems Services | Enterprise Services, MEDT1103
(713) 798-8469 | One Baylor Plaza
Internet: da...@bcm.tmc.edu | Houston, TX 77030
William Raphael Hix
David Smith (da...@bcm.tmc.edu) wrote:
: >On Mon, 22 Apr 1996, Stephen Jonke wrote:
: >In practice, the 603e systems run with a 32 bit memroy bus and are always
: >slower than their 601 counterparts with a 64 bit memory bus for pretty
: >much everything. I am sure that one day, Apple will get around to
: >releasing 603e based machines with decent hardware supporting the CPU, but
: >until they do, don't expect them to perform up to their SpecInt's.
:
: Seeing as how (I belive) that the 603 and 603e have 32 bit INTERFACES
: while the 601 and 604 have 64 bit interfaces it would be very hard, not
: to mention useless, to put a 64 bit bus on a 603. The 603 series is
: designed to low power and/or low cost uses, it would kinda of defeat use
: of the 603 in desktop use to put a "decent" and therefor expensive box
: around it when you could put a 604 in it.
Both the 603 and 603e can support either 32 bit or 64 bit memory buses.
Check out Motorlola's PPC site <http://www.mot.com/SPS/PowerPC/> for
more info.
Raph
------------------------------------------------------------------------
William Raphael Hix Department of Astronomy
ra...@astro.as.utexas.edu University of Texas
Voice: (512) 471-3412 R.L. Moore Hall
FAX: (512) 471-6016 Austin TX 78712
WWW: http://tycho.as.utexas.edu/~raph Room 17.210
------------------------------------------------------------------------
Ian Russell Ollmann
On 26 Apr 1996, William Raphael Hix wrote:
> David Smith (da...@bcm.tmc.edu) wrote:
> : >On Mon, 22 Apr 1996, Stephen Jonke wrote:
> : >In practice, the 603e systems run with a 32 bit memroy bus and are always
> : >slower than their 601 counterparts with a 64 bit memory bus for pretty
> : >much everything. I am sure that one day, Apple will get around to
> : >releasing 603e based machines with decent hardware supporting the CPU, but
> : >until they do, don't expect them to perform up to their SpecInt's.
> :
> : Seeing as how (I belive) that the 603 and 603e have 32 bit INTERFACES
> : while the 601 and 604 have 64 bit interfaces it would be very hard, not
> : to mention useless, to put a 64 bit bus on a 603. The 603 series is
> : designed to low power and/or low cost uses, it would kinda of defeat use
> : of the 603 in desktop use to put a "decent" and therefor expensive box
> : around it when you could put a 604 in it.
>
> Both the 603 and 603e can support either 32 bit or 64 bit memory buses.
> Check out Motorlola's PPC site <http://www.mot.com/SPS/PowerPC/> for
> more info.
Actually, I wrote the part above credited to Stephen Jonke. It was his
article that I was replying to. I am not sure exactly why Apple decided to
ship Performas with 32 bit memory busses, but a number of us have
privately speculated that it was so that the user wouldn't have to pair
the DIMMs in the two RAM slots given to the Performa motherboard. It would
have been nuts for Apple to sell a machine with two SIMM slots in which
both SIMMs had to be paired in order for it to work (i.e. Apple has to
ship the computer with both SIMM slots full and no place to add additional
RAM.) The two solutions would have been to either add another pair of SIMM
slots or solder 8 MB on the motherboard. In the end, the performas might
have sold better with a wider bus. They would have been less handicapped.
Even the 6300/100 with an L2 cache in it and everything is only as fast as
a 7200/90 w/o L2.
Ian Ollmann
Edo van der Kolk
In article <317DAC...@aug.com>,
Stephen Jonke wrote:
> True, although the 100MHz 603e has a significantly higher integer
> arithmetic rating than both the 100MHz and (obviously) 80Mhz 601, so I
> guess it depends on what you use the computer for. For instance, for
> games the 603e may actually be better since most games only use integer
> arithmetic. That's the theory, anyway, I'm not sure how it works out in
> practice....
Hello,
I have a performa 5300 and I like to play games. It's cpu and fpu
performance is better than that of a 8100 80mhz, but the video performance
is just a little worse than that of a 601:
from: http://www.macintouch.com/
ACE MORGAN, "Mac Detective"
The First Annual LOSER* Award
*pronounced: Loo-hoo-za-her
IMPORTANT NOTE BEFORE YOU READ ANY FURTHER: The buyers of the following
list of products are NOT losers in the sense that they have no sense but in
the sense that they have been ripped off by Apple, the third party vendors
and the computer press! It's a scam or a blunder or a conspiracy or a
crime. Whatever it is, it's bad news.
(And the envelope please.....)
1996 Loser Award: Unsuspecting buyers of any 603 based Power Mac.
For some reason, when then engineers at Apple designed this flavor of
PowerMac, they did something to hamstring the graphics. I ran benchmark
tests on 603 based PowerMacs (52xx, 5300, & 62xx) along side 601 based
PowerMacs (61xx, 7100, 7200, 7500). With matching cpu speeds (MHz), the 601
machines ran graphic functions quite a bit faster than their 603 based
counterparts!
You may not be aware of the difference unless you are using your 603 based
PowerMac next to a friend with a 601 or 604 based PowerMac. And believe me,
the difference is noticeable even without a stopwatch. In fact, a 6100
running 60MHz out-performed a 6230 running 75MHz by more than 100%! Add
level 2 cache and the 6100's advantage jumps to almost 200%!!
How can this be? I'm not completely sure but it's the fact, Jack. MacWorld
Magazine has contributed to the false impression that all PowerMacs are
created equal since their 'summary' performance rating includes "cpu
intensive, math intensive, and disk intensive" but EXCLUDES "graphic
intensive."
Why is "graphic intensive" important? A lot of the efficiency of your
Macintosh is related to how quickly the screen responds to changes. Whether
you are scrolling through the words a Word document or turning pages in
PageMaker or rotating photos in Photoshop or doing barrel rolls in Chuck
Yeager's Air Combat, you want the screen related functions to respond
smoothly and quickly. Whether you realize it or not, your computer's
ability to respond is independent of the cpu speed, or "megahertz," if you
will. It depends on the design of the graphics circuitry on the built-in
video or the addon video board inserted in the NUBUS or PCI slot.
If you are a game fanatic, you are going to be gravely disappointed when
you run Marathon or Havok on your 603 based machine. Jerky, jerky, jerky
and slow. If you are running QuickTime or QuickDraw 3D or video clips or
any drawing program, you will be majorly bummed. And don't forget the text
scrolling issue with MS Word and ClarisWorks.
Here's the data I collected from just my Speedometer tests.
Speedometer 8 bit Graphics Speed Scores
DESKTOP POWERMACS (bigger number is better)
MODEL RATING
5215/75 .90
5300cs/100 .63
6218/75 .78
6230/75 .79
6220/75 .90
IIci "hot rod"(1) 1.44
6100/60 1.76
7200/75 1.97
7100/80 2.05
6100/66(2) 2.13
6100/60(2) 2.17
7200/90 2.33
7100/80(2) 2.37
7500 2.85
7500(2) 3.11
(1) with Daystar 601/100 accelerator & Radius 24XP video card
(2) with level 2 cache
SO WHAT CAN I DO ABOUT IT?
If you just bought one of the 603 based PowerMacs at Computer City, take it
back. You have 30 days. If it's been more than 30 days, you have my
condolences. Try a letter to Apple. Complain that it's a lemon and you want
to trade for ANY 601 based machine.
If you are on the verge of buying a bargain price 603 based machine, you
might want to think twice. "But those prices are so tempting." I know but
here are some bargains on 601 based PowerMacs:
Apple has a bundle offer for parents of school kids starting April 15th
through Sears. It's a Performa 6100 for $1299 including 4X CD, 14" display,
AppleDesign keyboard, collection of software including ClarisWorks, and b/w
StyleWriter.
MacWorks is selling a factory refurbished Performa 6100/60CD for $999 (90
day warranty). If like most Performas, it includes the keyboard and
display, all you need to add is a color StyleWriter for $299 (Costco) and
you have a complete PowerMac 601 for $1298.
You can get a NEW Performa 6116 (60MHz PowerPC 601) with 2x CD, Apple 14
inch color display, AppleDesign keyboard, 14.4 modem and Color StyleWriter
printer for $1698 from MacMall catalog store.
Here's the hotest deal on a state of the art Mac from Mac Sale
International. This is on special because is overstocked on these and is
planning a faster version. It's a great machine that can be expanded &
accelerated later on: it's a 7200 (75MHz PowerPC 601) with 4x CD, 3 PCI
slots, 16 bit stereo sound, accelerated graphics, built-in Ethernet, 1MB
VRAM to which you can add 14 inch Apple color display , Apple Design
keyboard, and Apple color StyleWriter (from Costco) for total system price
of $1646. (Price does not include modem or software)
Take my advice and stay with 601 based PowerMacs (or 604 based if you can
afford them). One last suggestion: do your eyes an favor and get the Sony
15" multiscan display with a .25 dot pitch (versus .29 on the cheapo 14"
Apple display). Text and images are much sharper. It runs an additional
$200 and worth every penny.
Geoffrey Gautcher
Stephen Jonke wrote:
>
> In article <dmiller-2104...@dialup-211.minn.net>,
> dmi...@minn.net (david miller) wrote:
>
> >as a 100mz 603e. The 100mz 604 is 57% faster than the 100mz 603e.
>
> arithmetic rating than both the 100MHz and (obviously) 80Mhz 601, so I
> guess it depends on what you use the computer for. For instance, for
> games the 603e may actually be better since most games only use integer
> arithmetic. That's the theory, anyway, I'm not sure how it works out in
> practice....
>
>
> --
> Seen in recent computer peripheral ad: "User-friendly dip switches!"
You are correct; Macs using the 603e chip are slower--Hz for
Hz. Also, according to an article posted on the "Macintouch"
page (http://204.107.232.226/slow603.html) by Rob Morgan,
the 601 does a strangle hold on graphics. The article says
it all.
--
In Christ's Peace,
The "Mac Man"
Geoffrey Gautcher
Gladstone, Michigan
geo...@mich.com
http://www.grfn.org/~geoff/
Brian Cole
evd...@dataweb.nl (Edo van der Kolk) writes:
]
] ACE MORGAN, "Mac Detective"
] The First Annual LOSER* Award
]
] *pronounced: Loo-hoo-za-her
]
] IMPORTANT NOTE BEFORE YOU READ ANY FURTHER: The buyers of the following
] list of products are NOT losers in the sense that they have no sense but in
] the sense that they have been ripped off by Apple, the third party vendors
] and the computer press! It's a scam or a blunder or a conspiracy or a
] crime. Whatever it is, it's bad news.
]
] (And the envelope please.....)
] 1996 Loser Award: Unsuspecting buyers of any 603 based Power Mac.
I don't remember _how_ I knew, but I knew that the 5200/6200/5300/6300
machines had narrow data busses from the beginning. It wasn't a big
secret or anything. At least with the level-two cache it wasn't as
crippling as it could have been.
When the Classic II "replaced" the SE/30 people were disappointed in the
narrower bus, but people who bought one anyway didn't complain about scams
and conspiracies.
Now the IIvx (which had a _slower_ bus, not narrower, but it's a 50% reduction
in bandwidth either way) is a slightly different story. The "crippled" IIvx
bus _was_ a surprise that had not been disclosed in advance.
brian
PS. All this is from memory, don't don't take any of this as gospel without
double-checking first.
--
~~~~~~~~~~~~~~~~~ _______ _____ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Brian A. Cole |_ _ | | ___| Visit Brian's Repository of
U. Wisconsin | | | |__| |___ Macintosh Information at
t...@cs.wisc.edu |_| |__________| http://www.cs.wisc.edu/~tuc/mac/
Brian Cole
da...@bcm.tmc.edu (David Smith) writes:
]
] Seeing as how (I belive) that the 603 and 603e have 32 bit INTERFACES
] to mention useless, to put a 64 bit bus on a 603. The 603 series is
] designed to low power and/or low cost uses, it would kinda of defeat use
] of the 603 in desktop use to put a "decent" and therefor expensive box
] around it when you could put a 604 in it.
Well I'm not sure what you mean by "interfaces", especially when in all caps,
but the 603 and 603e support a 64-bit data bus just fine. It would be no
"harder" to put a 64-bit on a 603 than to put one on a 601. As for "useless",
take a look at a BeBox (http://www.be.com/) and see if you still think so.
You might be surprised how fast a 603 chip can be when it's not saddled with
legacy code that it must emulate.
geo...@mich.com (Geoffrey Gautcher) writes:
}
} You are correct; Macs using the 603e chip are slower--Hz for
} Hz. Also, according to an article posted on the "Macintouch"
} page (http://204.107.232.226/slow603.html) by Rob Morgan,
} the 601 does a strangle hold on graphics. The article says
} it all.
You are correct that existing Macs using the 603e chip are slower
Hz for Hz than existing Macs using the 601 chip, but that's not
a fair comparison of the chips themselves. If you put the two
chips on near-identical motherboards (same CPU speed, same bus
width, same bus clock divider, same level-two cache, same peripheral
bus, etc.) you will see the 603e edge out the 601 in all benchemarks
except for floating-point-intensive ones.
brian
David Smith
On Fri, 26 Apr 1996, Ian Russell Ollmann wrote:
[clip]
> Actually, I wrote the part above credited to Stephen Jonke. It was his
> article that I was replying to. I am not sure exactly why Apple decided to
> ship Performas with 32 bit memory busses, but a number of us have
> privately speculated that it was so that the user wouldn't have to pair
> the DIMMs in the two RAM slots given to the Performa motherboard. It would
> have been nuts for Apple to sell a machine with two SIMM slots in which
> both SIMMs had to be paired in order for it to work (i.e. Apple has to
> ship the computer with both SIMM slots full and no place to add additional
> RAM.) The two solutions would have been to either add another pair of SIMM
> slots or solder 8 MB on the motherboard. In the end, the performas might
> have sold better with a wider bus. They would have been less handicapped.
> Even the 6300/100 with an L2 cache in it and everything is only as fast as
> a 7200/90 w/o L2.
Sorry to have misquoted you. I must of missed a > somewhere.
DIMMs (the 168 pin as in the PCI mac's) are 64 bit chips, they would
only need to be upgraded in singles, 72 pin SIMMs are 32 bit and would
need to be upgraded in pairs, as on the nubus PPC "macintoshes" as
opposed to Performas.
Also my understanding is that a 64 bit motherboard costs quite a bit more to
make, and to make go fast, than a 32 bit bus, since you need twice as
many path's, and all the ASIC's on the bus have to deal with 64 bits of
data, along with things like having to upgrade the memory in sets (uggh,
anyone seen the new Sun UltraSparc servers? The have a 512 bit+parity data
path, you have to upgrade the 168pin DIMMs in sets of 8!!!).
I don't know how much more it costs to double the size of the
motherboard, but would the Performas have been nearly as popular if they
had cost say 20% more? Apple was aiming the Performas toward the lower
education market and intro home market, where cost is a major factor so a
lot of decisions were cost related. It also differinates them from the
"other" macintoshes.
BTW wern't the early PPC Performas (ex 6115) direct copies of their "big
brothers" with the 64 bit bus?
Steven Kan
David Smith wrote:
>
[big snip]
> BTW wern't the early PPC Performas (ex 6115) direct copies of their "big
> brothers" with the 64 bit bus?
Yup. Got one right here. The CPU of the Performa 611x computers is
almost exactly the same as that of the original 6100/60, except that
it ships without the L2 cache but with a lot more software.
--
A rich man who hailed from Seattle,
Wrote Win95 to do battle.
But Mac users pity
The masses not witty
Enough to know Wintel's for cattle.
Steven "Rocket Man" Kan
http://www-leland.stanford.edu/~stevenk
Never refuse a breath mint.