From Anandtech's IDF coverage, Intel showed some general performance expectations of their upcoming chips, including Merom:



Relative to what Apple is currently using in the Macbook Pro, the Core Duo, this chart indicates a 20% increase in power with about the same battery life. Notwithstanding the 64-bit architecture change, this seems to be a modest improvement from the Core Duo, without some of the earlier predictions of double the battery life and such. Perhaps that level of improvement will be seen one or two refinements after. And of course, this is just a simple graphic chart, but revealing since it's coming from Intel themselves.

It'll be cool to see laptops running for 6+ hours on a charge. Hopefully Merom will lead the way shortly to that kind of performance.

[julesstoop]

Maybe intel is not aiming at lowering power requirements even further, as their mobile platform is doing a fine job as is.
Besdides, they might know a thing or two more than we do about upcoming technologies like fuel cells, or more simply some next generation of traditional battery.

[Brave Ulysses]

I would think advancements in backlighting/screen technology, and mechanical parts would result in much greater battery life gains than future processor tech.

[Rolo]

Uh, the graphic indicates > (greater than) 20% increase. If the chip tops out at 2.66 GHz, that's a 33% increase in speed but even that doesn't tell the whole story. The whole chipset will be different, the CPU will be 64-bit, and the bus speed will be higher. What will the net result be for apps? What if there's a net 40% gain in overall performance, especially for pro apps? How much will the vector performance increase?

Bottom line: that simple graphic is way too simplistic so don't take it at face value. Intel isn't about to spill all the beans to AMD or do anything to hurt existing sales.

[Wickers]

Originally Posted by Rolo Bottom line: that simple graphic is way too simplistic so don't take it at face value. Intel isn't about to spill all the beans to AMD or do anything to hurt existing sales.

and probably dimming it down a bit for the people they are presenting this to... (you know, the ones that make their stock go up or down a little everyday)

IDF gets more attention from investors and media then most would like to admit. Funny how it's main focus takes a near second to it's desired result. Hype... start your engines!

[Partial]

Now, the question is whether I should wait for a Merom MBP or get a core duo MBP this summer for 200 less when I can get and sell a free iPod of some sort. Hmmm.. It will probably be worth the wait because they aren't going to add more cores or jump to 45nm for about about a year after merom's release. To me, that seems like a good time to jump rather than waiting for the latest and greatest thing.

[gsxrboy]

20% + gain for nothing? Sounds great to me. It will use lower power for the same performance as the current MBP resulting in better battery life (not just higher performance for the same power draw). Merom is where it all really starts, Yonah is just a transition for Apple & Intel.

[ZachPruckowski]

Conroe also looks good, as per this article. I think Intel might be turning a corner. Which is great news for us. Now Apple has a partner that's very intensely interested in desktop/laptop R&D, as opposed to IBM, who was mostly interested in server chip and game console chip R&D.

[Partial]

Originally Posted by ZachPruckowski Conroe also looks good, as per this article. I think Intel might be turning a corner. Which is great news for us. Now Apple has a partner that's very intensely interested in desktop/laptop R&D, as opposed to IBM, who was mostly interested in server chip and game console chip R&D.

To that I excitedly say, QFT!

[Aesahaettr]

Plus, they said since Conroe is more than six months away, then there will most likely be an even further increase in performance, and the Extreme Edition could clock to more than 3.0 ghz. I'm definitely looking forward to a 20 inch iMac with this in it!

[Partial]

what's next for notebooks after merom? When will we see quad or octo cores on each processor?

[Aesahaettr]

Hey, check this out too: http://appleinsider.com/article.php?id=1581

[Aesahaettr]

This means that computers with those chips would eventually load in a couple seconds after they had been developed further. I don't know how I feel about not having to wait for a computer to boot up. Without a wait, or at least a progress bar, I don't think there would be a sense of accomplishment for me when it was on.

[Baron Munchausen]

It could just be the very thing to paper over the cracks of a potentially glacial Vista boot experience. Even on XP2, I cannot believe people feel anything except concern when the desktop icons blink time and again as the system comes up, like some old lightbulb in a Soviet gulag.

We have been around this block before - a 1GB RAM should not break the bank and should provide a simple way to get a system up and running quickly.

I am unsure if the Intel example uses the flash memory to copy into normal RAM or it is run 'from' the flash. Methinks the former.

[oli84]

I would actually rather wait for the 45nm Merom, or at least the 800mhz bus one designed for the Santa Rosa platform. I think that's where the actual speed increase lies.

[WrestleEwe]

Originally Posted by Rolo Uh, the graphic indicates > (greater than) 20% increase. If the chip tops out at 2.66 GHz, that's a 33% increase in speed but even that doesn't tell the whole story. The whole chipset will be different, the CPU will be 64-bit, and the bus speed will be higher. What will the net result be for apps? What if there's a net 40% gain in overall performance, especially for pro apps? How much will the vector performance increase?

Yeah.. errr... hasn't your momma ever told you never to trust powerpoint slides?
"greater than 20%" means 20.00001%, simple as that. It's called statistics. Oh and my new iMac isn't 3 times faster than my G5 either, that's called marketing.

Originally Posted by Rolo Bottom line: that simple graphic is way too simplistic so don't take it at face value. Intel isn't about to spill all the beans to AMD or do anything to hurt existing sales.

Read this:
http://www.theregister.co.uk/2006/03/10/intel_heat/

I know it's The Register, so you'll have to take it with a little bit of salt, but the gist is there. Intel is lagging 5 YEARS behind in multi-core server processing, and they HAVE been downplaying BOTH performance/watt and multi-core processing until VERY recently.

The Core Duo is a fantastic laptop processor. There's no argument about that, but it DOESN'T mean that Intel is the One True Processor Mfgr in the world.

I really do not want to turn this thread in a Intel vs. AMD vs. IBM flame war, but I really think that some people here need to shake off at least a little bit of the Jobs Reality Distortion Field here and read up on some facts...

I'd enjoy discussing this with you guys (even I am wrong sometimes, and I'm not afraid to admit it), but if your argument is like the above qoute, I have one word for you: "Itanium", look it up.

[Anthem]

I wish I could dig up all of those quotes from Intel where they said "64-bit x86 is impossible. We're Intel, and we could do it if anybody could ."

[chucker]

Originally Posted by Anthem I wish I could dig up all of those quotes from Intel where they said "64-bit x86 is impossible. We're Intel, and we could do it if anybody could ."

I have to be one of the very few people who would prefer no 64-bit x86 over the craptacular AMD64/EM64T/x64-style 64-bit implementation.

[Wickers]

Originally Posted by chucker x64-style

x86_64

---

But yeah, it's just address space and marketing wrapped in a shit pile. At least one day everything will be '64-bit' and we will stop being bashed over the head with the bitness difference. Though, by then it will be some other new tech.

[chucker]

Originally Posted by Wickers x86_64

Microsoft calls it "x64", and I am thusly perfectly legitimized to call it the same.

http://en.wikipedia.org/wiki/x64

But yeah, it's just address space and marketing wrapped in a shit pile. At least one day everything will be '64-bit' and we will stop being bashed over the head with the bitness difference. Though, by then it will be some other new tech.

The main thing that royally pisses me off compared to the PowerPC's 64-bit implementation is how un-transparent the transition is gonna be. Unless Apple has some miracle plan they haven't told us about yet, it's gonna be universal binaries all over again. Can't run 32-bit and 64-bit code by side in the same way you can on PowerPC. Can't run both of them at native speed either, as 64-bit on x86 has a larger address space, which is why so many honchos think that "64-bit means faster".

You can, thankfully, have multiple binaries in one, which I suppose is what's gonna happen, but I really don't like Apple's eerie quietness over it so far.

But yeah. Tough times ahead again? Please no.

[Wickers]

Originally Posted by chucker Microsoft calls it "x64", and I am thusly perfectly legitimized to call it the same. http://en.wikipedia.org/wiki/x64

I was appending, not correcting.

Originally Posted by chucker The main thing that royally pisses me off compared to the PowerPC's 64-bit implementation is how un-transparent the transition is gonna be. Unless Apple has some miracle plan they haven't told us about yet, it's gonna be universal binaries all over again. Can't run 32-bit and 64-bit code by side in the same way you can on PowerPC. Can't run both of them at native speed either, as 64-bit on x86 has a larger address space, which is why so many honchos think that "64-bit means faster". You can, thankfully, have multiple binaries in one, which I suppose is what's gonna happen, but I really don't like Apple's eerie quietness over it so far. But yeah. Tough times ahead again? Please no.

Yeah, I've never thought of that before now. Apple will have to plan out either 'fatter' universal binaries or just run in 32-bit compatibility mode for a while. This must make developers everywhere just jump with joy over more porting... (as painless as it could end up being)

[chucker]

Originally Posted by Wickers I was appending, not correcting.

Oh.

Four different names for one and the same thing. Pretty bad, isn't it?

Yeah, I've never thought of that before now. Apple will have to plan out either 'fatter' universal binaries or just run in 32-bit compatibility mode for a while. This must make developers everywhere just jump with joy over more porting... (as painless as it could end up being)

Everything I know so far indicates that it won't be "painless" at all, sadly.

Oh well. Interesting times ahead.

[Dorian Gray]

Originally Posted by Brave Ulysses I would think advancements in backlighting/screen technology, and mechanical parts would result in much greater battery life gains than future processor tech.

The processor in my iBook uses more power than all the other components put together, and in the new MacBook Pros with Core Duo, that trend is intensified.

[Aesahaettr]

I've tried doing searches for a simple explanation of what the difference between a 32-bit and a 64-bit processor is online, but I can never find one that can explain it to me, who is new to the technical aspects of computers. Can anyone tell me in simple terms? Is it just faster with some programs than with others?

[julesstoop]

To put it very simple: It (a 64 bit processor) can count up to a higher number.
Because every location in your RAM is associated with a specific numerical address, a 64 bit chip can address much larger amounts of RAM.

[Aesahaettr]

Originally Posted by julesstoop To put it very simple: It (a 64 bit processor) can count up to a higher number. Because every location in your RAM is associated with a specific numerical address, a 64 bit chip can address much larger amounts of RAM.

So, more basically, are you saying that a 64-bit processor can have more things running for longer amounts of time?

[ZachPruckowski]

Sort of. It can address and use more RAM (up to like the petabytes or something incredibly high), and it can also get up to higher numbers when doing intense stuff, so for higher-end programs with 32-bit chips they have to do like work-arounds with 32-bit chips that they don't need with 64-bit, or something like that. I could be wrong, this is gleaned bit by bit from a variety of forums.

[WrestleEwe]

Originally Posted by TheWretched I've tried doing searches for a simple explanation of what the difference between a 32-bit and a 64-bit processor is online, but I can never find one that can explain it to me, who is new to the technical aspects of computers. Can anyone tell me in simple terms? Is it just faster with some programs than with others?

A quick Wikipedia search reveales a lot of information and the correct terminology to search further in "less biased" sources.

To make it easy:
Intels names:
1- IA64, server processing, used in Itanium and Itanium 2
2- EM64T, 64 bit extensions used in todays Pentium D and in the next rev. Core Duo
AMDs name:
1- AMD64,x86-64 and x64, used in all of AMD 64bit processors.

the short story:

IA64 was developed by Intel itself for the Itanium. You should pray to Jobs everyday so that Apple will not use that thing in their servers and Powermacs.

AMD released AMD64 a couple of years ago with their new Opteron processors, and now includes it in every proc they make.

Intel has developed EM64T to be compatible with AMD's arch. Mainly because AMD64 was becoming the de facto standard.

In the AMD64/EM64T you will not experience a performance hit when using 32bit programs (or OS's). However, you need a 64bit OS to run 64 bit binaries. With IA64 you will see a 50% performance hit when emulating 32bit.

What I predict is a 64bit version of OSX for the Powermacs and at most a few PRO apps that will have 64bit binaries included. This mainly because there are no performance benefits inherent in 64 vs. 32bit modes.

NB: 64bit programs have the disadvantage that they use more memory because things like pointers will now use 8 bytes in stead of 4.

[Robo]

Is AMD64 used in every processor AMD makes? I guess I assumed that Sempron was 32-bit.

[chucker]

What he meant was that when AMD uses 64-bit, they use AMD64. They don't have other implementations (at least in the PC market). Intel, however, has two different implementations.

That said, some newer Semprons indeed have 64-bit support.