Pentium M and Yonah Processors in Upcoming Macs?

[MontyZ]

Actually, I am one of the ones here who can touch-type.

Do most people not touch-type? I assumed more would than not. Wouldn't it take longer to get anything done typing with two fingers?

 

[egor]

Do most people not touch-type? I assumed more would than not. Wouldn't it take longer to get anything done typing with two fingers?

I think most do, anyone who has spent a fair amount of time at the keyboard two finger typing will suddenly realise that they no longer need to look and can use all their fingers anyway (like I did!).

 

[ShnikeJSB]

Hyperthreading...

Hyperthreading is a trick to compensate for the inefficiencies in the Pentium 4 / Netburst architecture. It's basically a way to utilize otherwise idle cycles due to branch mispredictions and cache misses inherent in the Netburst architecture with its insanely long pipeline.

Hyperthreading on the P-M architecture won't have as much benefit. That doesn't mean intel won't implement it, though.

I think you forget that when programs are made to take advantage of HyperThreading, there is actually a pretty big performance boost. More "processors" are ALWAYS better than one. The only time Hyperthreading slows things down is during synthetic single-task benchmarks, which is ridiculous. Hell, even Splinter Cell:Chaos Theory takes advantage of it, and it is a game! My friend's old 2.6GHz P4 with hyperthreading is definitely faster overall than my brand new Pentium-M 2.13GHz when running multiple things that we typically have open at the same time (especially when switching programs). We tend to have Photoshop, iTunes, Firefox w/ multiple tabs, firewall, anti-virus, anti-spyware, etcetera, open at the same time, and I must say that Hyperthreading DOES help, and I am sure it will help the P-M's when they get it too. Saying it won't is like saying Altivec doesn't help G5's because they are better than G4's. -JB

(EDIT: I mean my friend's computer FEELS faster because of the almost instantaneous response of the programs and switching and such while there is lag on my computer before anything is done, although it is done faster once gotten to LOL! Also, to mention the biggest pain in the butt, when a program is taking all the processing power and is hanging - on my computer it takes FOREVER and a day to force quit, sometimes even requiring a hard shutdown - while my friend's responds instantly because of Hyperthreading's more efficient use of spreading out processing power.)

 

[ShnikeJSB]

:::eating his foot:::

Well, I just looked at some more benchmarks, with both the 840D dual-core/non-hypertheaded chip AND the 840EE dual-core/Hyperthreaded chip, and in a few benchies the hyperthreaded Extreme Edition was actually slower than the NON-hyperthreaded chip... Apparently it is a problem with Windows and its suckage and not the actual processor/motherboard. So, MY HOPE is that OS X will be much better at handling all those available threads and take advantage of it as much as possible! I am betting yes, but who knows? Although I still stick with my OTHER comments in the last post about feeling faster and such. -JB

 

[wdlove]

I think most do, anyone who has spent a fair amount of time at the keyboard two finger typing will suddenly realize that they no longer need to look and can use all their fingers anyway (like I did!).

I use more than two fingers, but haven't been able to get away from looking. Still grateful for spell checker in Mac OS X.

 

[AidenShaw]

response time vs throughput

...and in a few benchies the hyperthreaded Extreme Edition was actually slower than the NON-hyperthreaded chip... Apparently it is a problem with Windows and its suckage and not the actual processor/motherboard.

Note that more often hyper-threading is a throughput booster, not a response-time enhancer.

I have a particular job that takes about 2 CPU hours on a system with HT disabled. It's 100% compute-bound, so it also takes 2 wall clock hours.

With HT enabled, I can run two copies of the job, but they each take 3 CPU hours (and the two finish in 3 wall clock hours).

So, in a day, I can run 12 jobs without hyper-threading, or 16 jobs with hyper-threading. My Opterons do about 13 per day per CPU. (3.6 GHz/1MiB Xeon, 2.6GHz Opteron)

Is it slower with HT - by one measure, yes.

Is it faster with HT - by a different measure, yes.

So, MY HOPE is that OS X will be much better at handling all those available threads and take advantage of it as much as possible! I am betting yes, but who knows?

Even the best available HT-aware schedulers will run slower on some benchies.... If those apps are what you run, leave HT off (there's a BIOS setting to enable/disable HT).

Although I still stick with my OTHER comments in the last post about feeling faster and such. -JB

Bingo! It is often the case that interactive programs are written to feel faster to improve the user experience - even if by a stop watch they're a bit slower.

Humans perceive "instantaneous" to be somewhere in the 30msec to 50msec range - anything less is instantaneous.

A program that has six 50 msec delays as it starts up will be perceived as faster than a program with one 200 msec delay - even though it's 50% slower. (By this I mean that the first program updates the screen a bit, pauses 50 msec, updates, pauses, updates.... The second program pauses 200 msec, then draws the whole window.)

HT can have the same effect. You feel like things are faster and more fluid - but if you pick the right benchmark and use an accurate timer you'll realize that it's a bit slower.

To me, though, it sounds like a win!

 

[~loserman~]

Power Books can be even thinner with Intel

The Powerbook line can become even thinner using Pentium M.
They also will have longer battery life and better performance than they currently have.
Another benifit will be the ability to run Windows Apps without having to use Virtual PC and Windows because of WINE.

WINE has come along way in the last few years and now that Apple is switching to X86 it becomes a possibility for the Mac

 

[ksz]

I don't know if this has been posted yet, I apologize if it has. Anyone doubting the Pentium M's prowess should really check out this article. Pentium M takes on the other top CPUs out there. Athlon FX 55 and Pentium 4 EE CPU's.

http://www.tomshardware.com/cpu/20050525/index.html

In short, the Pentium M whoops all kinds of ass over the P4 and Athlon FX. Imagine a dual core version of this CPU with a memory bus faster than 533Mhz. Wewp!

This surprised me as well. An overclocked Pentium M running at 2.5 GHz beating just about everything out there...and doing it while dissipating only 27 Watts (give or take a few for OCing).

 

[GuyClinch]

Here's a question, will the intel x86 chips for apple use a different socket than normal? I ask this because if one can stick any easily buyable intel chip into a mac then surely that drastically reduces the incentive to buy a new mac when you can forego apple and upgrade it yourself rather easily?

Two comments here:

1) Intel has been changing sockets quite frequently - so I am not so certain that you won't want to buy a new Mac even if they use standard intel sockets.

2) Also don't rule out that the "security" chip will insist that your bios downclock the "new" intel chip to it's old speeds even if it did work. Yes I realize that the New Macs might use EFI but the same idea holds.

For example with Dell you DO NOT have full control over your BIOS (and hence new chip options options) like you do with a custom built system. Most of the overclocking and/or processor switching is done on these kinds of systems NOT OEM systems. In otherwords Apple can simply engineer away this kind of "disincentive" to buy new Macs quite easily.

Pete

 

[GuyClinch]

This surprised me as well. An overclocked Pentium M running at 2.5 GHz beating just about everything out there...and doing it while dissipating only 27 Watts (give or take a few for OCing).

The Pentium M is a heck of a chip and with some tweaking it's as fast as anything out there. But in all honesty it's not really faster then the AMD64 series which has enough headroom to compete.

http://www.pcper.com/article.php?aid=133 FYI information the FX is 2.6 GHZ which is the "normal" overclock people are using on the PM on the desktop. The thing is though it's trivial to overclock one of the AMD64 series line to that speed as well.

Pete

 

[FFTT]

If you want to know why Apple didn't choose AMD, you might want to find out who is running their hardware division now and who was runing Motorola's
hardware division back when they hit the 500MHz wall and had supply problems with Motorola.

 

[blitzkrieg79]

The Pentium M is a heck of a chip and with some tweaking it's as fast as anything out there. But in all honesty it's not really faster then the AMD64 series which has enough headroom to compete.

http://www.pcper.com/article.php?aid=133 FYI information the FX is 2.6 GHZ which is the "normal" overclock people are using on the PM on the desktop. The thing is though it's trivial to overclock one of the AMD64 series line to that speed as well.

Pete

Don't forget that pentium M gives about the same performance as Athlon FX clock for clock yet its not as power hungry as Athlon FX, I am more than sure that a desktoptized version of Pentium M will dominate Athlon FX without any doubt...
Anyway, the thing about AMD is that they don't have any watt efficient mobile CPU on horizon and based on a fact that lately notebook sales surpassed desktop sales AMD might actually be in slight trouble...
Anyway, I don't have a good experience with AMD based desktops, mostly flawed chipset designs...

 

[Zaty]

I'm a bit surprised many people seem to believe Apple will stick a Pentium M into iBooks and Mac minis. I'm telling you the won't. iBooks and minis will get a Celeron M. Only PBs will get a Pentium M. This way they can establish a clearer performance advantage for the PowerBooks. The Pentium M is also(probably) too expensive for a mini.

 

[pubwvj]

The Powerbook line can become even thinner using Pentium M. They also will have longer battery life and better performance than they currently have.

Good news. I would not buy another desktop machine. I want something that can travel with me for all sorts of reasons. Longer battery life is _the_ most important improvement I want to see in a new PowerBook. Longer life as in more hours of runtime between charges and longer lifetime for the battery (years and charges it lasts).

 

[pubwvj]

Do most people not touch-type? I assumed more would than not. Wouldn't it take longer to get anything done typing with two fingers?

It surprises me how many people don't touch type. It is so easy to learn and so fast once you know how. You can also do it in the dark. I taught all of our kids to touch type. Fifteen minutes a day of practice and before you know it you're whizzing along faster than you could even write.

 

[GuyClinch]

Don't forget that pentium M gives about the same performance as Athlon FX clock for clock yet its not as power hungry as Athlon FX, I am more than sure that a desktoptized version of Pentium M will dominate Athlon FX without any doubt...

Why do you expect the P-M to "dominate"? We agree they have the same performance clock for clock about in many tasks. Though the AMD64 will naturally hammer the P-M in some tasks that would take advantage of it's "64-bitness"

Anyway, the thing about AMD is that they don't have any watt efficient mobile CPU on horizon and based on a fact that lately notebook sales surpassed desktop sales AMD might actually be in slight trouble...
Anyway, I don't have a good experience with AMD based desktops, mostly flawed chipset designs...

AMD doesn't make the chipsets. I have NFORCE3 - it's rock solid. And I have an intel 865BGF - that's also very reliable. Now I do think that Intel is absolutely the right choice. But intel has several years now to match AMD on the desktop and they have been unable to do it. The current X2 AMD processors are a superior implementation of dual cores - and much better then the NetBurst based Pentium Ds.

I do think your selling AMD a bit short. They do have a low power processor - the Turion. And in game performance the M rocks.

http://www.gamepc.com/labs/view_content.asp?id=turion64&page=9
http://www.gamepc.com/labs/view_content.asp?id=turion64&page=10

The Turion64’s low power consumption and heat production attributes finally give AMD a viable competitor to Intel’s Pentium-M processor, which has dominated the mobile arena with its efficient architecture. Battery life between Turion64 and Pentium-M systems should be very similar, assuming both mobile systems are equipped with similar hardware and batteries. While the Pentium-M delivers better performance per clock over the Turion64, the Turion64 does have some fairly significant features which should not be overlooked. The Turion64 has an integrated memory controller, for smoother application performance, in addition to DDR-400 memory support (opposed to DDR-333). The Turion64 also has the ability to execute 64-bit code, whereas the Pentium-M does not. One could pick up a Turion64 for a low-power, ultra-portable Windows XP X64 Edition 64-bit mobile platform, whereas the Pentium-M is stuck in 32-bit land at this time, and will likely continue to stay here until mid-2006 at the earliest.

But yes the Pentium M is a sweet processor. I have been thinking about building a desktop model to replace my current Northwood P4 (2.6). I would feel better though if I could use a standard cooler instead of those wimpy little ones the motherboards come with. I realize the M runs really cool but I dont like buzzy little fans.

Overall though I have been accused of being a windows troll - I love this deal for Apple. The G5 wasn't as fast as Apple made it out to be - I don't believe it's as fast at the P-M or AMD64 in most normal computing tasks (things other then photoshop filters). And now Apple doesn't have to WORRY about falling significantly behind.

That's really all that matters - you want your "lines" to be competitive with what is out there. Apple realized they weren't going to be on the winning end of any CPU war so they just went with the safe bet. Throw in the fact that Mac owners will now likely be able to play Windows games and other "as needed" windows only software (and there is ton of that out there) it's something that Apple should have done back in 1991.

Pete

 

[iWillard]

I think most do, anyone who has spent a fair amount of time at the keyboard two finger typing will suddenly realise that they no longer need to look and can use all their fingers anyway (like I did!).

Meh... I two finger type faster than most people can touch type... It was that darn Apple IIe that I had as a young young boy.... No one taught me the proper way to type, so I just learned it on my own... My middle school keyboarding teacher looked at me typing as said "Welp, you can type fast already, I'm not gonna make you change..."

I'd say two of the major reasons for people choosing a PowerBook over an iBook are Extra desktop (PB) vs. Mirroring (iB) and Bells and Whistles (scrolling trackpad/backlit KB/BT standard) vs. standard.... Tho most people don't use half the bells and whistles anyway, but they like to know they're there...

 

[MontyZ]

It surprises me how many people don't touch type. It is so easy to learn and so fast once you know how. You can also do it in the dark. I taught all of our kids to touch type. Fifteen minutes a day of practice and before you know it you're whizzing along faster than you could even write.

I was forced to take typing classes since the 8th grade, and I am now so glad I did. I don't know how I'd do my job if I coudln't touch type. Not having to look down at the keyboard while typing also means you can catch typos as the appear on screen immediately.

Aren't kids required to learn typing in school these days? I would think this should be required as much as learning how to read and write in today's computerized world.

 

[~Shard~]

I was forced to take typing classes since the 8th grade, and I am now so glad I did. I don't know how I'd do my job if I coudln't touch type. Not having to look down at the keyboard while typing also means you can catch typos as the appear on screen immediately.

Aren't kids required to learn typing in school these days? I would think this should be required as much as learning how to read and write in today's computerized world.

Along those lines, you might find this thread of interest.

 

[ClimbingTheLog]

I totally disagree. I think it's much easier to scratch a "polished" plastic surface like the iBook's than to dent an anodized aluminum one.

Yeah, I guess it depends how you handle the machines. Still, it's pretty easy to polish out a scratch in polycarbonate - I'd hate to try pulling an Al dent.

I guess some people would rather buy a dented car than a scratched one but most go the other way (dents betray impacts).

 

[mccoma]

I'm a bit surprised many people seem to believe Apple will stick a Pentium M into iBooks and Mac minis. I'm telling you the won't. iBooks and minis will get a Celeron M. Only PBs will get a Pentium M. This way they can establish a clearer performance advantage for the PowerBooks. The Pentium M is also(probably) too expensive for a mini.

That's a good possibility - the other is (given the timeframe of Intel releases) that the iBook will get the single core M and the Powerbook will get the dual core version.

 

[MacSA]

At the rate Apple are updating hardware, the next iBook WILL have the Intel chip.

 

[tdewey]

That's a good possibility - the other is (given the timeframe of Intel releases) that the iBook will get the single core M and the Powerbook will get the dual core version.

You don't think they'd just use a slower version of the Yonah for the iBook and mini?

http://www.digitimes.com/mobos/a20050609A2006.html

Only a $32.00 difference between the (low-end) 1.66 dual-core Yonah and the single-core version.

 

[kretzy]

What do people think the Intel chips will be named?....i mean will Apple keep their traditional Intel names - P-M, P4, Yonah etc. - or continue with the GX naming system?

For example, if they put a P-M in the new powerbook, will they name it a G5/G6 processor, or just P-M

 

[AidenShaw]

it won't matter

What do people think the Intel chips will be named?....i mean will Apple keep their traditional Intel names - P-M, P4, Yonah etc. - or continue with the GX naming system?

For example, if they put a P-M in the new powerbook, will they name it a G5/G6 processor, or just P-M

It won't matter, everyone will use Intel's names.

Think about how many people use "Velocity Engine" instead of "AltiVec"....