Apple using Intel processors

G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Gack, snore, snooze, mumbling in his sleep. What's that he's saying?
Itanium? Nah, never. You can safely bet that Intel has had Power
emulation for Itanium scoped for a long time.

RM
 

Ed

Distinguished
Apr 1, 2004
1,253
0
19,280
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 23 May 2005 13:04:54 -0400, Yousuf Khan <bbbl67@ezrs.com> wrote:

>Here's a rumour that's as old as the hills rivalling the "Dell will use
>AMD" rumours in longevity. The latest: "Apple will use Intel" rumour.
>This time the Wall Street Journal is propagating the rumour.
>
>Apple to explore Intel chips for Macs
>http://biz.yahoo.com/rb/050523/tech_apple_intel.html?.v=3
>
> Yousuf Khan


Recently, Apple was unable to fulfil a promise to have 3GHz G5 chips
within a year of the unveiling of the G5-based Power Mac.

Apple denies eyeing Intel chips
http://www.pcpro.co.uk/news/73057/apple-denies-eyeing-intel-chips.html


Wasn't Apple saying "MHz doesn't matter" just a couple years ago? The
MAC is/was all about features and ease of use, not MHz?
Ed
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 23 May 2005 14:42:06 -0500, Ed <spam@hotmail.com> wrote:

>On Mon, 23 May 2005 13:04:54 -0400, Yousuf Khan <bbbl67@ezrs.com> wrote:
>
>>Here's a rumour that's as old as the hills rivalling the "Dell will use
>>AMD" rumours in longevity. The latest: "Apple will use Intel" rumour.
>>This time the Wall Street Journal is propagating the rumour.
>>
>>Apple to explore Intel chips for Macs
>>http://biz.yahoo.com/rb/050523/tech_apple_intel.html?.v=3
>>
>> Yousuf Khan
>
>
>Recently, Apple was unable to fulfil a promise to have 3GHz G5 chips
>within a year of the unveiling of the G5-based Power Mac.
>
>Apple denies eyeing Intel chips
>http://www.pcpro.co.uk/news/73057/apple-denies-eyeing-intel-chips.html
>
>
>Wasn't Apple saying "MHz doesn't matter" just a couple years ago? The
>MAC is/was all about features and ease of use, not MHz?

I don't think the IBM/Apple relationship has been a happy one.

RM
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Del Cecchi wrote:
> "Yousuf Khan" <bbbl67@ezrs.com> wrote in message
> > Apple likes to lock its users into proprietary computer systems, so
it has
> > no problems taking on proprietary processors from single sources.
So I
> > don't see why it would even bother to try to obtain price
concessions from
> > IBM, it really wouldn't matter to its fanatical userbase -- they'll
pay
> > anything to get it anyways. Dell on the other hand has to extract
every
> > last drop out of Intel.
>
> Care to explain how an Itanium isn't a "proprietary processor from a
single
> source"? Hell, a Pentium is too for all practical purposes. What a
dough
> head.
>

Well, I was saying that the Itanium was also exactly the same as a
PowerPC from IBM.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 23 May 2005 13:04:54 -0400, Yousuf Khan wrote:

> Here's a rumour that's as old as the hills rivalling the "Dell will use
> AMD" rumours in longevity. The latest: "Apple will use Intel" rumour.
> This time the Wall Street Journal is propagating the rumour.
>
> Apple to explore Intel chips for Macs
> http://biz.yahoo.com/rb/050523/tech_apple_intel.html?.v=3
>
> Yousuf Khan

well if most of osx is bsd based i doubt it would be that hard to port and
with the success of the ipod they have a substantial consumer base willing
to try a new os which hasnt been designed by a pack of middle managers
trying to extend their monopoly as far as possible but rather an os that
actualy works.
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

epaton wrote:
> well if most of osx is bsd based i doubt it would be that hard to port and
> with the success of the ipod they have a substantial consumer base willing
> to try a new os which hasnt been designed by a pack of middle managers
> trying to extend their monopoly as far as possible but rather an os that
> actualy works.

Apple suffers from the exact same curse and blessing that Microsoft
suffers from -- it has locked its users into a proprietary operating
system, and the applications are therefore also proprietary. It's a
blessing because it allows it to charge huge amounts to its userbase.
It's a curse because you're stuck on one platform forever.

In theory, OS X is portable, in reality it is not. Similarly, Windows NT
and its offspring are also theoretically portable. Windows was ported to
everything from Alpha to Itanium, but in the end the only architecture
that it's going to evolve with is x86-64. OS X can similarly be ported
to many architectures, but in the end it's only the PowerPC that's going
to matter to it. The reason being that nobody bothers to port to the
other architectures. Just having a ballfield there doesn't mean that
players will come.

Solaris is finding a similar situation. It's finding that it's installed
base of Solaris for Sparc aren't all that interested in Solaris on x86.

Yousuf Khan
 

mygarbage2000

Distinguished
Jun 5, 2002
126
0
18,680
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 23 May 2005 13:04:54 -0400, Yousuf Khan <bbbl67@ezrs.com>
wrote:

>Here's a rumour that's as old as the hills rivalling the "Dell will use
>AMD" rumours in longevity. The latest: "Apple will use Intel" rumour.
>This time the Wall Street Journal is propagating the rumour.
>
>Apple to explore Intel chips for Macs
>http://biz.yahoo.com/rb/050523/tech_apple_intel.html?.v=3
>
> Yousuf Khan

_IF_ (and it's a huge IF) Steve Jobs decides to recompile Mac OS for
Pentium, it should be able to run on any generic PC (including
AMD-based). They may try to artificially prevent it from running on
anything but Apple-branded systems, but almost surely there soon will
be a hack or two to circumvent it. Unless Apple is eyeing Itanic,
which will keep Macs highly proprietary and obscenely overpriced.
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

nobody@nowhere.net wrote:
> _IF_ (and it's a huge IF) Steve Jobs decides to recompile Mac OS for
> Pentium, it should be able to run on any generic PC (including
> AMD-based). They may try to artificially prevent it from running on
> anything but Apple-branded systems, but almost surely there soon will
> be a hack or two to circumvent it. Unless Apple is eyeing Itanic,
> which will keep Macs highly proprietary and obscenely overpriced.

Plus, it's highly doubtful that Itanium would help Apple obtain
low-power mobile parts for its laptops. The low-powered versions of
PowerPC seems to be what's at the heart of the bad blood between Apple
and IBM.

Apple likes to lock its users into proprietary computer systems, so it
has no problems taking on proprietary processors from single sources. So
I don't see why it would even bother to try to obtain price concessions
from IBM, it really wouldn't matter to its fanatical userbase -- they'll
pay anything to get it anyways. Dell on the other hand has to extract
every last drop out of Intel.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

"Yousuf Khan" <bbbl67@ezrs.com> wrote in message
news:NEwke.3603$dZ5.396799@news20.bellglobal.com...
> nobody@nowhere.net wrote:
>> _IF_ (and it's a huge IF) Steve Jobs decides to recompile Mac OS for
>> Pentium, it should be able to run on any generic PC (including
>> AMD-based). They may try to artificially prevent it from running on
>> anything but Apple-branded systems, but almost surely there soon will
>> be a hack or two to circumvent it. Unless Apple is eyeing Itanic,
>> which will keep Macs highly proprietary and obscenely overpriced.
>
> Plus, it's highly doubtful that Itanium would help Apple obtain low-power
> mobile parts for its laptops. The low-powered versions of PowerPC seems to
> be what's at the heart of the bad blood between Apple and IBM.
>
> Apple likes to lock its users into proprietary computer systems, so it has
> no problems taking on proprietary processors from single sources. So I
> don't see why it would even bother to try to obtain price concessions from
> IBM, it really wouldn't matter to its fanatical userbase -- they'll pay
> anything to get it anyways. Dell on the other hand has to extract every
> last drop out of Intel.
>
> Yousuf Khan

Care to explain how an Itanium isn't a "proprietary processor from a single
source"? Hell, a Pentium is too for all practical purposes. What a dough
head.

del cecchi
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 23 May 2005 22:51:20 -0400, Yousuf Khan <bbbl67@ezrs.com>
wrote:

>nobody@nowhere.net wrote:
>> _IF_ (and it's a huge IF) Steve Jobs decides to recompile Mac OS for
>> Pentium, it should be able to run on any generic PC (including
>> AMD-based). They may try to artificially prevent it from running on
>> anything but Apple-branded systems, but almost surely there soon will
>> be a hack or two to circumvent it. Unless Apple is eyeing Itanic,
>> which will keep Macs highly proprietary and obscenely overpriced.
>
>Plus, it's highly doubtful that Itanium would help Apple obtain
>low-power mobile parts for its laptops. The low-powered versions of
>PowerPC seems to be what's at the heart of the bad blood between Apple
>and IBM.
>
The whole story sounds wildly implausible. The only reason Itanium
seems like a possibility is that I gather it would be much more easy
to make Itanium look like Power than it would be to make x86 look like
Power, and it's hard to imagine a Mac that wouldn't run existing
software. Who knows what Intel may have been doing with power
management in Itanium; it has to be something they're worried about
even for server applications.

Given what Intel has sunk into Itanium, I can't imagine that the price
at which they would sell the chip to Apple would be an issue. It
would probably be worth it to them as a marketing stunt.

It all still sounds wildly implausible, with or without Itanium.

RM
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Robert Myers wrote:
> The whole story sounds wildly implausible. The only reason Itanium
> seems like a possibility is that I gather it would be much more easy
> to make Itanium look like Power than it would be to make x86 look
like
> Power, and it's hard to imagine a Mac that wouldn't run existing
> software. Who knows what Intel may have been doing with power
> management in Itanium; it has to be something they're worried about
> even for server applications.

Actually none of the articles about this story have mentioned anything
about Itanium, I think most people assume x86. It's only this and other
forums which are speculating on Itanium.

I don't think it's got anything to do with how well Itanium can emulate
Power, because it's quite obvious that Itanium can't emulate anything
well. Witness its x86 emulation, the most emulated architecture in the
world, even Intel the inventor of x86 could not get Itanium to emulate
it well; also, x86 and Power are much closer to each other than Itanium
is, since CISC and RISC are much closer to each other than they are to
EPIC.

I think the speculation has arisen simply for marketing factors that
line up between the two products. Apple likes proprietary hardware to
maintain its proprietary systems franchise: Itanium is the most
proprietary processor in the world, by design. Apple will only need a
small number of processors equivalent to maybe 2% of the desktop market
at present -- chump change in the x86 world, but could be quite
significant for Itanium.

> Given what Intel has sunk into Itanium, I can't imagine that the
price
> at which they would sell the chip to Apple would be an issue. It
> would probably be worth it to them as a marketing stunt.

Probably. They could also cheapen the cost of Itanium enough for Apple
if they bring out a small-cache version for desktops.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On 5/23/2005 13:04, Yousuf Khan wrote:
> Here's a rumour that's as old as the hills rivalling the "Dell will use

Yea, we've heard it before but WSJ missed the boat by nearly a month this
time:

<http://www.windowsitpro.com/Article/ArticleID/46175/46175.html?Ad=1>
(not too far down; search page for "bizarre") Unfortunately, no
attribution of which sandwich's owner said this.

~Jason

--

Posted Via Usenet.com Premium Usenet Newsgroup Services
----------------------------------------------------------
** SPEED ** RETENTION ** COMPLETION ** ANONYMITY **
----------------------------------------------------------
http://www.usenet.com
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

> Solaris is finding a similar situation. It's finding that it's installed
> base of Solaris for Sparc aren't all that interested in Solaris on x86.
>
> Yousuf Khan

having installed a beta of solaris10 on x86 i can say theres probaly other
reasons as well, it was missing some crucial drivers like for a creative
labs sound card and even a 3com network card.

also the solaris install is horrible, in stead of asking you all the
questions up front it makes you wait an hour in front of the pc.

that said now solaris is opening up i recon somebody could make a decent
distro out it, portaris (gentoo portage on solaris) is meant to be pretty
good and im sure theres plenty of other gentoo tweekers who like the idea
of multiple kernels and it would avoid a lot of the porting probs since
users would be doing it themselves.
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Tue, 24 May 2005 12:25:30 -0400, Jason Gurtz <no@noemail.com> wrote:

>On 5/23/2005 13:04, Yousuf Khan wrote:
>> Here's a rumour that's as old as the hills rivalling the "Dell will use
>
>Yea, we've heard it before but WSJ missed the boat by nearly a month this
>time:
>
><http://www.windowsitpro.com/Article/ArticleID/46175/46175.html?Ad=1>
>(not too far down; search page for "bizarre") Unfortunately, no
>attribution of which sandwich's owner said this.

You mean M$ really does err, sandwiches for "lunch"... or do you have to
pay for your own?

The next para, after "bizarre", is an interesting one too, especially for
these NGs: 25 x64 servers to replace 250 32-bit ones - something for the
x64 Luddites to gnaw on!:) Jeez I wonder if the new MSN servers are all
AMD x64s??

--
Rgds, George Macdonald
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

According to Robert Myers <rmyers@rustuck.com>:
>
> I don't think the IBM/Apple relationship has been a happy one.

Perhaps, with high volume customers in the game console market lining up,
IBM needs the MHz-demanding (and thus research-resource-demanding, high
cost) Apple less and less. Apple is just looking at other options in case
they get dumped?

Just speculation.
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Robert Myers <rmyers1400@comcast.net> writes:

> The whole story sounds wildly implausible. The only reason Itanium
> seems like a possibility is that I gather it would be much more easy
> to make Itanium look like Power than it would be to make x86 look like
> Power, and it's hard to imagine a Mac that wouldn't run existing
> software. Who knows what Intel may have been doing with power
> management in Itanium; it has to be something they're worried about
> even for server applications.

That's all just a question of how much high-level code is in OSX
compared to assembly code. Beyond that, making either Itanium or
x86 look like Power seems.... let's just hope there's lots of
high-level code in there.

> Given what Intel has sunk into Itanium, I can't imagine that the price
> at which they would sell the chip to Apple would be an issue. It
> would probably be worth it to them as a marketing stunt.
>
> It all still sounds wildly implausible, with or without Itanium.

It seemsmuchmore likely with x86-64 than either x86 or Itanium.
--
Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605
Department of Computer Science FAX -- (505) 646-1002
New Mexico State University http://www.cs.nmsu.edu/~pfeiffer
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Tue, 24 May 2005 05:47:29 -0400, Robert Myers
<rmyers1400@comcast.net> wrote:

>On Mon, 23 May 2005 22:51:20 -0400, Yousuf Khan <bbbl67@ezrs.com>
>wrote:
>
>>nobody@nowhere.net wrote:
>>> _IF_ (and it's a huge IF) Steve Jobs decides to recompile Mac OS for
>>> Pentium, it should be able to run on any generic PC (including
>>> AMD-based). They may try to artificially prevent it from running on
>>> anything but Apple-branded systems, but almost surely there soon will
>>> be a hack or two to circumvent it. Unless Apple is eyeing Itanic,
>>> which will keep Macs highly proprietary and obscenely overpriced.
>>
>>Plus, it's highly doubtful that Itanium would help Apple obtain
>>low-power mobile parts for its laptops. The low-powered versions of
>>PowerPC seems to be what's at the heart of the bad blood between Apple
>>and IBM.
>>
>The whole story sounds wildly implausible. The only reason Itanium
>seems like a possibility is that I gather it would be much more easy
>to make Itanium look like Power than it would be to make x86 look like
>Power, and it's hard to imagine a Mac that wouldn't run existing
>software.

Uhh, it's a HELL of a lot easier to make Power look like Power than to
make Itanium look like Power though. And that's the real key. It
would be completely moronic for Apple to consider Itanium! They would
lose compatible for the sake of getting a slower, more expensive and
more power hungry processor; a lose-lose-lose situation!

> Who knows what Intel may have been doing with power
>management in Itanium; it has to be something they're worried about
>even for server applications.

Ohh, I'm quite sure that Intel HAS been working on this, they MUST be.
I've mentioned it many times before, performance/watt is quickly
becoming a much more important measure of performance for server
processors than raw performance alone. With dual-core chips we're now
looking at being able to cram 8 processor cores in a 1U server, IFF
the power consumption is in the right ballpark.

>Given what Intel has sunk into Itanium, I can't imagine that the price
>at which they would sell the chip to Apple would be an issue. It
>would probably be worth it to them as a marketing stunt.

The question is how bad of a performance hit the chip would take to
meet a break even (just on per-unit costs) price point? In most
situations a PowerPC chip with 512KB of cache will beat out an Itanium
with 1.5MB of cache with ease, and that's WITHOUT having to emulate
PowerPC instructions for all the applications out there!

>It all still sounds wildly implausible, with or without Itanium.

Apple ditching PPC for x86 is wildly implausible. Apple ditching PPC
for IA-64 is completely absurd! At least with x86 they would have
high performance, low powered processors that are incompatible. With
Itanium they would have lower performance, higher powered processors
that are incompatible.

Ohh, and it doesn't even solve the basic rumor that started all this
nonsense; that Apple is not happy with only having a single supplier
for their processors.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Yousuf Khan wrote:

> Apple suffers from the exact same curse and blessing that Microsoft
> suffers from -- it has locked its users into a proprietary operating
> system, and the applications are therefore also proprietary. It's a
> blessing because it allows it to charge huge amounts to its userbase.
> It's a curse because you're stuck on one platform forever.

What ARE yu talking about. Mac started running on a 68000 with an OS I
can't remember, then went to MacOS, then to Power CPUs, then to OS-X. It
may be proprietary, but it sure isn't "one playform" no matter how you
define it.
>
> In theory, OS X is portable, in reality it is not. Similarly, Windows NT
> and its offspring are also theoretically portable. Windows was ported to
> everything from Alpha to Itanium, but in the end the only architecture
> that it's going to evolve with is x86-64. OS X can similarly be ported
> to many architectures, but in the end it's only the PowerPC that's going
> to matter to it. The reason being that nobody bothers to port to the
> other architectures. Just having a ballfield there doesn't mean that
> players will come.

AFAIK OS-X is open source, although the graphical manager isn't. It's
based on BSD, and I think it was ported to Sparc as a proof of concept
(by two students on a one week break if I recall). And you can run (at
least) freeBSD and Linux on Mac, although I've only seen it on a laptop.
>
> Solaris is finding a similar situation. It's finding that it's installed
> base of Solaris for Sparc aren't all that interested in Solaris on x86.

If you buy hardware to be compatible it's okay, although the install was
painful last time I saw it. But Sun hardware has come down in price, why
bother?

Thought: Sun went to SPARC because they couldn't get Motorola to push
the 68k fast enough. I wonder if you could port OS-X to Sparc, which is
made by multiple vendors? How's that for a rumor starter?

--
bill davidsen (davidsen@darkstar.prodigy.com)
SBC/Prodigy Yorktown Heights NY data center
Project Leader, USENET news
http://newsgroups.news.prodigy.com
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Bill Davidsen wrote:
> Yousuf Khan wrote:
>
>> Apple suffers from the exact same curse and blessing that Microsoft
>> suffers from -- it has locked its users into a proprietary operating
>> system, and the applications are therefore also proprietary. It's a
>> blessing because it allows it to charge huge amounts to its userbase.
>> It's a curse because you're stuck on one platform forever.
>
>
> What ARE yu talking about. Mac started running on a 68000 with an OS I
> can't remember, then went to MacOS, then to Power CPUs, then to OS-X. It
> may be proprietary, but it sure isn't "one playform" no matter how you
> define it.

Well, yes Apple has been a lot more successful at pushing its userbase
around than Microsoft has in the past. But back then, the installed
userbase of Macs was a lot smaller than it is now too: a few million vs.
about 25 million now. The last major change really was the switchover
from 68xxx to PowerPC -- that was a hardware switchover. The switchover
from MacOS to OS X is trivial by comparison -- it's just software. In
the hardware switchover, Apple had to come up with hardware emulators
and stuff, whereas in the OS software switch, the majority of
applications worked between the OSes, with a few odd men out.

Microsoft's userbase in the 100's of millions, possibly approaching the
billion mark. Microsoft's userbase was already in the 100's of millions
when Mac made the jump from 68xxx to PPC. So really the size of the
transition that you can force down your locked-in users' throats is
inversely proportional to the size of your userbase. Apple couldn't hope
to make the 68xxx to PPC transition today, with its userbase as large as
it is today.

>> In theory, OS X is portable, in reality it is not. Similarly, Windows
>> NT and its offspring are also theoretically portable. Windows was
>> ported to everything from Alpha to Itanium, but in the end the only
>> architecture that it's going to evolve with is x86-64. OS X can
>> similarly be ported to many architectures, but in the end it's only
>> the PowerPC that's going to matter to it. The reason being that nobody
>> bothers to port to the other architectures. Just having a ballfield
>> there doesn't mean that players will come.
>
>
> AFAIK OS-X is open source, although the graphical manager isn't. It's
> based on BSD, and I think it was ported to Sparc as a proof of concept
> (by two students on a one week break if I recall). And you can run (at
> least) freeBSD and Linux on Mac, although I've only seen it on a laptop.

The reality is that Mac OS X is not portable, and it's got nothing to do
with what programming language it was written in, or if its close
relatives are running on other platforms. It's not portable, because its
userbase won't let it be ported. It's too entrenched. Everytime an OS
port is made, doesn't necessarily guarantee that the userbase will
follow into the new platform. HP-UX exists on PA-RISC and Itanium, the
PA-RISC crowd isn't easily following HP-UX into Itanium for HP. For Sun,
Solaris exists on Sparc and x86/x64, the Sparc crowd doesn't necessarily
adopt Solaris on x64 just because it's there.

>> Solaris is finding a similar situation. It's finding that it's
>> installed base of Solaris for Sparc aren't all that interested in
>> Solaris on x86.
>
>
> If you buy hardware to be compatible it's okay, although the install was
> painful last time I saw it. But Sun hardware has come down in price, why
> bother?

You're missing my point, it's not because of missing device drivers that
it hasn't picked up. It's because just having the same OS available on a
different platform is not enough. The OS by itself will not run all of
your application binaries on the other platform. You still have to wait
for applications to be ported too.

>
> Thought: Sun went to SPARC because they couldn't get Motorola to push
> the 68k fast enough. I wonder if you could port OS-X to Sparc, which is
> made by multiple vendors? How's that for a rumor starter?
>

Sparc is well on its way to becoming the next embedded processor core, a
a la MIPS, or ARM. The good old days of general purpose computing are
behind it.

Yousuf Khan
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Thu, 26 May 2005 02:59:05 GMT, Bill Davidsen
<davidsen@darkstar.prodigy.com> wrote:

>Yousuf Khan wrote:
>
>> In theory, OS X is portable, in reality it is not. Similarly, Windows NT
>> and its offspring are also theoretically portable. Windows was ported to
>> everything from Alpha to Itanium, but in the end the only architecture
>> that it's going to evolve with is x86-64. OS X can similarly be ported
>> to many architectures, but in the end it's only the PowerPC that's going
>> to matter to it. The reason being that nobody bothers to port to the
>> other architectures. Just having a ballfield there doesn't mean that
>> players will come.
>
>AFAIK OS-X is open source, although the graphical manager isn't. It's

Perhaps just semantics here, but OS-X is NOT open source at all.
Darwin is an open source OS licensed under the BSD license. OS-X is a
closed source OS that is built on top of Darwin. The GUI is the most
obvious difference between Darwin and OS-X, but it's not the only one.

FWIW though, Darwin HAS been ported to x86 and somewhat actively
maintained.

>based on BSD, and I think it was ported to Sparc as a proof of concept
>(by two students on a one week break if I recall). And you can run (at
>least) freeBSD and Linux on Mac, although I've only seen it on a laptop.

Darwin probably has been ported to a few different architectures,
though PowerPC and x86 are the only two that are actively maintained.

To the best of my knowledge, OS-X has never been ported to any
architecture other than PowerPC.

>> Solaris is finding a similar situation. It's finding that it's installed
>> base of Solaris for Sparc aren't all that interested in Solaris on x86.
>
>If you buy hardware to be compatible it's okay, although the install was
>painful last time I saw it. But Sun hardware has come down in price, why
>bother?
>
>Thought: Sun went to SPARC because they couldn't get Motorola to push
>the 68k fast enough. I wonder if you could port OS-X to Sparc, which is
>made by multiple vendors? How's that for a rumor starter?

Makes a bit more sense than switching over to IA-64, though still
seems like a pretty dumb idea IMO. Sun's processors aren't well
suited to desktop or workstation applications. Fujitsu's chips do
better, though they haven't been keeping up with x86 or PowerPC.
Pretty much everyone else has got out of the SPARC chip market or are
focusing at the low-end/embedded market, and even that seems to be on
the out as ARM chips replace most others.

-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Yousuf Khan wrote:
^
> Apple suffers from the exact same curse and blessing that Microsoft
> suffers from -- it has locked its users into a proprietary operating
> system, and the applications are therefore also proprietary. It's a
> blessing because it allows it to charge huge amounts to its userbase.
> It's a curse because you're stuck on one platform forever.

And this is in what respect different to every other platform? No matter
if Mac, Win or anything else, if You want to run applications for a
certain platform You're stuck with it (except if You want to use
emulators of course)...

> In theory, OS X is portable, in reality it is not.

No? MacOS X is based on FreeBSD. Isn't FreeBSD portable? And You
obviously miss that the MacOS X Base "Darvin" (which is MacOS X except
the Aqua desktop) runs on x86 for over five years now. Everytime a new
MacOS X version comes out there also is a new Darvin - for PPC and
x86...

> Similarly, Windows
> NT and its offspring are also theoretically portable. Windows was
> ported to everything from Alpha to Itanium, but in the end the only
> architecture that it's going to evolve with is x86-64.

But this is purely a market decision, not a technical reason....

> OS X can
> similarly be ported to many architectures, but in the end it's only
> the PowerPC that's going to matter to it. The reason being that
> nobody bothers to port to the other architectures. Just having a
> ballfield there doesn't mean that players will come.

You want to have a look here:
http://developer.apple.com/darwin/

This is MacOS X except the code Apple can't and/or doesn't want to make
Open Source...

> Solaris is finding a similar situation. It's finding that it's
> installed base of Solaris for Sparc aren't all that interested in
> Solaris on x86.

Certainly not. SPARC sales are dropping like hell. Sun invests more and
more of its ressources in Opteron-based systems, and these machines run
(You guess it) Solaris x86... ;-)

Solaris x86 once really was just a niche for Sun but today it is as
important as Solaris SPARC - probably even more. And traditional Sun
customers are happy that they can use x86-64 hardware which is a)
cheaper b) more powerful than even the fastest UltraSPARC and c) doesn't
need to be from Sun or the few OEMs like Axil...

Benjamin
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

Yousuf Khan wrote:
> Bill Davidsen wrote:
>
>>Yousuf Khan wrote:
>>
>>
>>>Apple suffers from the exact same curse and blessing that Microsoft
>>>suffers from -- it has locked its users into a proprietary operating
>>>system, and the applications are therefore also proprietary. It's a
>>>blessing because it allows it to charge huge amounts to its userbase.
>>>It's a curse because you're stuck on one platform forever.
>>
>>
>>What ARE yu talking about. Mac started running on a 68000 with an OS I
>>can't remember, then went to MacOS, then to Power CPUs, then to OS-X. It
>>may be proprietary, but it sure isn't "one playform" no matter how you
>>define it.
>
>
> Well, yes Apple has been a lot more successful at pushing its userbase
> around than Microsoft has in the past. But back then, the installed
> userbase of Macs was a lot smaller than it is now too: a few million vs.
> about 25 million now. The last major change really was the switchover
> from 68xxx to PowerPC -- that was a hardware switchover. The switchover
> from MacOS to OS X is trivial by comparison -- it's just software. In
> the hardware switchover, Apple had to come up with hardware emulators
> and stuff, whereas in the OS software switch, the majority of
> applications worked between the OSes, with a few odd men out.
>
> Microsoft's userbase in the 100's of millions, possibly approaching the
> billion mark. Microsoft's userbase was already in the 100's of millions
> when Mac made the jump from 68xxx to PPC. So really the size of the
> transition that you can force down your locked-in users' throats is
> inversely proportional to the size of your userbase. Apple couldn't hope
> to make the 68xxx to PPC transition today, with its userbase as large as
> it is today.

Humm. Apple ported from 68xxx to PowerPC, and ported from one operating
system to a completely different one. The processor move was long, but
not that long. Apple provided an emulator, and most vendors simply
recompiled around that limitation. Carbon did the rest.

I would assert the opposite. If apple changed CPUs again, their user
base and vendors would switch faster this time, because they are used to
it. If Apple switched once a year, then it would still work out. Vendors
would simply get used to changing CPUs with version numbers.

The OS API is far more important than the CPU type. That's as it should
be, high level languages were supposed to make that possible.

Now onto Microsoft. They have stayed put ? No. The move from 16 bit
windows to 32 bit windows was huge, both in terms of virtually recoding
everything, and ALSO in terms of changing the API, since they not only
widened everything, but 32 bit flat mode changed the way memory and
programs were handled entirely.

Then Microsoft went from Windows to Windows NT, a complete, from scratch
recoding of the operating system.

Now Microsoft is going from 32 bits to 64 bits, on the very same
processor. If they are "entrenched", there is a hell of a strong
current running through that trench !

>
>
>>>In theory, OS X is portable, in reality it is not. Similarly, Windows

It does not have to be. Only the API has to be stable, and the API is
evolving into a collection of APIs. OpenGL is in common to both Windows,
Apple and even Linux. Windows used to be WOW for 16 bit to 32 bit,
now its WOW32 for Windows 32 bit to 64 bit.

>>
>>
>>AFAIK OS-X is open source, although the graphical manager isn't. It's
>>based on BSD, and I think it was ported to Sparc as a proof of concept
>>(by two students on a one week break if I recall). And you can run (at
>>least) freeBSD and Linux on Mac, although I've only seen it on a laptop.
>

The Unix section is open source. Everything else is not. Thats like saying
that the statue of liberty's base is free for the taking. Try moving it.

>
> The reality is that Mac OS X is not portable, and it's got nothing to do
> with what programming language it was written in, or if its close
> relatives are running on other platforms. It's not portable, because its
> userbase won't let it be ported. It's too entrenched. Everytime an OS
> port is made, doesn't necessarily guarantee that the userbase will
> follow into the new platform. HP-UX exists on PA-RISC and Itanium, the
> PA-RISC crowd isn't easily following HP-UX into Itanium for HP. For Sun,
> Solaris exists on Sparc and x86/x64, the Sparc crowd doesn't necessarily
> adopt Solaris on x64 just because it's there.

All true, and who cares. Apple and Microsoft don't port operating systems,
they recode them entirely, and emulate the old API. Why ? Because by the
time it became necessary to move, the code was obsolete anyways.
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

George Macdonald wrote:
> On Tue, 24 May 2005 12:25:30 -0400, Jason Gurtz <no@noemail.com> wrote:
>
>
>>On 5/23/2005 13:04, Yousuf Khan wrote:
>>
>>>Here's a rumour that's as old as the hills rivalling the "Dell will use
>>
>>Yea, we've heard it before but WSJ missed the boat by nearly a month this
>>time:
>>
>><http://www.windowsitpro.com/Article/ArticleID/46175/46175.html?Ad=1>
>>(not too far down; search page for "bizarre") Unfortunately, no
>>attribution of which sandwich's owner said this.
>
>
> You mean M$ really does err, sandwiches for "lunch"... or do you have to
> pay for your own?
>
> The next para, after "bizarre", is an interesting one too, especially for
> these NGs: 25 x64 servers to replace 250 32-bit ones - something for the
> x64 Luddites to gnaw on!:) Jeez I wonder if the new MSN servers are all
> AMD x64s??
>
Brings to mind several things, among them why 64 bit makes much of a
difference in an application typically dominated by i/o rather than
anything CPU, if the new servers come with new software which would run
faster on any CPU, and if the 250 servers (what were they DOING?) were
loaded to capacity.

Servers and workstations have had 64 bit capability for a decade in the
RISC vendors, and I don't think there's much sales volume in this
market. I'd love to know more details.

--
bill davidsen
SBC/Prodigy Yorktown Heights NY data center
http://newsgroups.news.prodigy.com
 
G

Guest

Guest
Archived from groups: comp.sys.ibm.pc.hardware.chips,comp.sys.intel (More info?)

On Mon, 06 Jun 2005 19:36:49 GMT, Bill Davidsen
<davidsen@deathstar.prodigy.com> wrote:

>George Macdonald wrote:
>> On Tue, 24 May 2005 12:25:30 -0400, Jason Gurtz <no@noemail.com> wrote:
>>
>>
>>>On 5/23/2005 13:04, Yousuf Khan wrote:
>>>
>>>>Here's a rumour that's as old as the hills rivalling the "Dell will use
>>>
>>>Yea, we've heard it before but WSJ missed the boat by nearly a month this
>>>time:
>>>
>>><http://www.windowsitpro.com/Article/ArticleID/46175/46175.html?Ad=1>
>>>(not too far down; search page for "bizarre") Unfortunately, no
>>>attribution of which sandwich's owner said this.
>>
>>
>> You mean M$ really does err, sandwiches for "lunch"... or do you have to
>> pay for your own?
>>
>> The next para, after "bizarre", is an interesting one too, especially for
>> these NGs: 25 x64 servers to replace 250 32-bit ones - something for the
>> x64 Luddites to gnaw on!:) Jeez I wonder if the new MSN servers are all
>> AMD x64s??
>>
>Brings to mind several things, among them why 64 bit makes much of a
>difference in an application typically dominated by i/o rather than
>anything CPU, if the new servers come with new software which would run
>faster on any CPU, and if the 250 servers (what were they DOING?) were
>loaded to capacity.

Well we have to count that some of those 32-bit servers were old - say
800MHz (dual) Xeons. Even so a 10x reduction is impressive. As to 64-bit
making a difference, AIUI, there is less I/O since the 32-bit processes can
be >2GB of application space and the total memory can be >4GB so processes
can sit around in memory instead of being paged in & out. IOW the only I/O
is real transaction I/O.

>Servers and workstations have had 64 bit capability for a decade in the
>RISC vendors, and I don't think there's much sales volume in this
>market. I'd love to know more details.

If you go here http://72.3.250.42/amd_05e1/index.html (AMD bumph but be
patient) and look at the MS_Treasury article under Enterprise Success
Stories there's a user perspective which has some interesting info.

--
Rgds, George Macdonald