Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Fukui said:
I dont see why not. All most small-time devs need to do is click a check box...

Exactly. It's not like they have to go out of their way to compile for either architecture.
 
~loserman~ said:
YEP I sure have..... and Steve's too
Of course my post was tongue in cheek....
Since OS X has the worst memory subsystem of any OS on the planet.
.
Not sure I catch your drift with this statement.(are you referring to SMP kernels?)
My Dual and Quad Opteron boxes scale memory bandwidth almost perfectly.(with NUMA)

The AMD cpus unlike the INTELS at least don't have to exit the die and go to the FSB to see ea others memory. They have an internal crossbar between the cores besides hypertransport.
They are superior in every way to Intel's offering.(Intel guy thinking..."hey we can make a dual core chip! We will just take 2 Cores and throw them on the same die". Another Intel guy says " but that won't really be a dual core processor since they can't communicate with ea other". 1st Intel guy says " sshhhhhhhhh")
A dual dual core XEON based system will have to share the memory with all four cores. Talk about memory starvation.
At least with AMD's offering they will get close to twice the bandwidth.

By the way all my programs are single threaded CFD codes and we run all our Opteron Clusters with LINUX using a NUMA kernel. THATS why I smoke the AMD weed.....Instead of the Intel Crack Pipe

...and to simplify things the top Intels give about the same performance as top AMDs in everyday software... right now no one has any advantage in real world performance, you can run 3d benchmarks all you want but the reality is that Intel with the Pentium M architecture will take the marginal lead away from AMD, you talk about on chip mem controllers, and at what speed does the fastest RAM run these days? maybe it makes sense for servers but for multimedia comps and apps it doesn't...
Its not like Intel is so far behind, Athlons beat Intels in gaming benchmarks by couple of frames, big deal...
Check out the prices of the new dual core processors from Intel and AMD... Intels are way cheaper... Also take into account that as some people might not know but in the year of 2003 when AMD didnt post such high profits (actually they operated at a loss) guess who bough a couple of millions of shares in that company?
Intel...

Intel knows what they are doing, the partnership with Apple is more than just processors, Intel is finally out of grasp of Microsoft, Apple and Intel can work on things that Intel wanted to work for a long time but MS was the obstacle for them, even Intel knows current x86 architecture isn't going too far... With Apple they have a totally independent hardware and software (with its own OS) company (which is easily adaptable to new technologies/architectures and is not so crippled by PC standards)... I am willing to bet that sooner or later they will both experiment on new architectures (that won't be x86 nor PPC, something radically new)...
 
Brian0523 said:
This thread is proof that people with too little information become a danger to theirself and others around them!

Or they just get irritating. The real problem is I can't stop reading through the shlock.
 
Give Apple credit for some forethought, there will be a HARDWARE or similar solution so that OSX only boots on Apple machines.
 
chibianh said:
pfft.. cmon! If you watched the Keynote, you couldn't even TELL steve wasn't using a G5 until he actually told everyone. Just because it's a different processor doesn't mean it still won't be a Mac. give me a break.

That's exactly right. I'd be happy so long as I can get a dual-dual-core, a monster graphics card, and an inexpensive fiber channel card to the SAN in my basement.

NEWAY: I already run Linux on pretty much everything, and the whole point of 'Linux on pretty much everything' is that it highlights 'Linux' and not the specific features of the weird and varied hardware I run it on. If I want the Linux experience, I can have it, pretty much everywhere.

If I want the Apple OS X experience, I can now have it on an okay Intel motherboard, or a PPC. So long as my experience is super-sexy, I'm a happy puppy!

Here's what I hope from Apple on Intel:

- Lighter laptops with carbon fiber. My 12" is nearly perfect -- if only 2x the speed, 2x the rez, and 2x the battery life.
- A long-life palm-top akin to the Sharp Zarus (which I also own) but with BlueTooth, WiFi, and GPS. We'll call it iGo. or IAm.

While I'm going, here's the rest of my list:
- A concept of a 'media center' MacMini, with HDMI, and video recording. Should have a wireless capability to store information on an 'XSANmini' in the basement. Everyone should have terrabytes beside their furnace in the basement.

- MacGrid high density server blades. Apple would make great management apps for large grid clusters. Right now MicroWay does okay, but we can do much better. It would be awesome if we could stick these grids in our basements, beside the XSANmini's and our furnaces.

As for Apple Applications, here's what I want:

- Apple Life Planner - a health planner linked to a full medical database and holistica healing knowledgebase, with voip to doctors and health practitioners. It should also be able to use iPod technology as a personal health recorder, taking your heart rate, and other stress info (i.e. it should listen to your voice for stress patterns) and factor that input. It should also have a financial application that works intuitively like Garage Band. I can't deal with Quicken, etc... Their numbers never seem to add up. It should also have a library of recipes, and have logic to create grocery shopping lists (most food programs randomize recipes so lots of food you buy goes bad). It should let you transfer money into various types of savings, and provide an iTunes like front-end for various types of investment, with 99 cent trades. That's it for version 1.0. Essentially, it has to work on Maslow's hierarchy of needs. Food, shelter, etc... all the way up to self actualization. I could work on the design for this if someone is interested. doktorfuture@mac.com

It would truly be the iLife killer app.
 
Fukui said:
I dont see why not. All most small-time devs need to do is click a check box...

And many big-time devs can do that too! ;)

Example: Mathematica 5
"There are millions of lines of C code, Java code, ancient code that hasn't been changed since the Reagan administration. This isn't a little toy app, it's a beast". - Theo Gray of Wolfram Research at WWDC 2005

This is not that bad of a transition!
 
Mr Maui said:
I don't believe Steve Jobs wants to phase out the PPC. I believe IBM has left him with little option. They have become the Motorola of the 21st century. Apple placed their trust in IBMs promises to deliver large quantities of fast, powerful PPC processors, and committed to delivering faster and better ones in the future (most likely on a timetable ... i.e. "3 GHx in a year"), but IBM failed to deliver on their promises. I've read many posts over the past year with people complaining that Apple was not upgrading their systems fast enough, that Apple was not delivering a faster Powerbook solution, that Apple was neglecting its customers for the iPod, that Steve Jobs didn't care about the people that made Apple what it was ... but in truth, IBM was to blame. I rarely read posts saying that IBM let customers down, that IBM was falling short, or that IBM bore any responsibility for the failed deliveries of promises by Apple. I'm certain that Steve would have preferred to stay with PPC and not have to eat his own words (based on promises from IBM), but was left with little choice if he wanted to continue to grow his company into the future, especially in the areas of portability. Neither IBM, nor Freescale, has offered cool solutions for laptop expandability, the fastest growing area of computer purchases in the market, so Steve Jobs had to decide what was best for the future. It appears that the Pentium-M is the cool choice.

ok, but we will have the machines in 1 to 2 years from now and will have to mobilize a whole industry to do yet another transition.. by then, I'm sure IBM will have ramped-up its production of fast PPC'sx.. they (IBM) surely didn't invest huge amounts of $'s in a factory if that wasn't the ultimate goal .. I must be missing something... (built-in DRM ?)
 
A few things I hope Apple does.

1. Get rid of the BIOS, use something that will be even better than OF.

2. Use the common Intel CPU sockets, easy CPU upgrades ;)

3. Be a leader in technology. They were for a while with USB and Firewire until recently when the PCs passed them with PCI-EXPRESS etc. But now that they are teaming with Intel, PCI-EXPRESS, express-cards, WIMAX, DDR500, Blueray, etc should be a reality.
 
So, the story goes...

A tech reporter asked a high-level Intel exec at a recent press conference:

"Do you have your Yonahs up?"

The Intel exec declined to comment on unreleased products... but he blew a kiss to the reporter!
 
magi.sys said:
Get rid of the BIOS, use something that will be even better than OF.

If they are replacing OF, then it will undoubtedly be better. This is Apple we're talking about! ;)

But, lets clarify something about BIOS.

From Wikipedia:
In computing, the Basic Input-Output System or BIOS is primitive software code embedded within a computer hardware system whose main functions are: to provide a visual display of the system to a monitor on startup of the system, have basic key access to a keyboard and provide low-level communication services among hardware components.

Isn't Open Firmware then a type of BIOS? :confused:
 
spinko said:
ok, but we will have the machines in 1 to 2 years from now and will have to mobilize a whole industry to do yet another transition.. by then, I'm sure IBM will have ramped-up its production of fast PPC'sx.. they (IBM) surely didn't invest huge amounts of $'s in a factory if that wasn't the ultimate goal .. I must be missing something... (built-in DRM ?)
What production needs to be ramped up? IBM's Fishkill, NY fabrication plant produces multiple millions of chips. Apple accounts for between 2% and 5% of total production from the plant. depending on who is your source of information. Production capability is not the real issue for IBM. Production DESIRE and INTEREST is problem. IBM is interested in producing lots of chips for Microsoft, Sony and Nintendo for the gameboxes and not producing newer, better, or additional chips for Apple. IBM is content to produce what they can make lots of money on with little or no R&D at this point. Apple is too small of a fish in the IBM pond to make any effort for. IBM made a commitment and then broke it. Period. Not because they were incapable, but because they were disinterested. They may produce faster and better chips in the future, and Apple may or may not use some of them, but why now? Why in the future? Why not a year ago? Then Apple would still be with IBM and this discussion would not exist.
 
Since Powerbooks will probably use Pentium-M CPUs, maybe we'll be able to switch it with the rumored VIA socket 479 CPU. Now that would be great since it will have Padlock (HW security engine) and will be able to SSH/SSL/IPSEC at full network speeds. That would be very interesting.
 
magi.sys said:
A few things I hope Apple does.

1. Get rid of the BIOS, use something that will be even better than OF.

2. Use the common Intel CPU sockets, easy CPU upgrades ;)

3. Be a leader in technology. They were for a while with USB and Firewire until recently when the PCs passed them with PCI-EXPRESS etc. But now that they are teaming with Intel, PCI-EXPRESS, express-cards, WIMAX, DDR500, Blueray, etc should be a reality.

Please note, that by asking them to remove #1, you are in fact precluding the use of #2. Using a PC bios would be the BEST thing for compatibility and future upgrades. Whereas a custom Apple "OF" or whatever could be limiting in what it will allow to boot the machine / etc.

A bios is "Basic Input Output System" Why in the world is anyone getting romantic over a bios? Has anyone here seen the new BIOS on a modern PC motherboard? The number of options and configuration settings is mind boggling. And to me, that is a GOOD thing! I like options, and tweaking. I don't expect to see as much of that in an apple, but please don't ask apple to hamper and tie down the machines!
 
Quartz Extreme said:
But, lets clarify something about BIOS.

Isn't Open Firmware then a type of BIOS? :confused:

Yes. Yes. Yes. Yes. Yes. Yes.

YES!

Everyone read his post again.

And please, please, stop it with the "I don't want my Mac to have a BIOS" because it clearly shows people's ignorance on the subject. You might as well be saying "I won't buy any Mac with a keyboard, I don't want no Windows logo on any of my keys!"
 
iMeowbot said:
That's not the way any architecture or OS change has happened, in the little Apple universe or anywhere else. The smaller devs are the ones who will take "forever" to get onto the new architecture. Those are the ones who haven't got around to migrating to Xcode, and don't have budgets to spend on all the latest and greatest Macs. They'll still be offering only PowerPC binaries and you'll be screaming at them to get with the times.

Nah, they'll click the checkboxes and ship both just because it's frankly pretty cool.

It'll just need some good beta-testing if they don't have one of the platforms.

NeXT programmers (especially share/freeware) made their programs four-way fat, even though most of them would never even *see* NeXTSTEP running on an HP or Sun workstation, which ran up to over $20,000, and weren't about to pay to fly out to NeXT to test things there.
 
Correct me if I'm wrong, but isn't running x86 hardware and software going to make it EASIER to port an APP to OSX? I mean, eventually they won't have to bother porting it to PPC... and since OSX is such an efficient and stable OS, why WOULDN'T macs become great gaming computers? I mean, I can't tell you how many times I've been playing WOW and had a party member have to drop out, reboot, and re-join, whereas I've NEVER ONCE had to do so.

Unless your talking about some large cunk of assembly, really the answer is going to be no. For instance look at linux applications, most of them nicly recompile on alpha/sparc/powerpc/x86/arm etc.

with the speed of todays computers hardly anyone is programming in assembly making very little of a application arch spersific. Actully you would prolly get minimal to none in speed increase from using assembly above C since compilers like gcc are so well optermized.
 
dicklacara said:
Hey Loserman ... long time no, whatever.

I read somewhere that OS X Server couldn't compare to Linux because it does a sucky job creating threads....

Was that one of your issues?

Did you ever convert to Linux?

What do you think of ditch & switch?

Dick

Funny thing is happening on the OS X to Linux switch....
It turns out that if we ditch OS X's TCP/IP stack then OS X runs fine.
For some reason unknown to us OS X's TCP/IP stack introduces around 15 seconds latency to our App at every iteration( a cycle of our code)
While Linux's TCP/IP stack only adds .6 seconds to an iteration.

We tested on Myrinet so that we could avoid running TCP/IP and OS X is actually running a little faster than Linux now.

Weird huh?

I am on both sides of the fence on the PPC to Intel Switch.
I understand why Apple is doing it.....
But
It means we won't be buying any more Xserves.
Our whole reason for buying Xserves was the price performance value of the double precision floating point provided by the 970.
Since Apple will be moving to Intel they wont be a player for our HPC dollars.

I'll add this too.
We were very surprised to here this news. We had just had 2 Apple VP's at our office giving us an NDA briefing less than a month ago on their Server/Processor roadmap/Future product releases and they didn't say a word or hint about this.
In hindsight the briefing they gave us makes sense now because we came away from that briefing feeling like we hadn't even had one.
 
spinko said:
ok, but we will have the machines in 1 to 2 years from now and will have to mobilize a whole industry to do yet another transition.. by then, I'm sure IBM will have ramped-up its production of fast PPC'sx.. they (IBM) surely didn't invest huge amounts of $'s in a factory if that wasn't the ultimate goal .. I must be missing something... (built-in DRM ?)

hmmm, that's what i was thinking 2 years ago: "I'm sure in 1 year IBM will ramp up and have gobs of superfast G5's pouring out of fishkill in time to make steve's promise come true". It didn't happen then; now with even less demand for a PC-targeted G5 chip why should we think it's going to happen now?
 
pont said:
Unless your talking about some large cunk of assembly, really the answer is going to be no. For instance look at linux applications, most of them nicly recompile on alpha/sparc/powerpc/x86/arm etc.

with the speed of todays computers hardly anyone is programming in assembly making very little of a application arch spersific. Actully you would prolly get minimal to none in speed increase from using assembly above C since compilers like gcc are so well optermized.

I agree. For the most part it is the OS that is important as far as developing for different architectures. Generally if An OS runs on multiple platforms then its usually a simple recompile of your code because your App is already using predefined libraries and APIs of the OS.

I would not be so bold to describe gcc as being so well optimized. It does an OK job... but can't compare to native compilers such as XLC on Power or ICC on Intel. Although gcc 4.0 does show some promise.
 
daver969 said:
hmmm, that's what i was thinking 2 years ago: "I'm sure in 1 year IBM will ramp up and have gobs of superfast G5's pouring out of fishkill in time to make steve's promise come true". It didn't happen then; now with even less demand for a PC-targeted G5 chip why should we think it's going to happen now?

Here is a funny tale....
We recently talked to IBM and they told us they will be shipping up to 2.5Ghz dual core 970's by January in their JS20 blade servers.
 
1) I'm surprised no one has coined the term before, this transition has been planned since the very beginning. It always been a Mac[int]osh :) We're just so used to using the short form.

2) Due to the custom motherbard and processor, its been difficult to price the components in the machines of today, since the G5's retail price is not readily available, Apple can charge what they want. I believe that the margin that Apple enjoys today will diminish. Apple has tried making money on per unit sales an that strategy has failed. Lower margin on higher volume is more likely. Apple will have a very hard time justifying prices once one can calculate the costs of the machine based on off the shelf components [it is irrelevant that
the board Apple uses may be custom, if I can get a similar board from Asus that can power the same CPU, why am I paying apple so much more for it.

3) A significant portion of the revenues enjoyed in the PC industry is sales of OEM components. Apple has not been able to profit from the tweaker market. If I can get a faster CPU for my Mac[int]osh 12 months after buying it, Intel benefits. Apple looses nothing as there is just no way that I would have bought a new machine at that point. Apple will reap the benefits of consistent sales [a Mac user will still buy a machine every X years] as well as some incremental updates. As it stands today, what can we really do with our machines/
 
~loserman~ said:
Here is a funny tale....
We recently talked to IBM and they told us they will be shipping up to 2.5Ghz dual core 970's by January in their JS20 blade servers.

ah, well maybe if they're their own customer they'll be more motivated to live up to that deadline.
 
Fender said:
A wonderful piece of work by a dude named Hywel Thomas:

http://homepage.mac.com/hywel.thoma...s/flowchart.pdf

Many, Many thanks for this. I'm greatful for it. Now I can rest camly and buy a older 1.8 Dual G5, or iMac G5 Rev B; love the design of both and not fear of apps for a few years. After that there is always Yellow Dog Linux http://lists.terrasoftsolutions.com/pipermail/yellowdog-announce/2005-June/000093.html

As for Intels DRM, if it affects CD's theres always iTunes Music Store - just got my first credit card and its waiting for iTMS tracks I've been longing for almost 4 years. Also iPod Mini for the wife, and for only two weeks the MiniDisc Recorder Player for me - 15 discs of old music that I must keep to transfer onto HDD - then U2 iPod.

Now the only worry I still have is, although Apple had a secret project of OS X for x86; does that mean that its been kept up to date by purchasing the high-end PC to run it on to see if code needs to be altered to take advantage of things like HyperThreading or whatever Intel called the 1st P4 Extreme Edition. Also, what of Intels dual-core implementation and x86-64bit technology; that is all brandnew technology.

Will OS X x86 take FULL advantage of this? I really hope so. Oh yeah if they keep the PowerMac design with its 1st Intel cpu/board, which I highly doubt, will it be able to have more HDD's!!??

To all the coding experts, and Mac faithful. Thank you very much for informing me on being positive and why to stay that way. :D Acceptance stage acheived. I just hope that apple keeps the daughter card like cpu drop in; dont know why I like it.

Future thought. say in 4 years would it be safe to say that IBM may sell to 3rd party aftermarket cpu upgrades newer G5's like they did for G3's??
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.