2006 vs 2016 MacBook Pro announcement

[sunapple]

I came across the announcement video of the original MacBook Pro from Macworld 2006. This was ten years ago and now that the new 2016 MacBook Pro has been announced I thought it would be interesting to compare these two product launches. Especially since people seem to be convince Apple has lost it's core values. The Macworld keynote didn't look all that different from the Apple we know now though...

So in 2006 Apple started the transition to Intel with the iMac G5.

But there was One More Thing... While they weren't able to update the PowerBook G4 to a G5, they did announce the MacBook Pro with Intel processors. I wasn't an Apple user back then, so I don't really know how people felt about this switch (although I believe some people still want that PowerBook G5 for some reason). The numbers didn't lie though, performance per watt was 4.5x higher than with a G5 and it made this new MacBook seriously quick.

The MacBook Pro had a 15" screen which was Ultra Bright. As bright as a Cinema Display!

Of course, while being the fastest, it was also the thinnest and lightest MacBook Pro yet. The highlighted features include: iSight camera for conference calling, IR-receiver for Front Row and the MagSafe port.

Initially, these two configuration were available...

What struck me were the similarities with the 2016 MacBook Pro announcement. On this moment the machines have not yet shipped, but the forums are filled with hatred and disappointment.

The 2016 MacBook Pro is the thinnest and lightest yet. The display is lots brighter than before. Performance has increased, not just processors but also in SSD read/write speeds. It has a new, more stable keyboard and a 2x larger trackpad. The USB-C ports support a very wide range of I/O. Speakers are better. Internal cooling has improved. Touch Bar changes input methods on laptops for the first time in... decades?

The announcement from 2006 had Steve Jobs, a brilliant speaker who actually did the whole announcement including demos. Everything carefully coordinated and timed. The new features were presented as really practical, you could immediately see how this could be something you'd want to have.

I don't think the two machines differ that much if you look at the amount of new features and improvements. The 2016 keynote itself wasn't anywhere near as good and exciting, that had a lot to do with presentation. The machine did have some really nice improvements, but we've seen similar improvements for years now across all products. Thinner, lighter, faster. I totally think Steve would have presented the exact same machine if he were still alive, he would just make it seem more exciting than it actually is.

Yes, the new MacBook Pro doesn't have all ports available out-of-the-box and the prices are much higher than before (don't know how they compare to the 2006 model?). But is it not a "Pro" machine anymore just because they showed the Touch Bar displaying emoticons? I think the 2006 presentation included just as much examples for normal consumers and that they didn't change too much from that.

Of course, in 2006 they still offered a tower Mac Pro, but at Macworld they also announced... iPod Remote. iPod accessories must have been the Watch bands from the pre-iPhone-era.

Maybe I'm being not critical enough, I'm sure someone will correct me if I missed the point. This was just my opinion, something I noticed while watching the Macworld keynote. Let me know what you think!

 

[psingh01]

Well if today's Pro was 4-5 times faster than the last gen people would be pretty happy despite the ports. But its not.

 

[sunapple]

Well if today's Pro was 4-5 times faster than the last gen people would be pretty happy despite the ports. But its not.

That's not really realistic as long as they stick with Intel. There might be a switch underway, but I wouldn't expect that kind of improvement. Not as long as you update every year at least.

 

[jjlannoo]

What they needed was Craig on stage more. He should head the keynotes, but I know that won't happen. He brings a energy and humor the rest of the gang lacks. He also comes off honest whether that is the actual case or not.
[doublepost=1478452878][/doublepost]

That's not really realistic as long as they stick with Intel. There might be a switch underway, but I wouldn't expect that kind of improvement. Not as long as you update every year at least.

Yeah i think they will move to their own processors eventually but I think we are several years away.

 

[hj576]

Well if today's Pro was 4-5 times faster than the last gen people would be pretty happy despite the ports. But its not.

That would kinda defy physics.....
Its not really possible anymore. Moores law does not apply any more. You cant really have 4-5 times faster CPU between generations. The best you can hope for, is a bit better performance, and better efficiency

The only way we are going to have a major improvement in performance is through quantum computing (so not any time soon )

Seriously expecting something that is not possible is absurd, and getting mad because it didnt happen is even more absurd

 

[bjet767]

The law of speed increases in CPU is over so what will be the advancements for the future?

Relatively speaking it was far easier for Jobs to be technically "innovative" in 2006 than it is today.

Today's band wagon is full seamless integration of devices and not just at home or in the offfice, but world wide.

In 2006 one couldn't use their iPhone to check in for their flight and now it's also your boarding pass.

 

[jeremiah256]

I don't believe there was any sacrifice of ports going from a PowerBook to a MacBook Pro or they were minor. More power, MagSafe, and same ports verses what some are upset about with the 2016 MBPs.

As for catering to regular consumers, regular consumers, buying multicolor iMacs, clamshell iBooks and iPods are what saved Apple. I have nothing against putting regular consumer needs first. But:
- I've been using 16GB in my 2011 MBP since about late 2013 and would benefit from 32GBs,
- PCs are the platform of choice for VR (VR is something I want to explore)
- the only current Thunderbolt 3 docks guaranteed to work with Sierra are expensive monitors,
- most of the software I use is cross platform,
- this would be the first time I'd be purchasing a Mac I can't upgrade, so I'd have to spend double immediately to future proof my computer (and it still won't have 32GB RAM)...
I think it is not unreasonable that I take time to assess my options and decide whether my particular needs can compete with what is best for Apple's primary customer, the regular consumers.

 

[cmaier]

That would kinda defy physics.....
Its not really possible anymore. Moores law does not apply any more. You cant really have 4-5 times faster CPU between generations. The best you can hope for, is a bit better performance, and better efficiency

The only way we are going to have a major improvement in performance is through quantum computing (so not any time soon )

Seriously expecting something that is not possible is absurd, and getting mad because it didnt happen is even more absurd

You actually can make much bigger leaps in performance. Intel just can't (or won't). You don't need quantum computing. You can do wonders with lithography (which intel is now falling behind tsmc), low-k dielectrics, strained silicon, soi, etc. You also need to stop using asic design and use hand design like AMD used to do (and which I assume apple is doing based on who works there), which buys you 20% right there.

 

[bjet767]

You actually can make much bigger leaps in performance. Intel just can't (or won't).

I always love it when blog posters know more than engineers whose job is to provide stable and reliable products, like fast CPUs, to the innocent public users.

 

[cmaier]

I always love it when blog posters know more than engineers whose job is to provide stable and reliable products, like fast CPUs, to the innocent public users.

I was an engineer whose job was to provide fast CPUs. See, e.g. Maier, Cliff A. et al. (1997). "A 533-MHz BiCMOS Superscalar RISC Microprocessor". IEEE Journal of Solid-State Circuits, Volume 32, Number 11, pp. 1625–1634, and https://www.ecse.rpi.edu/frisc/theses/MaierThesis/.

I also worked at AMD for 10 years and designed many a microprocessor.

 

[knudsen2]

rofl

 

[bjet767]

I was an engineer whose job was to provide fast CPUs. See, e.g. Maier, Cliff A. et al. (1997). "A 533-MHz BiCMOS Superscalar RISC Microprocessor". IEEE Journal of Solid-State Circuits, Volume 32, Number 11, pp. 1625–1634, and

Ok then its all management's fault that they turn out inferior CPUs. If it was left up to you engineers everything would be perfect.

Give me a break. Egos.

 

[cmaier]

Ok then its all management's fault that they turn out inferior CPUs. If it was left up to you engineers everything would be perfect.

Give me a break. Egos.

Not sure what your problem is, but your entire thesis seems to revolve on all companies being equally good at CPU design. They are not. Times have changed. Most of the best VLSI engineers are working at places like Apple, which is why their chips are improving at such a rapid rate. And intel no longer has to worry about competition from AMD so they aren't incentivized to push as hard as they used to.

The end result being that intel chips could be much better than they are, and the reasons they are not are not due to some fundamental physics limitation, which is what the original argument was.

As you point out, engineers understand this stuff much better than random forum posters, so I have provided my engineering perspective as someone who designed chips ranging from the exponential x704 to ultra sparc v to AMD K6 to the original opteron. I understand you're embarrassed, but claiming my forum contribution, which provided very specific suggestions as to how intel could get performance improvements, is "ego" is uncalled for.

 

[psingh01]

That's not really realistic as long as they stick with Intel. There might be a switch underway, but I wouldn't expect that kind of improvement. Not as long as you update every year at least.

No, but you painted a picture as if both announcements were similar when they clearly aren't. 2006 had a huge performance increase that cannot be ignored.

Yet it is still possible to improve performance, stick another CPU in it. That would get in the way of them trying to make the laptop more like an iPad but I'd like to see real innovation here...

 

[Pootmatoot]

I came across the announcement video of the original MacBook Pro from Macworld 2006. This was ten years ago and now that the new 2016 MacBook Pro has been announced I thought it would be interesting to compare these two product launches. Especially since people seem to be convince Apple has lost it's core values. The Macworld keynote didn't look all that different from the Apple we know now though...

So in 2006 Apple started the transition to Intel with the iMac G5.

View attachment 670638

But there was One More Thing... While they weren't able to update the PowerBook G4 to a G5, they did announce the MacBook Pro with Intel processors. I wasn't an Apple user back then, so I don't really know how people felt about this switch (although I believe some people still want that PowerBook G5 for some reason). The numbers didn't lie though, performance per watt was 4.5x higher than with a G5 and it made this new MacBook seriously quick.

View attachment 670640

The MacBook Pro had a 15" screen which was Ultra Bright. As bright as a Cinema Display!

View attachment 670641

Of course, while being the fastest, it was also the thinnest and lightest MacBook Pro yet. The highlighted features include: iSight camera for conference calling, IR-receiver for Front Row and the MagSafe port.

View attachment 670642

Initially, these two configuration were available...

View attachment 670643

What struck me were the similarities with the 2016 MacBook Pro announcement. On this moment the machines have not yet shipped, but the forums are filled with hatred and disappointment.

The 2016 MacBook Pro is the thinnest and lightest yet. The display is lots brighter than before. Performance has increased, not just processors but also in SSD read/write speeds. It has a new, more stable keyboard and a 2x larger trackpad. The USB-C ports support a very wide range of I/O. Speakers are better. Internal cooling has improved. Touch Bar changes input methods on laptops for the first time in... decades?

The announcement from 2006 had Steve Jobs, a brilliant speaker who actually did the whole announcement including demos. Everything carefully coordinated and timed. The new features were presented as really practical, you could immediately see how this could be something you'd want to have.

I don't think the two machines differ that much if you look at the amount of new features and improvements. The 2016 keynote itself wasn't anywhere near as good and exciting, that had a lot to do with presentation. The machine did have some really nice improvements, but we've seen similar improvements for years now across all products. Thinner, lighter, faster. I totally think Steve would have presented the exact same machine if he were still alive, he would just make it seem more exciting than it actually is.

Yes, the new MacBook Pro doesn't have all ports available out-of-the-box and the prices are much higher than before (don't know how they compare to the 2006 model?). But is it not a "Pro" machine anymore just because they showed the Touch Bar displaying emoticons? I think the 2006 presentation included just as much examples for normal consumers and that they didn't change too much from that.

Of course, in 2006 they still offered a tower Mac Pro, but at Macworld they also announced... iPod Remote. iPod accessories must have been the Watch bands from the pre-iPhone-era.

Maybe I'm being not critical enough, I'm sure someone will correct me if I missed the point. This was just my opinion, something I noticed while watching the Macworld keynote. Let me know what you think!

It would be more similar to 2006 if in 2006 - like in 2016 - 3 or 4 other companies were releasing faster, cheaper, objectively "better" machines at the same time.

Apple is acting like it is still the only game in town when it comes to the high end, and therefore has the freedom to tapdance for a while. It's isn't. It's arguably now 4th in the world of laptops, with Microsoft, Dell, and Razer creating the sort of products Apple would launch in 2006.

 

[catportal]

You actually can make much bigger leaps in performance. Intel just can't (or won't). You don't need quantum computing. You can do wonders with lithography (which intel is now falling behind tsmc), low-k dielectrics, strained silicon, soi, etc. You also need to stop using asic design and use hand design like AMD used to do (and which I assume apple is doing based on who works there), which buys you 20% right there.

WTF are you saying? I can't believe you were a CPU architect.

1) Intel's lithography process is not falling behind TSMC, Intel is the one that funds 99% of the research into lithography techniques.

2) Intel has been using strained silicon since 2004 (?) around when it was 90nm processing technology. At this point, there is very little to gain from increasing silicon stress vs yields.

3) "Stop using ASIC design" - first, I think you mean don't use automated layout, and CPU cores are not ASIC, anyways assuming what you meant, it's easy to do when we were talking millions of transistors, now we are at 4 billion transistors, it isn't feasible to layout by hand. Not to mention, the newer algorithms implemented by Cadence are around 90% performance or higher compared to custom layout.

Any significant performance improvement will require a new dielectric altogether. Silicon has more or less reached its limits. There is very little incentive for Intel to change from current silicon technology because current performance is actually fine for 99.99% of users, and anyone who really needs high performance will be using a cluster or accelerated computing, not a personal machine. Their effort is better spent reducing power consumption and adding more fixed hw for common tasks (VP9, HEVC, crypto).


---
Unrelated: quantum computing will do very little for personal computing. Quantum computing is ideal for solving NP-hard problems.
---

 

[cmaier]

Take a look at what's going on in the 10nm node. Intel has fallen behind.

As for "asic" I refer to synthesis/P&R. Synopsys costs you 20%.

As for strained silicon there is still performance to be gained. Just compare the globalfoundries/IBM device to intel.

WTF are you saying? I can't believe you were a CPU architect.

1) Intel's lithography process is not falling behind TSMC, Intel is the one that funds 99% of the research into lithography techniques.

2) Intel has been using strained silicon since 2004 (?) around when it was 90nm processing technology. At this point, there is very little to gain from increasing silicon stress vs yields.

3) "Stop using ASIC design" - first, I think you mean don't use automated layout, and CPU cores are not ASIC, anyways assuming what you meant, it's easy to do when we were talking millions of transistors, now we are at 4 billion transistors, it isn't feasible to layout by hand. Not to mention, the newer algorithms implemented by Cadence are around 90% performance or higher compared to custom layout.

Any significant performance improvement will require a new dielectric altogether. Silicon has more or less reached its limits. There is very little incentive for Intel to change from current silicon technology because current performance is actually fine for 99.99% of users, and anyone who really needs high performance will be using a cluster or accelerated computing, not a personal machine. Their effort is better spent reducing power consumption and adding more fixed hw for common tasks (VP9, HEVC, crypto).


---
Unrelated: quantum computing will do very little for personal computing. Quantum computing is ideal for solving NP-hard problems.
---

 

[catportal]

Take a look at what's going on in the 10nm node. Intel has fallen behind.

As for "asic" I refer to synthesis/P&R. Synopsys costs you 20%.

As for strained silicon there is still performance to be gained. Just compare the globalfoundries/IBM device to intel.

What is going on in the 10 nm node? Intel's Sandybridge 14 nm current leakage is less than TSMC's 10 nm. Intel has 10 nm foundries already, they won't be using it until Cannon/Coffeelake.

Strained silicon perf to yield ratio is not the same for x86_64 vs aarch64 or powerpc. 10% cpu perf gain is inconsequential to modern pc computing.

Like I previously said, modern P&R is within 90% of custom layout, which is not practically feasible for a 4+bn transistor SoC. Also, synthesis and P&R are completely independent of the acronym "ASIC", so I still have no idea what you are talking about.

None of the methods you have mentioned would come anywhere close to giving 2x perf let alone 4-5x perf when apple initially went from power arch to x86. Just stop trolling.

Ok then its all management's fault that they turn out inferior CPUs. If it was left up to you engineers everything would be perfect.

Give me a break. Egos.

this cmaier guy is just baiting, there is no point

 

[mikepro]

It was fun watching that. Made me sort of miss some of the stuff you used to be able to do, like make photo books and calendars directly in iPhoto. Also interesting seeing the ancestor of icloud sharing.

 

[bjet767]

Not sure what your problem is, but your entire thesis seems to revolve on all companies being equally good at CPU design.

My problem is with egos who post thing like you did, which wasn't an innocent "They can improve things." What you stated was a refute to the poster who stated "Moore's Law" for CPUs is dated and the gains of the past in performance are over. If you can do better, do it!

OK, I'll write it for you, "WHAT EVER!"

I buy Apple over the old days of Windows, DOS, Linux, CPM, and ... because it works, inspite of all the flaws you all seem to point out.

 

[sunapple]

It would be more similar to 2006 if in 2006 - like in 2016 - 3 or 4 other companies were releasing faster, cheaper, objectively "better" machines at the same time.

Apple is acting like it is still the only game in town when it comes to the high end, and therefore has the freedom to tapdance for a while. It's isn't. It's arguably now 4th in the world of laptops, with Microsoft, Dell, and Razer creating the sort of products Apple would launch in 2006.

Other than price (although PCs from Microsoft seem to be getting expensive as well) the MacBook Pro is still one of the best if you want a real laptop. I mean, SurfaceBook seems like a great alternative, but it's only a 13" laptop with dual core processors. Razor makes gaming laptops I believe, so you'd have to deal with 5 hours of battery life tops. Dell makes some interesting machines indeed, haven't looked in to them recently.

I don't think Apple is that far behind in this category, but I agree the competition is stronger than it ever was. The next four years will decide if people really want convertible machines or actually stick with normal laptops, right now the options have just gotten more diverse.
[doublepost=1478505660][/doublepost]

No, but you painted a picture as if both announcements were similar when they clearly aren't. 2006 had a huge performance increase that cannot be ignored.

Yet it is still possible to improve performance, stick another CPU in it. That would get in the way of them trying to make the laptop more like an iPad but I'd like to see real innovation here...

That's a difference, you're right. What I don't know for sure however is if Apple was only catching up to the competition or that the MacBook Pro was the fastes one out there. While Apple had been using a G4 processor for five years, the competition might have been using those faster Intel chips already.

So a 4-5x increase isn't realistic now when everyone uses the same gen processors, but the speed compared to the competition might have been comparable.