AMD new licensee, Apple

I saw those early benchmarks, and they did seem to show that the DDR RAM was no help. But in MacAddict's latest issue (if memory serves) contains a review and benchmarks that suggest otherwise. MacAddict contends that there is a small but significant improvement in the new DP 1 gHz models, despite them having less level 3 cache (again, from memory).

Personally, I'm looking for a 1gHz iMac, thanx. Santa???
 
Originally posted by Nalmar

6- And last but not least, discussion came to 64bit over 32bit processor. One thing I want to make clear is that 64bit IS A DISADVANTAVE. It will be necessary one day to overcome the 4gig addressing space limit but it WILL BE A SLOWDOWN. 64bit means bigger memory requirement ( memory is cheap but also think that it also occupies more space in caches ie less element in cache so more cache miss ) slower throughput ( a 128bit bus can move 4 32bit chunks or 2 64bit chunks ) and 64bit integer treatment is mostly useless on consumer machines. Is there a programmer here that can tell me he uses 64bit integer intensively in a consumer application ? The today 32bit G4 uses 32bit address, 32bit instruction and 32bit integer registers and ALU ( arithmetic and logic unit ) but has 64bit floating-point unit and register and 128bit vector ( altivec ) unit and register.
Click to expand...

And all the naysayers wondered why we moved from 16 bit to 24 bit Macintosh System Software. They also said we didn't need to go from 24 to 32 bit.

What happened in the move from 16 to 32 and all between? We got more colors on our monitors, a better finder, bigger hard drives, video playback, more desktop publishing abilities, 3D rendering and the list could go on.

We do not and cannot know what the advantages of 64 bit really are until we move and someone makes the killer APP to take advantage of it. What about PCs? Remember 16 bit windows? See how it got so much better when MS went to 32 bit and Intel moved from 486 to Pentium ? Thought the 486 was a 32bit chip, it didn't have a large enough bus to handle a true 32bit Operating system. Once we went 32bit, we saw preemptive multitasking, threading and more.

One thing that would see an immediate boost are databases right? Well what about swap files and journaled file systems? Aren't those essentially databases of memory and file systems? I don't know about you but Mach and the current HFS implementation on MacOS 10 is slow. Something has to be an issue there and I doubt it is the IDE controller or slow drives. The drives are the same ones tht ran on MacOS 9 and yet MacOS 9 has a faster disk I/O.

No, all those improvements like a jfs, Aqua, Quartz are not a disadvantage just because they make the current system slow. Why does everyone think that just because something bogs down your cpu it is bad? It's good but yes it will be slow and that is why it's time to move on and leave 32 bit in the past. Good riddens. I want to see what wonderful new inventions come out of 64bit and I bet you we have a new revolution in killer apps and User Interfaces for the GUI or whatever will replace the GUI. The desktop paradigm is dated and old and frankly, it deserves to exist only in one place:

My PDA be it a pocketPC or PalmOS. I think its laughable that those things are catching up to the desktop. 200-400 mhz 32bit ARM risc processing? Why don't we just all start carrying PDA's and forget using computers anymore and just have a VGA port coming out of the PDA? Geez. At least it would require less cooling.
 
i am a lowly college student who doesnt actually own a mac, but i do use them at work. dont hurt me. i'm saving up for one.

personally, i am against moving to the x86 platform. for one, i think of mac folk of more of a community than any pc users, windows or other. and i think that comes mainly from the fact that everything is so different from the hardware up. once you go x86, you lose that and apple loses their mystique.

there is an upside in that the price of mac hardware would go way down. but you know what else? so does your resale value. macs kill in resale value. x86's are a dime a dozen. not to mention that you will then have joe hacker putting osx on his ugly as hell box that he built from 8 different junk pcs. not the image mac wants. half the reason i want a mac is because of the way it looks. take that away, and i'll make due with whatever is handy. actually, i'll probably just borrow my friends copy of osx and put it on my gateway laptop (not really, but you see my point).

all of a sudden apple is nothing unique. just another niche os in a windows dominated market. stick a fork in em. they're done.
 
In response to Ed's request for reasons why excessive speed is critical. Here's why I need it:

1. Lightwave rendering. Yes, A same MHz G4 is actuallyslightly faster, but guess what? The PCs are over twice as fast now.

2. Photoshop. In spite of what Apple tries to shove down our throats, I've run a battery of tests and the MHz myth is not a myth. It's a reality. Pound for pound, test after test, a faster Intel chip will win virtually any speed test. For those living in their fantasy world that believe otherwise, I would question whether they've spent the time that I have doing these tests. I work on several hundred megabyte files, I need every ounce of speed I can get.

3. Video rendering. I have not edited video on a PC, but I've got to believe that a PC that's twice as fast is going to be faster than the G4, Altivec or not. Having said that, my DP machines are pretty impressive at rendering video. DV especially.

The rest of my work is 2D design and web design, so I have to particular need for increased speed, though it can always be faster to launch apps etc. Livemotion runs like crap. I'm switching to Flash after last night's edit. Piece of junk software. I hear Flash development is way faster on a PC.

I also do quite a bit of audio engineering, but currently my demands are easily handled by a mid-range Mac. CuBase SX BTW runs great in X! Very smooth, very snappy. Good example of how an X app should run. A bit buggy yet, but it's a total rewrite, so I'll cut em some slack for 1.0

So, in terms of need for speed, I can only think of three areas, but they are critical for me, so MHz is a big deal. I shouldn't say MHz. I could care less what the MHz is as long as it's fast. If you can make a CPU that runs at 200 MHz, but has a wide bus, DDR ram and all the fixin's including very optimized code, that runs faster than a DP 1 gig, bring it on.

I'm giving Apple two years. (I try to hold on to a machine for two years before buying a new one). If, after two years, there has not been a significant improvement in closing the speed gap, I will very seriously consider bailing. I want a lightning fast GUI, quick launch apps, and fast rendering/processing. If Apple can't give that to me, I'll find someone who will. (This comes from a 15 year Mac vet, who literally loves his machines).
 
Originally posted by MacLuv
Hi Ed :)

I was just going through the thread again and taking it all in, I thought I'd use your post here to illustrate a very important part of business ...
:)
Click to expand...

Hi Macluv :)

nicely presented and i wouldn't argue with it. believe it or not, i once earned a 2 year degree in merchandising/marketing so none of this is new to me.

but i think apple makes a lot more profit on their low end normally. only recently has the low end become an old style imac 500 for $649 (source: latest maczone catalog). which by logic tells me that apple must have made quite a bit of profit off me when i bought the low end imac 400 for $1299 2 years ago. i'm pretty sure that maczone can't sell them for less than what they pay apple for them and that apple isn't selling anything below cost yet. Apple has never shown signs of ofering 'loss leaders'. add to this the fact that you all are telling me there is nothing so special about mac hardware other than it's ability to run apple systems and i have to figure that even the low end models are making a fair share of profit.

and note that the current advertising budget is being spent to target consumers. apple isn't marketing speed. they're marketing something muchless tangible and much more valuable - a simple and effective user experience. and it seems to be working. we get people here regularly (like cf25) who are looking into macs and os x. all the mac sites are seeing these potential switchers poking around and trying to learn more.
They don't ask questions like "how many mhz does a mac have, or how fast can a mac render a complex photoshop task. they ask if it can do a certain common task they are used to on windows, or whether they should buy now or later.

mindbend - thanks for your input - also very well presented. i 'm still waiting to hear the same from the people who are arguing so loudly here. i can appreciate your stance, but i would ask, as you seem to be asking yourself, how much faster does it have to be to satisfy you? twice as fast, 3x, what? or is it a matter of shaving a certain amount of time off of certain tasks you regularly perform? at what point does the trade off between speed and love of all the goood things about macs tilt towards the pc side? i'm curious.
 
Originally posted by MacLuv
So are you suggesting that Apple just hides behind a bush for the rest of our lives in hope that people will someday "come around" and buy Apple products? We've been down this road before.
...
Click to expand...

No... I'm not saying that apple shouldn't do something... I say that going x86 comes with downsides.
1- Getting in trouble with a powerful enemy : microsoft witch has control on an important part of OSX : office. As of now, OSX wouldn't have survived without office or without photoshop.

2- Risking of losing a part of it's hardware sales. Every buyer or would-have-been buyer of powercomputing and motorola clones would buy a 400$ walmart pc and put OSX on it. I know I would.


Originally posted by terran74
And all the naysayers wondered why we moved from 16 bit to 24 bit Macintosh System Software. They also said we didn't need to go from 24 to 32 bit.
Click to expand...

Macintoshes were always 32-bit. The first macintoshes used 68000 cpu which were 32-bit processors but had a 16bit external data bus ( they were referred as 32-16 ) but were full 32 bit internally. The thing with 24bit was that they used the upper 8-bit of the addresses for status things ( I'm not exactly sure what was their purpose ) leaving 24-bit for addressing (16megs) thought to be more than enough at the time ( like the 1meg limit in original PCs ). Some day, they came to that limit and had to expand. But right now, 32-bit addressing permit 4gig of address space. Are you already choked for memory ?

And for the rest of the cpu, everything is at least 64bit except integer units and registers. And unless you see enough calculation requiring 64bit treatment, the wasted power will result in performance degradation. Databases yes but not swap files and journaled file systems like terran74 said, more like humane genome database or universe modeling.
 
a snippet i found while doing my 24 hr check on this thread -

from macluv
With all of the posts in this community, especially the ones regarding OSX vs. XP, I can, without a doubt, say that nobody can argue that OSX on the G4 platform isn't crap.
Click to expand...

well, i would argue that it's not crap on my G3 platform. so i would imagine i wouldn't see it as crap on a G4 either. i bring this up becasue it hits at the heart of this discussion in several ways. one - people assuming that everyone's perceptions are or should be the same as theirs. this quote is obiviously an opinion - the kind that is often referred to as flamebait. there is little doubt in my mind that this single statement started the flare of agressions in this thread. and yet it has no factual basis. it is only one person's opinion based upon their perceptions and expectations. two - the confusions that always arise when we start to discuss hardware. systems and software always seem to override the discussion and become the new focal points. this always makes the points unclear and leads the discussion far from the subject - which in this case is whether apple and amd will partner for further projects. not whether windows are better than osx or even if pc's are better than macs. the real question that is being raised is whether this is a realistic way for apple to move forward.

but to be clear, i'm not pointing a finger at MacLuv. i'm pointing more towards the differences in perceptions that we all bring to this discussion without having experienced the others' perceptions.
 
as a graphic designer, i'll keep it simple...

apples hardware is behind

its slow and outdated for the most part

i have a 700mhz pc running winxp sitting right next to my 800mhz g4

the pc runs pretty damned close in any heavy testing (photoshop etc)

but for you casual users, it puts the g4 and osx to shame running "everday normal stuff" everything is snapier and faster

true its not as animated or as pretty, but i couldnt care less about that, i want my tool to work and work well thats all ;)

i use macs cause i like them simply, but i sure as hell wish they were faster :(
 
Posted by krinkle cut..,.
“
I'm late to this thread, but I just had to respond to this. You'd rather NOT optimise code and pray for faster hardware? That's the solution? Doesn't that seem a little cyclical to you? Develop sloppy software so I can develop new technologies that need newer hardware, so I can develop new technologies, that will need newer hardware, so I can develop new technologies (etc. ad nauseum).”

Information technology has “competitive advantage”. Its all about having better, faster and more reliable software than your competitor so that your consumer can beat their competitors. Having the same technology as your competitor (albeit more optimized/reliable) is not neccesirly going to create a significant “competitive advantage” for your company. The other company can work more hours and match your output and so on.

However, if you can do stuff “differently” and “better” then your competitor cant compete with you, therefore you have “competitive advantage”. The software industry has found that “new technology” is what makes that happen. That’s why research and development are so important. If you don’t spend a lot on r&d you are always playing catch up and your advantage dwindles and you leave the industry (go bust).
You may develop technology so good that no one has a hope in hell of catching you and therefore the license it from you (Quake/Doom graphics engine). Its so much better to be ahead of the game.

As a company you have limited resources, so you have to decide whether your gonna spend time & money optimizing or researching new stuff. The IT industry has found that hardware speed grows at a significant rate every year (unless you’re a mac user!!) so its best for them to spend more time developing new stuff. That’s the only common sense decision you can make right now. Until hardware development slows down and your competitive advantage comes from optimizing.

Another reason against constant optimization is that your code has to be supported by other programmers who may join the company later. Optimized code tends to use short cuts and less straight forward ways of doing things just to get the speed up. You can then end up with code that is less readable and structured, making it difficult to maintain.

I’m not saying optimizing is bad, but in this environment at the moment it doesn’t make business sense to spend to much time and money on it because I piece of hardware will come out next month that fixes the problem for you. Nobody writes in machine code anymore (even though its much faster) because processors are so fast there’s no need to put anyone through that type of pain! That’s just the way it is.

An example of the struggle is this …

Apple developed OS X with 2ghz + proccessors in my mind. Theres no way they would have put so much processor intensive stuff in it if that wasn’t the case. They wanted “competitive advantage” over Windows. So they took a gamble.
Now they cant get those chips the OS looks slow compared to its competitors. All the advantages that Apple have in OS X are marred by slow response in the OS. Now they spend all their time fixing and optimizing stuff (i.e. Quartz Extreme) so that they can be on a level playing field with Windows. If Apple had gotten 2ghz + chips at the time OS X was released Apple would be laughing. Now it all looks quite silly. The r+d gamble hasn’t paid off in this instance. But how else could they beat Windows without taking that gamble? That is why they need AMD or something equivalent by the first quarter of next year or the party is over …..
 
Hi There folks :)

I personaly think that AMD is going to make RISC based processors with Altivec
AND Apple is not going to delevop MACOSX based on a Stupid X86 processor
they cant do that it will cost so much money to drop the whole PPC .

Maybe if Apple goes bankrupt they can think about the X86 Architecture...
for there last Hope..

AMD had RISC based prosescors way in the past ....
 
I think you can rule out that AMD will provide PowerPC compatible processors. AMD would have to license the technology from either Motorola or IBM. Unless one of those WANT to get out of this business, this is a no-go, since Apple is the only real buyer for desktop processors of the PowerPC variety. Yes, IBM builds their own for their own machines, but they make the desktop processors for the Macs exclusively right now, they don't have other buyers for the G3 desktop processors. This might change with the PPC 970, but I don't see that happening yet.
 
Originally posted by BuddahBobb
G3 is in the gamecube?
Click to expand...

No, it's not. The PowerPC in the gamecube is a special version of the G4 for embedded machines. It runs slightly cooler and doesn't have Altivec and performs special functions for the gamecube.
 
You must log in or register to reply here.
Back
Top