Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Quick! Everyone go look at the Apple web site

1 view
Skip to first unread message

Chris

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
Loads of new and extremely cool stuff!!

Dual CPU G4's
8" cube PowerMac G4 entire computer!
Apple "Pro" USB keyboard (no more small keys)
Apple Optical Mouse (I'm almost positive it's a slightly modified
Microsoft IntelliMouse Explorer)
5 New iMac colors/configs
Gigabit Ethernet option (been around for a while, but oh well)


Alright *now* who's ready to finish MI multiprocessor support :-)~~~


Chris


John Klos

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
> Dual CPU G4's

Two cpu packages on a daughercard. Not ZIF compatible. Not a dual-core
G4. 133 mhz memory bus. Cardbus for airport. 4x AGP. Gigabit ethernet
standard. Same price for dual processor configs as was the previous single
processor configs ($2500 for dual 450, $3500 for dual 500 mhz (USD)).

> 8" cube PowerMac G4 entire computer!

100 mhz memory bus, full sized DIMMs (three slots), 2x AGP with 16 meg ATI
(same as older G3 / G4), 10/100-base-T, gigabit ethernet or airport is
optional. $1800 USD.

> Apple "Pro" USB keyboard (no more small keys)

With sound volume buttons and an "eject" button, and programmable function
keys.

> Apple Optical Mouse (I'm almost positive it's a slightly modified
> Microsoft IntelliMouse Explorer)

Not an Intellimouse at all. Totally optical, no buttons. Button function
is performed by pressing down on the top end of the mouse. Pressure
sensitivity is adjustable.

> 5 New iMac colors/configs
> Gigabit Ethernet option (been around for a while, but oh well)

Standard in all new G4s... are drivers in NetBSD?

> Alright *now* who's ready to finish MI multiprocessor support :-)~~~

The dual G4s run with OS 9, so they shouldn't be significantly different
than the dual/quad 604e machines.

Speaking of G4s: how far along is gcc with Altivec in NetBSD? I met an
interesting fellow from Yellow Dog Linux (or whoever their parent company
is) who gave me some interesting info on both Altivec and on booting from
MacOS. I'll be sending him an email pretty soon.

John Klos


David A. Gatwood

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
On Thu, 20 Jul 2000, Chris wrote:

> >> Gigabit Ethernet option (been around for a while, but oh well)
> >
> > Standard in all new G4s... are drivers in NetBSD?
>

> They probably are, but I would doubt they are working in macppc. I
> haven't heard of anyone here having a gigabit card to even try. Heck I think
> half of the 100BaseT drivers are still flaky/don't work on macppc.

Well, if you have ethernet at all on the current AGP G4s, I'm told that
ethernet should also work on the new ones, or nearly so. My sources say
it's the same chip, just with different support hardware to enable the
gigabit part of the chip. Note that I have no way to verify the accuracy
of that info.


> True, but uh we can't use the other CPU(s) for anything significant in
> NetBSD other than what is cutting edge development in -current. A Quad or 8
> CPU 604e would make for some very mean BeOS power though :)

Heh. A quad-G4 would be pretty sweet. :-)


> As far as Airport goes, isn't it just a matter of bringing the cardbus
> driver from i386 into macppc? Anyone know if the WaveLan driver will run an
> Airport card?

Dunno. Should be pretty close, in theory....


David

---------------------------------------------------------------------
A brief Haiku:

Microsoft is bad.
It seems secure at first glance.
Then you read your mail.


Chris

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
on 7/20/00 12:53 AM, John Klos at jo...@klos.com wrote something like:

>> Apple "Pro" USB keyboard (no more small keys)
>
> With sound volume buttons and an "eject" button, and programmable function
> keys.

I still want them to bring back a better incarnation of the Apple
Adjustable Keyboard (the anoying tiny buttons on the numeric keypad are very
bothersome), I'd be using the one off my Quadra if I could fit it on this
tiny desk heh.

>> Apple Optical Mouse (I'm almost positive it's a slightly modified
>> Microsoft IntelliMouse Explorer)
>
> Not an Intellimouse at all. Totally optical, no buttons. Button function
> is performed by pressing down on the top end of the mouse. Pressure
> sensitivity is adjustable.

The Intellimouse Explorer is optical, it's the next step down from the
IntelliEye (two less buttons I think) this one has four (plus scroll and
scoll button) and the Eye had 6. Pressure sensitive buttons sounds very
cool, no more clicking noises!

>> Gigabit Ethernet option (been around for a while, but oh well)
>
> Standard in all new G4s... are drivers in NetBSD?

They probably are, but I would doubt they are working in macppc. I
haven't heard of anyone here having a gigabit card to even try. Heck I think
half of the 100BaseT drivers are still flaky/don't work on macppc.

>> Alright *now* who's ready to finish MI multiprocessor support :-)~~~
>
> The dual G4s run with OS 9, so they shouldn't be significantly different
> than the dual/quad 604e machines.

True, but uh we can't use the other CPU(s) for anything significant in


NetBSD other than what is cutting edge development in -current. A Quad or 8
CPU 604e would make for some very mean BeOS power though :)

As far as Airport goes, isn't it just a matter of bringing the cardbus


driver from i386 into macppc? Anyone know if the WaveLan driver will run an
Airport card?

Just curious,


Chris


Matthew Reilly

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
on 7/20/00 1:51 AM, David A. Gatwood at dgat...@deepspace.mklinux.org
wrote:

> On Thu, 20 Jul 2000, Chris wrote:
>

>> As far as Airport goes, isn't it just a matter of bringing the cardbus
>> driver from i386 into macppc? Anyone know if the WaveLan driver will run an
>> Airport card?
>

> Dunno. Should be pretty close, in theory....

It's my understanding that the Airport interface is actually some sort of
ATA bus, not PC card despite the similar form factor. I could be wrong
through.

cheers,

Matthew


Henry B. Hotz

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to

The card is identical to what's sold for PC's. I don't think there
is a PC standard for ATA-in-PC-card-form-factor so I don't think so.
I could be wrong though.


Signature failed Preliminary Design Review.
Feasibility of a new signature is currently being evaluated.
h.b....@jpl.nasa.gov, or hbh...@oxy.edu

Matthew Reilly

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to

This URL at Apple show the block diagram of the iBook and the Airport card
is shown on the same ATA bus at the CD-ROM.

http://developer.apple.com/techpubs/hardware/Developer_Notes/Macintosh_CPUs-
G3/iBook/index.html

The diagram of the PowerBook shows it coming off the Key Largo chip and not
the Cardbus bridge but doesn't say anything about ATA.

http://developer.apple.com/techpubs/hardware/Developer_Notes/Macintosh_CPUs-
G3/PowerBook/index.html

Is the PC card spec close to or based on ATA? It makes me wonder, it seems
strange for Apple to go this proprietary route for seemingly little gain.
But hey, we are talking about Apple here.

cheers,

Matthew


Henry B. Hotz

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
At 5:19 PM -0400 7/20/00, Matthew Reilly wrote:
>Is the PC card spec close to or based on ATA? It makes me wonder, it seems
>strange for Apple to go this proprietary route for seemingly little gain.
>But hey, we are talking about Apple here.

I can only say that there are url's which describe how to use the
full-strength encryption card sold for PC's with Mac's in place of
the export-encryption-only version that Apple sells. You can even
buy a PC card (looks like it anyway) and replace the PC card inside
the AirPort. It even works with Apple software except that you have
to put the keys in on a PC. They are a standard product from Lucent.

Matthew Reilly

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to

Oh OK, I was confused. I thought you were talking about the card that went
inside the macs. The Airport base station has a Lucen Wavelan PC card inside
plugged into some sort of single chip AMD 486 that handles the NAT and DHCP.

cheers,

Matthew

Henry B. Hotz

unread,
Jul 20, 2000, 3:00:00 AM7/20/00
to
At 5:59 PM -0400 7/20/00, Matthew Reilly wrote:
>on 7/20/00 5:24 PM, Henry B. Hotz at ho...@jpl.nasa.gov wrote:
>
> > At 5:19 PM -0400 7/20/00, Matthew Reilly wrote:
> >> Is the PC card spec close to or based on ATA? It makes me wonder, it seems
> >> strange for Apple to go this proprietary route for seemingly little gain.
> >> But hey, we are talking about Apple here.
> >
> > I can only say that there are url's which describe how to use the
> > full-strength encryption card sold for PC's with Mac's in place of
> > the export-encryption-only version that Apple sells. You can even
> > buy a PC card (looks like it anyway) and replace the PC card inside
> > the AirPort. It even works with Apple software except that you have
> > to put the keys in on a PC. They are a standard product from Lucent.
> >
>Oh OK, I was confused. I thought you were talking about the card that went
>inside the macs. The Airport base station has a Lucen Wavelan PC card inside
>plugged into some sort of single chip AMD 486 that handles the NAT and DHCP.

Perhaps we were both confused because I thought the cards were the
same in that respect.

David A. Gatwood

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On Thu, 20 Jul 2000, Matthew Reilly wrote:

> It's my understanding that the Airport interface is actually some sort of
> ATA bus, not PC card despite the similar form factor. I could be wrong
> through.

Nope. It's a PCMCIA card. They made it without the antenna so you
couldn't buy the (much cheaper) airport cards and stick them in a normal
machine. Similarly, the version in the base stations is modified to have
a different pin order for the same reason, IIRC.

David A. Gatwood

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On Thu, 20 Jul 2000, Matthew Reilly wrote:

> This URL at Apple show the block diagram of the iBook and the Airport card
> is shown on the same ATA bus at the CD-ROM.
>
> http://developer.apple.com/techpubs/hardware/Developer_Notes/Macintosh_CPUs-
> G3/iBook/index.html
>
> The diagram of the PowerBook shows it coming off the Key Largo chip and not
> the Cardbus bridge but doesn't say anything about ATA.

They want you to think it is drastically different because otherwise
they'd have to charge as much for the Airport as the other ? 802.11 ?
boards. IIRC, it was a licensing term or some such.... At least that's
what I keep hearing from various sources.

David A. Gatwood

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On Thu, 20 Jul 2000, Matthew Reilly wrote:

> Oh OK, I was confused. I thought you were talking about the card that went
> inside the macs. The Airport base station has a Lucen Wavelan PC card inside
> plugged into some sort of single chip AMD 486 that handles the NAT and DHCP.

Yeah. Like I said in another message, that one's basically PCMCIA, but
they made some slight changes to the spec so it won't work with a
computer... either changing pin order or something on that scale....

Bill Studenmund

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On Thu, 20 Jul 2000, John Klos wrote:

Q> > Alright *now* who's ready to finish MI multiprocessor support


:-)~~~
>
> The dual G4s run with OS 9, so they shouldn't be significantly different
> than the dual/quad 604e machines.

We have spin-up code for the 604e machines (I think I saw tsubai put that
in). Adding spin-up code for these shouldn't be too hard.

> Speaking of G4s: how far along is gcc with Altivec in NetBSD? I met an
> interesting fellow from Yellow Dog Linux (or whoever their parent company
> is) who gave me some interesting info on both Altivec and on booting from
> MacOS. I'll be sending him an email pretty soon.

Apple has patches to gcc to support Altivec. I'm not sure if they've made
it out yet or not.

Take care,

Bill


Nathan J. Williams

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
<wrst...@zembu.com> (Bill Studenmund) writes:

> Apple has patches to gcc to support Altivec. I'm not sure if they've made
> it out yet or not.
>

What kind of "support"? Making effective use of superword-level
parallelism in compilers is far from being a solved problem...

- Nathan

Bill Studenmund

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to

Rudimentary assembler support I believe, and knowledge that those
registers exist. I'm not sure if there's much else.

Take care,

Bill

Henry B. Hotz

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
At 3:23 PM -0400 7/21/00, Nathan J. Williams wrote:
><wrst...@zembu.com> (Bill Studenmund) writes:
>
> > Apple has patches to gcc to support Altivec. I'm not sure if they've made
> > it out yet or not.
> >
>
>What kind of "support"? Making effective use of superword-level
>parallelism in compilers is far from being a solved problem...

I assume what you mean having the compiler automatically figure out
when it can make use of altivec for ordinary programs. I'd be happy
with it just having assembly support.

J.T. Conklin

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
>>>>> "Bill" == Bill Studenmund <wrst...@zembu.com> writes:
Bill> Apple has patches to gcc to support Altivec. I'm not sure if they've made
Bill> it out yet or not.

As I understand things, there are copyright assignment issues that
prevent the apple/motorola patches being integrated.

After I bought my G4, I discovered gas didn't have altivec support
(having assembler support is good enough for me). At the time, I
didn't know that the apple patches existed, so I added support for the
new instructions on my own. Since there was no copyright assignment
issues for my patches, they have been integrated into the development
binutils sources.

Unfortunately, I haven't had enough time to actually play with the
vector unit since then.

--jtc

--
J.T. Conklin
RedBack Networks

John Klos

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
Time for a new thread!

First, I was told by Troy, who is a developer for Terra Soft Solutions
(Yellow Dog and Black Lab Linux for PowerPC), and who also helped to get
Altivec support into gcc, that one place to go is altivec.org; they have
lots of software and information.

He also offered to help me (us) to get the latest diffs and mentioned that
he knows a lot of the pitfalls that the current gcc with Altivec has, and
is working on getting a clean set of diffs that generate a gcc that can
regenerate itself.

As soon as I have a little more time, I'm going to start working on this,
but I need to ask: has anyone built gcc 2.95.2 on NetBSD ppc? Who out
there has a G4 in case we do build a compiler and want to test it?

Ah... the luxury of time... And I still have a lot of documentation work
to finish...

John Klos


Nathan J. Williams

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
"Henry B. Hotz" <ho...@jpl.nasa.gov> writes:

> >What kind of "support"? Making effective use of superword-level
> >parallelism in compilers is far from being a solved problem...
>
> I assume what you mean having the compiler automatically figure out
> when it can make use of altivec for ordinary programs.

Yes, that's what I had in mind. See for example
http://www.cag.lcs.mit.edu/slp/ for some papers on the
subject. There's some good work, but it's mostly only applicable to
Fortran code.

> I'd be happy with it just having assembly support.

Sure, that's useful too.

- Nathan

Bill Studenmund

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On 21 Jul 2000, J.T. Conklin wrote:

> After I bought my G4, I discovered gas didn't have altivec support
> (having assembler support is good enough for me). At the time, I
> didn't know that the apple patches existed, so I added support for the
> new instructions on my own. Since there was no copyright assignment
> issues for my patches, they have been integrated into the development
> binutils sources.

Cool.

What's a good reference on the new altivec opcodes?

Take care,

Bill


Bill Studenmund

unread,
Jul 21, 2000, 3:00:00 AM7/21/00
to
On 21 Jul 2000, Nathan J. Williams wrote:

> Yes, that's what I had in mind. See for example
> http://www.cag.lcs.mit.edu/slp/ for some papers on the
> subject. There's some good work, but it's mostly only applicable to
> Fortran code.

well, we do have a fortran compiler... :-)

Take care,

Bill


Benjamin Herrenschmidt

unread,
Jul 24, 2000, 3:00:00 AM7/24/00
to
On Thu, Jul 20, 2000, Matthew Reilly <mjre...@inch.com> wrote:

>>
>>> As far as Airport goes, isn't it just a matter of bringing the cardbus
>>> driver from i386 into macppc? Anyone know if the WaveLan driver will
run an
>>> Airport card?
>>
>> Dunno. Should be pretty close, in theory....
>

>It's my understanding that the Airport interface is actually some sort of
>ATA bus, not PC card despite the similar form factor. I could be wrong
>through.

I got the Apple airport card working in linux/ppc recently. You can look
at my page (ppclinux.apple.com/~benh) for the latest hacked driver. I
basically had to hack the wvlan_cs driver of the Lucent card. I removed
the PCMCIA stuff, grabbed the card's address from the OF device tree, and
added a hack to various KeyLargo IOs in order to get the card up (the
hack was done after tracing MacOS code for powering up the card).

It looks like the card is actually a PCMCIA one (slightly hacked). Apple
makes a pseudo-PCMCIA bus by using a 16 bits IO port that goes out of the
KeyLargo ASIC, along with some of KL GPIOs in order to drive some PCMCIA
probe/control signals.

I also had to fix some endian bugs in the original linux driver, I don't
know if the BSD driver has the same bugs, after quickly looking at it, it
appears to be cleaner.

Ben.


0 new messages