CeBit: PPC 970's up to 1.8GHz

[ktlx]

Originally posted by Scottgfx
Being that the industry is targeting SATA (Serial ATA) to replace IDE/ATA, expect the drives to progress in a similar fashion at standard (parallel) ATA/IDE drives.

That depends upon the company and the market they are after. Seagate seems to be going after parallel ATA replacement and so their drives look to be parallel ATA drives with serial ATA interfaces. Western Digital seems to be going after SCSI replacement so their drives look like updated SCSI drives with serial ATA interfaces.

My guess is that we will continue to see both types of drives for the time being. Serial ATA is attractive for 1U servers and it will not suprise me if other companies put serial ATA interfaces on their SCSI drives (or new drives with similar specs) for that market.

 

[GPTurismo]

I don't expect to see scsi go anywhere for a while. It's to much of an industry standard and to trusted in high end business.

Maybe better scsi, which I am hoping for :S

 

[nuckinfutz]

It's ridiculous that Apple has yet to make public a CPU roadmap for the Mac. I won't buy another Mac until they do.

I don't think you'd see this. But rather I'd more prefer an updated Hardware Roadmap. But then again that would take all the fun out of computer advancement on the Mac. It's the thrilf of NOT knowing that's exciting sometimes. I don't think you'll be buying a Mac for a long time coder if you hold fast to that statement.

I'm buying a new laptop next month, and I'm 90% decided it's going to be a cheap Intel machine and not a new PowerBook. I won't enjoy it as much, but I'll be out a lot less cash when it becomes obsolete.

Dude...98 was a WAYS ago. If you're Wallstreet has lasted that long then you definitely got mileage out of it. Now go back and compare the Wintel Notebooks from that era and see how many are in uses today. The "CHEAP" Wintel Portable that will save your time and money is mythical.

 

[Hattig]

It is odd that someone wants to buy a WinTel laptop ahead of a new Apple laptop.

Apple laptops are the best part of Apples lineup, in my opinion - much more competitive in power and price against PC laptops, when weight and battery life are factored in. Coupled with Airport, nice looks, thin cases, etc, and you are on a winner. The iBooks are very affordable indeed, and the PowerBooks aren't too bad against similarly configured laptops from Dell or Compaq (although the AMD based ones do have a definite price advantage).

I have a 1998 era PII266 based HP Omnibook. It still works fine in Linux with a minimal window manager, and it is probably worth about £300. The next laptop I buy will certainly be an Apple of some kind - I'm waiting for the next generation though

 

[Dave K]

Originally posted by JavaCoder
It's ridiculous that Apple has yet to make public a CPU roadmap for the Mac. I won't buy another Mac until they do.

If you applied this logic to the rest of the industry, you'd never buy another computer.

Well, maybe a Sun...

Remember folks, Apple no more makes the processors they'll be using than Dell, HP/Compaq, Sony, or the vast majority of the rest. You want your CPU roadmaps, scream at the manufacturers of CPU's, not the end users.

It's not like these roadmaps are anything more than rough estimates anyways. I've got a MaximumPC at home with official roadmaps from 2 years back that says I should have been running AMD Hammers almost 12 months ago...

 

[JavaCoder]

Dude...98 was a WAYS ago. If you're Wallstreet has lasted that long then you definitely got mileage out of it. Now go back and compare the Wintel Notebooks from that era and see how many are in uses today. The "CHEAP" Wintel Portable that will save your time and money is mythical.

I know that buying a comodity wintel laptop is a totally different value proposition, but I could buy one today with what I still OWE on the WallStreet (which, after many years of faithful service, does no longer boot).

But my point really is that Apple has to be on the cusp of making an major, evolutionary change to a different CPU. Wheter it's a departure from the PPC family or the switch to 64 bit. This is a different, more fundamental, change than just roadmaping where, for example, commodity x86 chips are going to be a year down the road.

I liken this change to the switch to OS X. Once apple committed to the OS X strategy, the did a good job letting thier customers and developers know what's ahead -- remember "all life will be Carbon based"? Deciding to spend more money on an Apple than on a wintel machine, requires you have faith that that Mac is going to live up to what you expect. Past performance is no guarantee of future returns, as they say. I'd like to base my decision on, at least, Apple's promises than on rumors.

 

[kenohki]

Originally posted by JavaCoder
I know that buying a comodity wintel laptop is a totally different value proposition, but I could buy one today with what I still OWE on the WallStreet (which, after many years of faithful service, does no longer boot).

I know this sounds insensitive but you but you can't blame Apple for the decision you made on financing the thing.

But my point really is that Apple has to be on the cusp of making an major, evolutionary change to a different CPU. Wheter it's a departure from the PPC family or the switch to 64 bit. This is a different, more fundamental, change than just roadmaping where, for example, commodity x86 chips are going to be a year down the road.

Number one, I don't think Apple is going to move away from the PowerPC architecture. Number two, the "switch to 64 bit" isn't really that big of a deal. Sure, you can use more memory and do "big math" with large integers. However, the PowerPC specification has been 64 bit from the get go. It has just been the case that all implementations of PowerPC so far have been 32 bit. The move to a 64 bit implementation isn't inherently "better". Heck, you could have a 64 bit G4 if Mot would design one. It would work with large integers and large addresses but it would still perform the same as the trusty 32 bit G4 we have right now.

The magic that's going on right now is that IBM has taken an interest again in the PowerPC platform. Sure sure, they had been making revisions to the PPC750FX but their big R&D money was going into the POWER series where they make the big money. All of the sudden, blade computing is in vogue and they decide to let the POWER4 technology trickle down into the PowerPC line.

As long as IBM keeps letting this trickle-down effect happen, PowerPC should stay competitive. Heck, I've read interviews with members of the POWER design team where they've said their goal is not to let Itanium survive. That's a pretty hefty task but IBM hasn't been behind the curve thus far, so I'm fairly confident in their abilities to provide top notch performance from the POWER/PowerPC line.

Other than POWER/PowerPC, the only other viable long-term choices are x86(-64) and Itanium. If Apple moves to x86 or x86-64, there is little chance for them to differentiate themselves from commodity hardware on a value proposition. There is also no chance in hell that they'll ever get ahead of the commodity market in performance (take into account their past performance up against Power Computing). And, if they moved to Itanium, you'd be paying the price of a loaded G4 just for the damn microprocessor. With PowerPC though, they have a chance to be different, and they have a chance to regain some ground performance-wise. Plus, if IBM's claims about the POWER5 performance are true, PowerPC may even take back the speed crown in the desktop arena someday.

And this stuff is all taken out of the platform context. Just imagine the amount of flack from developers if Apple were to try and do a multi-platform strategy with OS X. It just doesn't make sense for them to move anywhere with their current business model.

 

[-hh]

Re: geez...

Originally posted by nickgold
To the guy working on large (several foot) graphic files at a high resolution, complaining that your processing tasks go slowly...

If this was aimed at my prior comment, the " symbol means inches, not feet. As in an 8 inch by 12 inch high quality photographic print done digitally instead of via analog chemistry.

The reason I do this is quasi-financial: assuming that my time and the PC are sunk costs, I can have more control over the printing process, at a lower (admittedly "perceived by wife") cost. My alternative is to continue to pay the big bucks for an Ilfachrome (formerly Ciebachrome) process print. This is an older and not-particulary-standard print enlargement that is made directly from a slide without the resolution losses resulting from the use of an internegative. Putting the technical nuances aside, the important take-away is to realize that they are quite expensive: $50 for an 8x12 inch print, and $100 for an 11x14" print are *good* prices.

A mere five 11x14 prints recoupe the cost of my medium format HQ inkjet printer. A dozen more replaces my exisiting Nikon LS-1000 scanner with a brand new Firewire Nikon LS-4000 film scanner. And a score more buys a new Mac.

Since the Mac & printer are "dual use" items, a good chunk of this hobby investment is pretty easily rationalized.

It's like me buying my ECHO, and instead of being happy to have got a reliable, perky little efficient car, totally whining about the fact that it doesn't drive like a Porsche.

vroooooooooooom!

No offense intended, and I mean that. But come on. My first computer was a C64. And when you compare what we have today to what we had 15 years ago or so, well, frankly, the differences between a G4 1.4 dual and a P IV 3 GHz don't really seem all that significant. Sometimes, a little more perspective is in order.

Sure. And what has changed for me is that money that I would have been previously tempted to dump into my local camera store - - or not done because I couldn't afford it - - is redirected into my PC system as physical assets that I can potentially use to get more bangs for the same net amount of bucks.

I know that the temptation is the "we do it this way because we can", and while this is true, its really more correct to say that the advancements in technology gives us additional options (solutions) from which we can then select that which is the lowest cost, best value, or some other metric that we place value upon.

As this all relates to the Mac, the real-or-perceived hardware performance lag versus PC's makes it an issue of frustration. The 970 is today's "promised land" on the horizon, but most of us have been disappointed enough times by Apple to not get too optimistic until that ship actually arrives.


-hh

 

[BenRoethig]

Originally posted by JavaCoder
I know that buying a comodity wintel laptop is a totally different value proposition, but I could buy one today with what I still OWE on the WallStreet (which, after many years of faithful service, does no longer boot).

But my point really is that Apple has to be on the cusp of making an major, evolutionary change to a different CPU. Wheter it's a departure from the PPC family or the switch to 64 bit. This is a different, more fundamental, change than just roadmaping where, for example, commodity x86 chips are going to be a year down the road.

I liken this change to the switch to OS X. Once apple committed to the OS X strategy, the did a good job letting thier customers and developers know what's ahead -- remember "all life will be Carbon based"? Deciding to spend more money on an Apple than on a wintel machine, requires you have faith that that Mac is going to live up to what you expect. Past performance is no guarantee of future returns, as they say. I'd like to base my decision on, at least, Apple's promises than on rumors.

Very true. You don't see Dell or Gateway mapping out their future. You don't know if they're going to use Intel, AMD, or VIA processor until they formally announce the product. All you can do is look at the roadmaps of the CPU manufacturers and guess. The Mac side is the same way. IBM and Motorola are the processor suppliers. They make the processors and Apple chooses which ones to use.

 

[copperpipe]

funny...

<B>That is a little odd
my computer dies when im working on my artworks. 20X30in at 266 dpi

i have a dual 1.0ghz
with 1.5 gigs of ram and im working off a raid

the photoshop file alone takes up 1.6gb of ram when working on it (not file size)

and the whole system ends up using the swap and my computer goes to a hult, but only 10% cpu power is being used


i dont see how you you can cruse with that machine on that big of a file, you only have a max of 2 gb ram.</B>

Weird. I was just working on my latest project, an image 24 x 36 inches at 300 dpi (in RGB of course), total file size 222.5 mb. I have a duel 867 with 768 Ram. And this image is really intense, no room for good compression. I was using my Wacom and it worked fine, and I applied some fairly intense filters to the entire image and they didn't take very long at all. Maybe it's a perspective thing. 30 - 40 seconds doesn't seem long to me. Especially when you design smart, you work out the process on a smaller image, and that way you only apply the filters to the large image once. I can wiat 30 seconds I suppose. I never had my machine go to a halt.

Anyway, everyone is jumping all over me for my comments. Sure speed is good and all, and I always want a faster machine too. Maybe I'm more patient than most people. Or maybe I get fed up by how noone is ever satisfied with what they got. I'm just happy as anything with my machine, - it's fast, the OS is unbeleiveably great, it never crashes (a wet dream of only two years ago), and I ENJOY it. Smell the roses.

And whats, funny, is that I just realized that while i was working on that huge graphic, I also had Safari, iTunes, and InDesign running at the same time. Whoops, those filters could've been shaved off by probably 10 seconds each. Oh well. So much for me designing smart, eh?

 

[Scottgfx]

Originally posted by ktlx
My guess is that we will continue to see both types of drives for the time being. Serial ATA is attractive for 1U servers and it will not suprise me if other companies put serial ATA interfaces on their SCSI drives (or new drives with similar specs) for that market.

Sure, sure. But what I've heard so far makes SATA to really be an equal to ATA in performance and still only able to have two drives on a single chain. I don't see that as a SCSI replacement.

 

[jettredmont]

Re: geez...

Originally posted by nickgold
Let's take the several foot high res graphic files as an example... Do you think they would really be THAT much faster to process on a high-end Pentium, vs. a top of the line Mac? Let's assume for argument's sake that the Pentium is twice as fast as the Mac...

In tasks I do every day, the difference between a 733 G4 and a 1.5 P4 can be as significant as 6x (30-minute compile on the G4 takes 5 on the PC), or as little as 1.5x.

This is why I don't understand these speed arguments at all... To me, when I'm in the market to get a new machine, a 50%, 100%, or even 200% increase in efficiency is NOT enough to get me to buy a new machine. When I see a MAJOR leap in performance, so I can now do totally new things, then I upgrade. But I am just not all that excited about being able to do the same old things a bit faster.

As a professional, if $4000 (say) spent every 2 years increases your average productivity by 100%, that is a damned good investment!

As a home consumer, I can see your argument; efficiency isn't that big of a deal. But IMHO Apple needs the pro market to survive, and having trailing-edge processors is killing that market.

Besides which, the 970 puts out easily double the SPEC scores per GHz as the G4, and so in a single revision your performance gap (ie, what you have now vs what is available) might double. To me, that signals that it is time to start seriously looking at upgrading (perhaps wait for the sevond rev of the hardware).

But you know what? Tons of "pros" are still using beige G3s in their workflow, and get the job done! Sometimes being a pro is about knowing how to get the most out of the tools you already have -- not wishing for the next thing that's always going to be right around the corner.

Absolutely. However, when new tools become available, a significant part of being a pro means doing a cost-benefit analysis to see if it fits. If beige G3s can keep up with you, then use them! No reason to upgrade! But if you're spending more time waiting on your G3 than you are creating, and haven't upgraded because new computers cost too much, you are engaging in short-sighted economics at best.

 

[BenRoethig]

Originally posted by mariner77
Don't forget, one of the key advantage of a 64 bit CPU is that it can support a LOT more memory! I hope Apple's new mobo takes at least 8GB of RAM. Yes, I need it. Modern day CCD arrays can produce HUGE image files (in astronomical applicaitons) of over 400MB per image. If I have a couple of those open there goes several gigs. It only gets worse. To me, the fast speed of PPC970 is nice and all but equally important if not more is the super fast bus!!! Let the images (video, sound, or whatever) flow!

For the desktops, you're going to have 4 DIMM slots. If Apple offers a workstation, you may see up to 8 DIMM slots.

 

[php]

Re: The real question is....

Originally posted by jamilecrire
Has anyone tried to price comparable 64bit platforms? A dual 1.05GHz Sun Blade 2000 is $19,995. My guess is this is going to hurt Sun (sunw) bigtime.

Oracle is in RC 2 stage (forever) but I would be willing to bet the cost analysis of an XServe with dual 1.8Ghz versus a comparable Sun machine will scare the hell out of them (sun).

Sun is developing UltraSPARC V and VI processors that far out perform the PPC970. It's not like Sun's R&D is standing still. You're comparing what Sun is currently selling to what might happen if Apple even decides to go that route.

 

[Frobozz]

Re: yeah but

Originally posted by copperpipe
What type of work are you doing on a mac that requires more speed than that? There must be something, but I just can't figure out what it is...

3D and Video. Until we're calculating things at the speed of light, you will never have enough speed for either of these.

 

[php]

I work with digital audio, video and multiple software synthesizers simultaneously, all of which require as much processing power as possible. If I run a DAW such as MOTU Digital Performer with several softsynths and samplers while scoring to a QuickTime video, my current dual processor G4 just can't handle the load.

 

[astrocity20]

Originally posted by barkmonster
I agree.

If you configure an entry level G4 with an 80Gb HD and the dual 1.25Ghz model with a GeForce4 MX, making the only difference the CPU module, the price difference is about £320 in the UK. Without the HD increase to match things it's just over £355 more.

If it was possible to compromise and have the single 1Ghz cpu swapped for a single 1.4 GHz cpu with 2Mb L3 (not 1.42 due to the 133Mhz bus), I'm sure people would be buying the entry level system in droves if it was only another £200 or something.

A lot of people just want the fastest mac for the price, offering the fastest CPUs in a single configuration as a CTO option wouldn't exactly kill sales of the higher end hardware because they'll always be people who need that much power and can afford it.

Agreed. Apple ought to sell Macs configurable, perhaps not barebones but configurable. Regardless More people would have Macs forcing developers to devlop for that market place. Perhaps Apple should strike up deals with a few big companies and start selling processor upgrades and mobos and the likes. This would extensively broaden their market and availablilty easily. A lot of people think Macs are great it's just they can't afford a top of the line G4, selling different lines that are expandable would solve this and the added ability to upgrade to a different mobo and processor exclusively from Apple would work easily. iMacs are pretty sure but I don't see why you can't slap a different color on a G series computer and market it as a cheaper model that has expandability instead of everything being fixed and unupgradable in some aspects which can scare away potential customers.

 

[Dont Hurt Me]

astrocity makes a good point, i love the imac, why not a upgradable daughter card in it? I see apple allways competing with itself. Iam sure there are people out there that have said i want a imac but i cant upgrade it. I want a powermac because you can upgrade it but cant afford it and a display. Therefore they run out and buy some wintel crap because they get the display and a computer that is upgradable and can afford it. And Apple looses another sale because it competes with itself.

 

[Rincewind42]

Originally posted by Dont Hurt Me
astrocity makes a good point, i love the imac, why not a upgradable daughter card in it? I see apple allways competing with itself. Iam sure there are people out there that have said i want a imac but i cant upgrade it. I want a powermac because you can upgrade it but cant afford it and a display. Therefore they run out and buy some wintel crap because they get the display and a computer that is upgradable and can afford it. And Apple looses another sale because it competes with itself.

While I generally agree with you, there are reasons to not have consumer level machines customizable. Cost & Inventory. As the iMacs & iBooks stand now, Apple has a known quantity when dealing with both. Designing the iMac to be upgradable would add to it's cost to the point that it would probably break the 2k barrier on the high end.

What Apple would need is a machine that is more Pro-sumer than the PowerMac and the iMac. Something that can take 1 or 2 expansion cards and upgradable CPUs. Unfortunately, there doesn't seem to be room for it on the current cost setup, as the top of the line iMac configured for expansion (i.e. 512MB ram) is $600 less than the bottom of the line PowerMac similarly configured. Unfortunately, you need more like a $1000 difference to consider adding a category that wouldn't canibalize sales of the PowerMac. The best solution currently would be to make the PowerMac more afforable.

 

[astrocity20]

That's sorta what I was saying Making a different machine that'd look like a PowerMac throw a different color on it and make it more affordable. Mac users and people who want to be mac users only have two options at the moment the iMac that may not have the power they need or want without upgradability or... and over priced PowerMac that adds expandibility. High price just to expand. I know iMacs are more or less made to be allinone no expandability but the fact that people want an affordable Mac that they can expand goes without saying. Generally I think for iMacs to be more customizable they'd have to make a bunch of different variations or the board, which could be hard, but then again I think it's a bit to easy to throw one of 4 variations of a board in there at such a price. I don't have an imac but I know in PCs you can just get the processor and throw it out for a new one at a higher clockspeed as long as the mobo supports it. But I dunno.

 

[maradong]

The question might indeed be stupid, but as a future switcher i wonder what the 970 cpu will bring?
Is it something like a faster g4, i mean at a higher fsb or anything like that. Or has it got an entirly new core?
Last is it the future g5 ?

thx for the help.

p.s. planning to buy a 17 inch pb. Shall i wait for the 970 or buy right now ?

 

[Rincewind42]

Originally posted by maradong
The question might indeed be stupid, but as a future switcher i wonder what the 970 cpu will bring?
Is it something like a faster g4, i mean at a higher fsb or anything like that. Or has it got an entirly new core?
Last is it the future g5 ?

thx for the help.

p.s. planning to buy a 17 inch pb. Shall i wait for the 970 or buy right now ?

The PowerPC 970 is an entirely new chip, faster core, faster fsb, pretty much everything faster. It is also a 64-bit chip, so it can access more than 4GB of memory easily (but will probably not have that particular feature enabled initially). It is not the G5 as classically reffered to, but will probably end up carrying that moniker should Apple use the chip (which is generally agreed). As for if you should wait or get the 17" PB, expect that there is first a possiblity that all of the 970 hype has been for nothing and Apple will not use the chip, also that Apple will not have a 970 capable portable chipset ready for the 970, thus causing a wait for PowerBooks to use the chip, and that the 17", regardless of the delays, will be available within the next 4 weeks, while a 970 based powerbook is at least 4 months away.

And as you are a switcher to be, welcome to the Light .

 

[Dont Hurt Me]

Well cebit is almost over, i thought ibm was going to introduce a 970 server or protype or something. has anyone heard any info yet as to the 970 at cebit?