Intel's 2006 Processor Roadmap

[wdlove]

The Intel 2006 Roadmap is certainly good news for Apple. So far Steve is on target. Go Intel. Now to see the actual proof in the pudding as the saying goes.

 

[minimax]

***REALITY CHECK***

There are plenty of press releases of Jobs promising 3Ghz G5 PM years ago. In fact, I remember how many people on this forum were finaly puffed up enough to brag about Apple's superior performance and how IBM was going to destroy intel. After all the PPC RISC architecture is a vastly superior design compared the the x86 legacy CISC crap strewn together. I am 100% positive that Jobs was indeed shown a very impressive IBM roadmap otherwise he would have never made such bold claims.

As far as your description of the superiority of the Pentium-M over the G4, You are overstating reality. The G4 is only slightly slower than the Pentium-M in tests which rely heavily on FSB or GPU related tasks. However the G4 beats the Pentium-M in tasks which rely heavily on Alti-vec such as Photoshop filtering, and DVD encoding.

I believe that the PB is in desperate need of an upgrade with newer technology. But the Pentium-M is not that great. My personal experience with the applications I use put the 1.67 G4 right on par with the Pentium-M 2.1 ghz. My business partner has a Dell with this chip using Windows XP SP2. Guess what? HE is envious of my PB. Setting side by side I can encode DVD's much faster, and I can perform Photoshop tasks faster.

Finally someone sane in this thread. Those P-M-kicks-the-living-crap-out-of-G-4 claims have no ground whatsoever. All decent tests I have seen support what digitalbiker states. People are totally staring themselves blind on the slow FSB of the G4. Show me one proper test (not that vague Barefeats test) to support your claims or I call troll.

 

[akac]

When i see a rev b Macintel PB running all of the specialized audio software that i use natively, effectively, faster than ever and cheaper, yes. Until then i am upgrading the ram on my current PB in an attempt to give it a few more years of vitality.

The switch is going to be great initially for the average mac user; the person who surfs the net, listens to itunes, writes emails, etc. The people that are going to suffer are the professional level user's that require fast cpu speeds that are a necessity for their fields, not to mention all of the specialized software that may, or may not be ported to x86 in a timely matter (without bugs!)

I remember how long it took some of my audio software to make the switch from OS 9 to OS X. If a major OS upgrade took almost a year, then how long will the upgrade to a major chipset take?

Changing APIs via an OS change is FAR more difficult for a developer than a chipset change. In fact, chipset changes are dramatically easier. In many cases its a recompile. In many others, its an intern's day job for a few weeks. For a select few, its a few months of intense work.

OS changes, on the other hand, could involve a few man-years of development. This should be one of the easiest transitions, frankly.

 

[Applespider]

Either IBM was way behind on the project and has since caught up or they're lying to protect their reputation. Im betting on the latter.

Or they're telling the truth that they've produced it but their yields are so low they can only make one a week... 😉

You'll lose that bet. The press have criticised Apple for years with regards to speed (perceived or real) against the PC industry. There is no way that Apple are going to buy up a chip that no-one else wants and add fuel to that fire. I think things will become clearer after the next round of PB updates - major change and we're not going Yonah in Jan. Minor change and that looks more likely.

 

[Websnapx2]

There are plenty of press releases of Jobs promising 3Ghz G5 PM years ago. In fact, I remember how many people on this forum were finaly puffed up enough to brag about Apple's superior performance and how IBM was going to destroy intel.

That was a road map from almost three years ago, when IBM said they had 90nm down pat. I'm sure when Apple got a revised look at their road map this year, it either got worse or IBM was promising the same things that were already supposed to have arrived. Apple was supposed to have had 3 Ghz available to them in summer 2004 and they are still not here. IBM says they have all these high end chips for the powerbook and the powermac available now, but where? Have they been put in Lenovo Machines? No. They don't exist. IBM may have designs, but they can't mass fabricate them in high yields without destroying batches. Low power 90nm chips was the wall IBM was facing, and it looks like they are starting to get over it now, but Intel is prepping release of 65nm in I'm sure higher yields than IBM requires/can produce in initial shipment release.

 

[SiliconAddict]

There are plenty of press releases of Jobs promising 3Ghz G5 PM years ago. In fact, I remember how many people on this forum were finaly puffed up enough to brag about Apple's superior performance and how IBM was going to destroy intel. After all the PPC RISC architecture is a vastly superior design compared the the x86 legacy CISC crap strewn together. I am 100% positive that Jobs was indeed shown a very impressive IBM roadmap otherwise he would have never made such bold claims.

As far as your description of the superiority of the Pentium-M over the G4, You are overstating reality. The G4 is only slightly slower than the Pentium-M in tests which rely heavily on FSB or GPU related tasks. However the G4 beats the Pentium-M in tasks which rely heavily on Alti-vec such as Photoshop filtering, and DVD encoding.

I believe that the PB is in desperate need of an upgrade with newer technology. But the Pentium-M is not that great. My personal experience with the applications I use put the 1.67 G4 right on par with the Pentium-M 2.1 ghz. My business partner has a Dell with this chip using Windows XP SP2. Guess what? HE is envious of my PB. Setting side by side I can encode DVD's much faster, and I can perform Photoshop tasks faster.

SO how can you say

This just doesn't apply in my case.

And how does that tell you that IBM gave steve some glorious roadmap? You are pulling that out of the air. For all we know Steve could have gone to IBM and said HEY. When can we expect a 3Ghz G5? Next year? Cool. Thanks. Per your previous post....

How soon we forget that IBM showed Jobs an amazing roadmap too with all kinds of low power fast chips that were being tested

And again I ask where it says they..

-were testing these new chips
-Showed Jobs an amazing roadmap (Which they obviously didn't live up to.)


Show me the facts where they showed Jobs this information. If they were really up to the point of testing don't you think Apple would already have samples? Don't you think they would be working on a G5 PowerBook. Don't you think they MIGHT have held off this move for a year if IBM had their crap together?

As for the G4 vs. P-M. Please. 🙄

There is no debate on this because we haven't seen ANY recent benchmarks comparing the two chips but as I stated in a previous post go here: Link see the difference between the two chips two years ago. Then look at the benchmarks of those PowerBooks vs. newer powerbooks on that site. Then look at benchmarks comparing those Pentium M chips to newer Pentium M chips. You can extrapolate a rough ballpark figure and it shows that the M is kicking the snot out of the G4 simple because each speedbump isn't doing anything for the system other then trying to push a firehose worth of data through the PowerBooks straw sized FSB. Simply clock for clock the G4 is getting thrashed by the M. The only thing saving the G4 at this point that somewhat plays catchup like a horse trying to chase a street racer is the AltiVec Engine. What I would love to see is SSE3 extension apps go up against AltiVec Engine apps. I'm betting it would be a close race. AV is prob more "tuned" then SSE3 but the inherent performance advantage would probably be enough.
Sorry but the simple fact is that the G4 is a CPU that hit its wall about 2-3 years ago. Moto's craptastic FSB limitations is where the road met the rubber and where the PowerBook spun out. Can the car still drive? Sure but grandma in her walker is passing us by at this point.

 

[SiliconAddict]

You'll lose that bet. The press have criticised Apple for years with regards to speed (perceived or real) against the PC industry. There is no way that Apple are going to buy up a chip that no-one else wants and add fuel to that fire. I think things will become clearer after the next round of PB updates - major change and we're not going Yonah in Jan. Minor change and that looks more likely.

I wouldn't be so sure. Apple doesn't give a crap about the name itself. They may call the thing G-Mobile-lite or something. Celeron is a name. That is all. If the tech behind the name is solid they aren't going to care. Personally I think the chips they will put in the Mac Minis and the iBooks will be ULV single core Pentium M's until Fall of '06 when they update everything to 64-bit. Not that 64-bit is a big deal at this point but Apple will prob get their wares set across the board to 64-bit as fast as possible and if the chips are there and waiting....so much the better. Its somewhat humbling to see how many possible options Apple has now. Right now Apple is probably scared pissless about this tranition but once its over the possibilities will be huge for them. The number of various types of chips that can go into a Mac will increase significantly. I think the engineers at Apple prob stared wide eyed at the close door sessions with Intel when they saw the PowerPoint with their hidden roadmap. I can see some of then tittering like a little schoolgirl. I want that one and that one and that one and OH! That one over there looks really nice! 😀

 

[digitalbiker]

I am not defending IBM.

And how does that tell you that IBM gave steve some glorious roadmap? You are pulling that out of the air. For all we know Steve could have gone to IBM and said HEY. When can we expect a 3Ghz G5? Next year? Cool. Thanks. Per your previous post....

-Showed Jobs an amazing roadmap (Which they obviously didn't live up to.)

Look I am not trying to defend IBM. I think that IBM seriously let us down. However I remember being on this forum several years ago and when the first G5 PM rolled out, IBM was gold, IBM was king, Apple even came out with a TV ad showing the PM blowing away people at their desks. There were several sites with articles on how IBM with the PPC was poised to dominate.

The vaporware roadmaps that were going around the boards never materialized. I think Job's is a jerk (I'm a Woz fan) but he is a very intelligent business man. He would never publicly announce a promised 3Gz G5 powermac without a clear roadmap from IBM. Just like he never would have anounced the "Performance per watt" gauge without a clear roadmap from Intel.

However like "stockscalper" posted first and my post was a response to, "How do we know Intel will deliver on their projections". Intel has a worse track record than IBM for not performing to roadmaps. That was the point of my post. I was not suggesting that IBM has a better current roadmap. In fact, it looks like IBM muffed a great opportunity to take the lead.

 

[willyjsimmons]

Good points

SiliconAddict has some good points.

I must say that P4-Ms are a little different than just P-Ms.

P-Ms have a bigger cache, a faster FSB and are available in the 90mn architecture.

If Apple really wants to penetrate the mobile market, the ibook must drop in price. One way or another.

 

[GilGrissom]

I must say that P4-Ms are a little different than just P-Ms.

P-Ms have a bigger cache, a faster FSB and are available in the 90mn architecture.

If Apple really wants to penetrate the mobile market, the ibook must drop in price. One way or another.

I agree with you here. I have used P4-Ms and P-Ms, personally I think the P-Ms are far greater. They appear to have been thought out better and perform much better for laptops. In all the machines I've used and provided support for the P-Ms are much better. I also agree with the iBook. What that means in processors wil continue to be debated! I think its too early to say. I think everyone has good points, even the contradicting people, it could really go either way, intentionally or wrongfully forced!

 

[minimax]

As for the G4 vs. P-M. Please. 🙄

There is no debate on this because we haven't seen ANY recent benchmarks comparing the two chips but as I stated in a previous post go here: Link see the difference between the two chips two years ago. Then look at the benchmarks of those PowerBooks vs. newer powerbooks on that site. Then look at benchmarks comparing those Pentium M chips to newer Pentium M chips. You can extrapolate a rough ballpark figure and it shows that the M is kicking the snot out of the G4 simple because each speedbump isn't doing anything for the system other then trying to push a firehose worth of data through the PowerBooks straw sized FSB. Simply clock for clock the G4 is getting thrashed by the M. The only thing saving the G4 at this point that somewhat plays catchup like a horse trying to chase a street racer is the AltiVec Engine. What I would love to see is SSE3 extension apps go up against AltiVec Engine apps. I'm betting it would be a close race. AV is prob more "tuned" then SSE3 but the inherent performance advantage would probably be enough.
Sorry but the simple fact is that the G4 is a CPU that hit its wall about 2-3 years ago. Moto's craptastic FSB limitations is where the road met the rubber and where the PowerBook spun out. Can the car still drive? Sure but grandma in her walker is passing us by at this point.

So is that pathetic test (only ONE application, just ONE benchmark and NO specific system details) all you build your case on? please... 🙄 perhaps you can impress your grandmother with this brilliant and well supported argument. "tuned" haha good one, and for the firehose argument, that's such a grotesk exaggeration it shows just how objective you are. Marvellous 'paper expertise' 😀
Anybody making such claims before the first Macintel test results are in are so full of crap i shouldnt waste my time on it.

 

[willyjsimmons]

What that means in processors wil continue to be debated! I think its too early to say.

You're right about that.

The cheap me says, roll out ibooks/minis/emac with current P-Ms.
Roll out Power Macs with Pentium Ds for the power users.

Wait for the dual-core/64bit Yonahs to update the powerbooks/imac

That will give a clear separation in performance among the product line, and help to justify a premuim price on the 'top of the line' 17in.

When prices for the new chipsets adjust, move the ibook/mini/emac to single-core Yonah.

Then move the Power Mac to the Pentium EE and/or provide alternate configs still featuring the Pentium D.

The bugaboo in this is the imac.

It's slim form factor almost demands the Yonah.

The mini would almost certainly feature a 'Celeron' Yonah.

Smaller cache and slower FSB. Cheap, cheap, cheap.

 

[Surreal]

interesting points have been made about altivec and it's efficiency in progs like photoshop.

that and this info about the PM's floating point calculation make me think the PM will lead to bettor books in the overall sense..but not for Audio/video/Digital media sense...at least not for a little while.

that makes me sad. only a little...cause i still want a better front side bus than the 200mhz the e600 promises. with that said... what i wouldnt give for a 1.67 G5 processor in a powerbook. i am confused as to why this hasnt happened yet..i might even consider a slightly lower clocked proc for heats sake for the better FSB and memory addressing.

i hope one of two things happen before the end of january (i will have had enough moneyu to buy a powerbook for a couple months at that point) either a G5 powerbook--which isnt that likely-- or an intel powerbook --which on top of not being likely, might not be so great for Logic 7.

sucks for me. P.S. i know that the current powerbooks work rather well with logic..but i do 80-90 percent of my work in the box. i use reason and would use the sound generators in logic to create the bulk of my sounds. that is why i want a fat bus, that and i want the 896HD...with all of it's 192khz glory (overkill, yes..but o so gorgeous)

 

[willyjsimmons]

Developers will set the tone.

I suppose it all depends on how far along the developers are in converting their code.

If things are on or ahead of shedule, you'll more than likely see an MacIntel by Q1 2006.

If developers are dragging their feet, Apple might just have to hold on to the PPC until Q3 or later.

I would assume that the major software suites have set a high priority on completing their conversions, and don't really want to rely on rosetta.

It would just leave a bad taste in the end users mouth.

 

[SiliconAddict]

So is that pathetic test (only ONE application, just ONE benchmark and NO specific system details) all you build your case on? please... 🙄 perhaps you can impress your grandmother with this brilliant and well supported argument. "tuned" haha good one, and for the firehose argument, that's such a grotesk exaggeration it shows just how objective you are. Marvellous 'paper expertise' 😀
Anybody making such claims before the first Macintel test results are in are so full of crap i shouldnt waste my time on it.

Note how you didn't contradict anything I said with supporting facts. I could rehash previous arguments but why beat a dead horse again. Let me guess you are a PowerBook owner? I'm sorry I'm trashing your purchase. I'm not trying to be harsh, but lets face facts. The PowerBook from the hardware side of things isn't anything special anymore. Beautiful design but that is about it. Software is another matter. But as has been stated time and again software doesn't make the computer, hardware doesn't make the computer its the combination of strong software and strong hardware that makes a superior computing experience. People keep throwing out that "well it works perfectly fine for me." Well sure it does. You have been using that environment since day one. Show me someone who has gone from a top of the line Alienware laptop or a top of the line Thinkpad that has been properly tweaked to a top of the line PowerBook and lets talk about how happy they are with performance. Look I'm not trying to sidetrack the thread but in this reality the PowerBook's performance blows.

You can prove this just by watching how Apple's products always move around on their hardware page when they get "updated".
http://www.apple.com/hardware/

With the last year to a year and a half of speedbumps the PowerBooks haven't moved in the list. Its: Desktops, Digital Life, Portables.

PowerBooks haven't been at the top for a LONG time. I will bet you a new pair of iPod inear headphones that when the Macintel PowerBooks show up they will once again lead the list. Why? Because Apple knows they finally have something to crow about. As it stands. They don't have jack.

 

[beatle888]

I suppose it all depends on how far along the developers are in converting their code.

If things are on or ahead of shedule, you'll more than likely see an MacIntel by Q1 2006.

If developers are dragging their feet, Apple might just have to hold on to the PPC until Q3 or later.

I would assume that the major software suites have set a high priority on completing their conversions, and don't really want to rely on rosetta.

It would just leave a bad taste in the end users mouth.

im sure Quark will drag their feet. they almost left the mac platform during the OSX transition. i actually hope they do it again so more people will continue the push with indesign.

 

[SiliconAddict]

I suppose it all depends on how far along the developers are in converting their code.

If things are on or ahead of shedule, you'll more than likely see an MacIntel by Q1 2006.

Well yes and no. Apple isn't going to wait forever on developers. This is why I'm thinking Announcement at MW in January with a 2-3 month wait on shippng time. Sort of Apple's way of cracking the whip at the developers. When its all said and done I'm guessing Apple will have given developers nearly a year (But not 1 full year.) to convert their apps before the first x86 system hits consumer's hands.
Do we even know where some of the developers are at with converting their apps? Someone needs to start a page that lists all the popular apps on Apple with with a red yellow green light next to them to designate the stage where the app is in regards to PPC - x86 conversion. maybe www.Macx86status.com or something.

 

[willyjsimmons]

That sounds logical

This is why I'm thinking Announcement at MW in January with a 2-3 month wait on shippng time. Sort of Apple's way of cracking the whip at the developers.

That makes sense.

A thread tracking the progress of conversions would be cool.

Adobe will be the first, I'll bet.

Followed by others who already support x86.

Adobe already has x84_64 products in the 'pipeline' so they're not starting from scratch.

 

[digitalbiker]

Best thing I have heard you say

Do we even know where some of the developers are at with converting their apps? Someone needs to start a page that lists all the popular apps on Apple with with a red yellow green light next to them to designate the stage where the app is in regards to PPC - x86 conversion. maybe www.Macx86status.com or something.

This is the best idea I have heard spewn forth from this thread. It would be nice if one of the rumor sites could get the inside scoop on major software vendors and maintain a page like this.

It would allow all Apple users to make informed decissions on switching to Intel. Something I am still a bit concerned about considering that Apple now really will be competing "Apples to Apples" when comparing Apple to Dell, Sony, Hp, etc.

 

[Websnapx2]

BBEdit Just released an update with Universal Binaries
See

 

[aegisdesign]

BBEdit Just released an update with Universal Binaries
See

SubEthaEdit beat them to it. v2.2 was universal.

I'm sure there's more universal binaries out there. It'd be cool if MacUpdate or VersionTracker picked up the baton and added a special section/flag so you could search on this.

The Intel roadmap looks ok for mobiles, particularly as G4 was taking too long and G5 was nowhere, but I'm still more than underwhelmed about the desktop lineup. Massive caches are nice but sacrificing the G5's media crunching potential for better laptops isn't going to sit well with pros having to wait longer for a DVD to encode. If IBM makes it out with dual-core G5s with bigger caches then an expensive Yonah based Xeon or Pentium D is poor recompense for being as fast as the other guys but no more.

If only St. Jobs had proclaimed the transition as being a transition to a multi-platform future and not a route-march into the land of the Intel-ites.

 

[xStep]

Adobe will be the first, I'll bet.

Already some developers have released PPC/Intel fat binaries. So, nope, Adobe will NOT be first.

 

[Websnapx2]

I guess at the end of the day, It's tough to see what anything will perform like since we don't have any info other than a Chip manufacturer and a hunch. We have no guarantees for speed, chip family, MOBO manufacturer/configuration. We're grasping at straws, because it fun. But if the software doesn't work on it, I can get a Intel PB. Everything I use requires Altivec (apple's pro apps will be fine, I'm sure, but what about photshop. I just got CS2). When they deal with that hump (as in a solution for my current software set up to run at a respective speed) this can become more than just entertaining banter.

 

[Surreal]

If only St. Jobs had proclaimed the transition as being a transition to a multi-platform future and not a route-march into the land of the Intel-ites.

the problem with this would be what to optimize for?

if you optimize for ppc because you develop a media app you might in essence be saying "buy a desktop comp if you want worthwhile perfromance from this app." i know it isnt that simple, but i would be a little sour if it came to "this program is snappy if you are on a desktop" or vice versa

 

[willyjsimmons]

Thought we were talking about 'pro' apps

Already some developers have released PPC/Intel fat binaries. So, nope, Adobe will NOT be first.

We were talking about 'pro' warez, not rinky dink editors.

Jobs already said that some apps would simply require a recompile.