Intel's Napa Platform with Yonah

[Randall]

Of course these tools slow the system. While you are correct that these items have an impact of a varying nature, the impact is still there. The effect is most pronounced when tools are utilized that perform "real time scanning", as these tools recheck system components each time a new process is started or a new file is installed.

Amen to that. Real-time virus scanning can have significant impact on performance, yet it is necessary in the Windows world as you all know.

Deep pockets, heavy intellectual resources, and a solid processor roadmap are all good reasons to say, I think, that the "Mactel" decision was a very good one for the future of Apple.

It was an excellent choice IMO as well. Hopefully all the PPC fanbois and nay-sayers will be on board after January too. Just because Apple is no longer using a unique architecture doesn't mean that innovation is dead. In fact I think this move to x86 gives new life to innovation.

 

[AidenShaw]

Of course these tools slow the system. While you are correct that these items have an impact of a varying nature, the impact is still there. The effect is most pronounced when tools are utilized that perform "real time scanning", as these tools recheck system components each time a new process is started or a new file is installed.

Sorry to disturb your myths with actual data, but note the following real-time scanning chart:

http://www.eset.us/compare/

This is pretty much a worst case test - opening and closing 200 different Excel spreadsheet files as quickly as possible.

The slowest case McAfee product slowed each access by 66 msec. Norton and others were less than 20 msec slowdown.

So, if the "most pronounced case" with Norton is 20 msec, I'll stand by my claim that it's "virtually unnoticeable". (Human perception can't discern below about 30-40 msec - anything faster is perceptually instantaneous.)

 

[backdraft]

Oh really?? Care to post some hard numbers to back up that claim?

Sure:

Here's a temperature test for the old dual PowerPC G5 @2.0 (130nm) couldn't find current. Temp are in degrees Celsius... (2 cpu's)

http://domino.research.ibm.com/library/cyberdig.nsf/papers/E6EBD3C859FB49F785256ED8006A3F4A/$File/rc23276.pdf

And Check out the new P4 temps in Celsius (90nm) you'd think they would be considerable less... (1 cpu)

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2026&p=3

And if u want to measure the temperature on your Mac here's an app: http://bresink.de/osx/HardwareMonitor.html

It uses the sensors on your mobo

-backdraft

 

[Sunrunner]

Note the following real-time scanning chart:

http://www.eset.us/compare/

This is pretty much a worst case test - opening and closing 200 different Excel spreadsheet files as quickly as possible.

The slowest case McAfee product slowed each access by 66 msec. Norton and others were less than 20 msec slowdown.

So, if the "most pronounced case" with Norton is 20 msec, I'll stand by my claim that it's "virtually unnoticeable". (Human perception can't discern below about 30-40 msec - anything faster is perceptually instantaneous.)

You missed the three key falacies to this test:

A) Older systems would have more pronounced result
B) Larger files would result in much larger scan times (the average excel file is only like 100k)
C) User time devoted to "helper" program installation/upgrade/configuration must also be computed and averaged over the specified timeframe.

 

[Sunrunner]

All this less watts/heat issue is BS, the X86 architecture is archaic. The PowerPC uses less watts/heat NOW. Stick 2 Intel chips in a tower and see how much heat they produce and how many watts they used. Once PPC goes 65nm you'll see. Face it Apple is TOO cheap to pay IBM to develop the PPC. They make billions and can't spare a couple million come on!
View attachment 36942

Its apparent you dont understand the intricacies of the matter. PPC is NOT a very efficient chipset. Also, Apple was basically forced to switch as a result of IBMs decision to not develop the G5 into a laptop chip. Of course, if you WANT to carry around an 8lb, 2-inch thick G5 powerbook, be my guest. IBM lost interest in Apple's processor needs when they picked up the Xbox 360 contract.

 

[aegisdesign]

Amen to that. Real-time virus scanning can have significant impact on performance, yet it is necessary in the Windows world as you all know.

No it's not .Just keep it away from the internet.

It was an excellent choice IMO as well. Hopefully all the PPC fanbois and nay-sayers will be on board after January too. Just because Apple is no longer using a unique architecture doesn't mean that innovation is dead. In fact I think this move to x86 gives new life to innovation.

Sorry, but I find Yonah as just so-so and I'm disheartened that Apple went with yesterdays 32bit architecture instead of waiting to transition in one step to 64bit intel/amd.

The real deal is Merom/Conroe. Until then the G5 rules (and Opterons too or pretty much any AMD desktop/server chip) for real work. If they do a version of OSX for the OQO now though, I'm there. That's all I want in a 'laptop'.

 

[Randall]

Sorry to disturb your myths with actual data, but note the following real-time scanning chart:

http://www.eset.us/compare/

This is pretty much a worst case test - opening and closing 200 different Excel spreadsheet files as quickly as possible.

The slowest case McAfee product slowed each access by 66 msec. Norton and others were less than 20 msec slowdown.

So, if the "most pronounced case" with Norton is 20 msec, I'll stand by my claim that it's "virtually unnoticeable". (Human perception can't discern below about 30-40 msec - anything faster is perceptually instantaneous.)

You forgot about all the spyware running in the background stealing all the CPU time and memory. Sad to say, but spyware is a serious problem to the average windows user. Furthermore, that access time is in Seconds, and I see a HUGE performace hit in that chart with McAfee.

 

[Randall]

No it's not .Just keep it away from the internet.




Sorry, but I find Yonah as just so-so and I'm disheartened that Apple went with yesterdays 32bit architecture instead of waiting to transition in one step to 64bit intel/amd.

The real deal is Merom/Conroe. Until then the G5 rules (and Opterons too or pretty much any AMD desktop/server chip) for real work. If they do a version of OSX for the OQO now though, I'm there. That's all I want in a 'laptop'.

Apple couldn't wait any longer to transition. Another year of G4's could have sunk them. Yes the G4 sucks that much. As for the Intel insides, to quote National Lampoon's Vacation... "You hate it now, but wait 'til you drive it!"

 

[backdraft]

Its apparent you dont understand the intricacies of the matter. PPC is NOT a very efficient chipset. Also, Apple was basically forced to switch as a result of IBMs decision to not develop the G5 into a laptop chip. Of course, if you WANT to carry around an 8lb, 2-inch thick G5 powerbook, be my guest. IBM lost interest in Apple's processor needs when they picked up the Xbox 360 contract.

Pit an X86 with equivalent ghz against PPC and you'll see how the PPC is more efficient and outperforms X86. Not to mention the top ranking supercomputers use PPC chips. Research/Development and Military Applications require PPC not X86.

"IBMs decision to not develop the G5 into a laptop chip"

IBM decided not to develop the laptop version of the G5 because Apple didn't want to pay. They asked Apple for $$$ so they can give them the product, what's wrong with that? IBM contracts chip designs its how they make $$$. Besides IBM wanted a couple million, Apple would recover the $$$ in no time. Spend a millions to make billions.

What about 64bit support? Altivec vs SSE2? DRM garbage in Intel chips? Intel's bigger die size? X86 exploits that are non-existent on PPC?

 

[Randall]

Pit an X86 with equivalent ghz against PPC and you'll see how the PPC is more efficient and outperforms X86. Not to mention the top ranking supercomputers use PPC chips. Research/Development and Military Applications require PPC not X86.

"IBMs decision to not develop the G5 into a laptop chip"

IBM can decided not to develop the laptop version of the G5 because Apple didn't want to pay. They asked Apple for $$$ so they can give them the product, what's wrong with that? IDM contracts chip designs its how they make $$$. Besides IBM wanted a couple million, Apple would recover the $$$ in no time. Spend a millions to make billions.

What about 64bit support? Altivec vs SSE2? DRM garbage in Intel chips? Intel's bigger die size? X86 exploits that are non-existent on PPC?

People please. I am so sick on GHz being compaired as an actual performance test. IT MEANS NOTHING. The bottom line is that the G4 rocked in it's hayday, but that was years ago. Now the G4 sucks a$$. The move to x86 will be the best thing that's ever happened to Apple. This opens up unlimitless 3rd party hardware options. I could care less about my laptop harware being the same architecture type as the Military uses. They probably want to switch to x86 too. And what about Altivec vs SSE3... That's what I thought.

P.S. 64 bit support for the laptop will be rolled out in Intel's next line of processors, due out at the end of 2006.

 

[backdraft]

People please. I am so sick on GHz being compaired as an actual performance test. IT MEANS NOTHING. The bottom line is that the G4 rocked in it's hayday, but that was years ago. Now the G4 sucks a$$. The move to x86 will be the best thing that's ever happened to Apple. This opens up unlimitless 3rd party hardware options.

Who said anything about the G4? G5 is the way to go.

"This opens up unlimitless 3rd party hardware options"

Give me the best hardware I don't need the rest, screw 3rd parties.

 

[SiliconAddict]

It's interesting to see the myth about how malware tools slow the system down turning into "accepted fact" on these boards....

Sure, the occasional full scan puts a load on a system, but the real-time stuff that's always on is virtually unnoticeable.

Depends on the software. Norton's consumer products are MASSIVE resource hogs. Their corp offerings (The only Symantec products I run.) are MUCH more refined and hardly touch the CPU or memory footprint.

 

[Randall]

Who said anything about the G4? G5 is the way to go.

"This opens up unlimitless 3rd party hardware options"

Give me the best hardware I don't need the rest, screw 3rd parties.

Yeah you're right, who needs top of the line ATi graphics cards for x86. Use of COTS video cards is a major perk with the move to intel.


I was talking about the G4. We're concerned with Apple's laptop line. The desktop line with it's dual G5's are perfectly fine for the time being.

 

[SiliconAddict]

Who said anything about the G4? G5 is the way to go.

"This opens up unlimitless 3rd party hardware options"

Give me the best hardware I don't need the rest, screw 3rd parties.

Unless you've figured out how to shoehorn a G5 into a PowerBook and still get 5 hours of battery life the G4 is very much is relevant.

 

[egsaxy]

um, yes and no.

Before people get carried away with dual core 2Ghz iBooks, a dose of reality.

Here's the Yonah pricing...

http://www.theinquirer.net/?article=26062

The top end Yonah is almost ten times the price of a G4. A Yonah as fast as the current G4 is 2-3 times as expensive.

1) Everyone knows the intel chips are more expensive to a small degree.
2) These are dual core chips so the 1.6 and 1.8 ghz chips at (241 and 295) respectively are for 2 chips
3) With simple logical deduction (but not 100 percent acuracy) these individual chips really cost 120.5 and 147.5 respectively
4) compared to an outdated G4 chip at 75 dollars (from postings elsewhere in this forum) the faster intel chip is slightly under 2x a chip and the same speed chip is only 1.5x a chip
5) We'd be getting 2 chips inside instead of one. even at the same speed this will be a powerful improvement.

at same speed these chips aren't too much more expensive. asumming apple and our consumers understand they've crammed two chips into a machine and apple absorbs some (not all) of the cost we'll have a good situation.

my 2 cents.

 

[backdraft]

People please. I am so sick on GHz being compaired as an actual performance test. IT MEANS NOTHING. The bottom line is that the G4 rocked in it's hayday, but that was years ago. Now the G4 sucks a$$. The move to x86 will be the best thing that's ever happened to Apple. This opens up unlimitless 3rd party hardware options. I could care less about my laptop harware being the same architecture type as the Military uses. They probably want to switch to x86 too. And what about Altivec vs SSE3... That's what I thought.

P.S. 64 bit support for the laptop will be rolled out in Intel's next line of processors, due out at the end of 2006.

Altivec vs SSE3 u know what I mean... Altivec blows SSE[insert number here] away...

"I could care less about my laptop harware being the same architecture type as the Military uses."

You just want "good enough" I don't, and neither do most Mac users, else they would be using X86 and Windows.

 

[Randall]

1) Everyone knows the intel chips are more expensive to a small degree.
2) These are dual core chips so the 1.6 and 1.8 ghz chips at (241 and 295) respectively are for 2 chips
3) With simple logical deduction (but not 100 percent acuracy) these individual chips really cost 120.5 and 147.5 respectively
4) compared to an outdated G4 chip at 75 dollars (from postings elsewhere in this forum) the faster intel chip is slightly under 2x a chip and the same speed chip is only 1.5x a chip
5) We'd be getting 2 chips inside instead of one. even at the same speed this will be a powerful improvement.

at same speed these chips aren't too much more expensive. asumming apple and our consumers understand they've crammed two chips into a machine and apple absorbs some (not all) of the cost we'll have a good situation.

my 2 cents.

Let's not forget the fact that the FSB on Yonah is 4x faster then the G4. That will make a HUGE difference.

Altivec vs SSE3 u know what I mean... Altivec blows SSE[insert number here] away...

"I could care less about my laptop harware being the same architecture type as the Military uses."

You just want "good enough" I don't, and neither do most Mac users, else they would be using X86 and Windows.

LMAO of course I want top of the line, but you won't be getting it with G4. Don't worry buddy, you'll be using x86 soon enough...

 

[backdraft]

Yeah you're right, who needs top of the line ATi graphics cards for x86. Use of COTS video cards is a major perk with the move to intel.


I was talking about the G4. We're concerned with Apple's laptop line. The desktop line with it's dual G5's are perfectly fine for the time being.

Go with NVIDIA... Plus I'd rather see the CELL CPU as a coprocessor to handle video.

Unless you've figured out how to shoehorn a G5 into a PowerBook and still get 5 hours of battery life the G4 is very much is relevant.

IBM can Apple has to cough up $$$ though. Oh, switch to Intel means bye bye Hypertransport. mobile dual core G5 + Hypertransport would smoke anything...

 

[Randall]

Go with NVIDIA... Plus I'd rather see the CELL CPU as a coprocessor to handle video.

I don't think that the CELL CPU will be able to compare with high end GPUs

IBM can Apple has to cough up $$$ though. Oh, switch to Intel means bye bye Hypertransport.

If that's the case, then Apple was planning the switch to x86 this whole time.

 

[SiliconAddict]

The Pentium M only really smokes a G4 in integer based benchmarks and anything very memory intensive that has to go through the CPU. That 2MB cache comes in handy. I've a 1.7Ghz Pentium M laptop and it's a lot slower at encoding in iTunes and transcoding video than a G4 and much slower than a G5. AltiVec makes a difference there and in some Photoshop filters at least until you get bandwidth bound instead of CPU bound.

It'd be interesting to see how the new Powerbooks do versus a Pentium M as they now run the Intrepid2 controller at 333Mhz instead of the previous 167Mhz direct FSB 1:1 lock. In theory since a lot of the time the CPU isn't involved with moving data about, the 333Mhz Intrepid2 has twice the memory access bandwidth available so is able for instance to DMA data off disk to memory twice as fast as the old controller.

Where the Pentium-M Windows laptops score well though is in graphics performance and in particular OpenGL. And that says more about Apple's slow OpenGL implementation than the processor. I'd be surprised if that changes with a MacIntel so expect us to still lose miserably in Cinebench. Dissing Windows performance is going to come back and haunt us if given the same hardware Windows is faster.

Of course the 1.67Ghz G4 v 1.7Ghz Pentium-M comparison is academic as there's a dual core 2.16Ghz Pentium-M in January and more to follow and no faster G4s or a laptop G5 on the horizon.

What about a 1.67 vs. a 2.26Ghz Pentium M? Or top of the line vs. top of the line? Its already been stated that the M's FP calculations ability is its weakest point. There is no contention on this. It doesn't suck but its not in the same class as the G5 or even the G4 AFAIK. (Could be wrong on that last point.)
In regards to overall benchmarks barefeats.com has shown (only once mind you.) that 2 years ago the G4 was trailing behind the Pentium M. Now it doesn't take a rocket scientist to look at further benchmarks of newer PowerBooks vs. older PowerBooks to see minimal progression then look at 2 year old M's vs. the latest and greatest seen the progress the've made both in the CPU and updated chipset and know the gap hasn't narrowed.

Everyone keeps bringing up the Intrepid2 controller. Again go to barefeats.com look at the benchmarks. (they have new PowerBook benchmarks from October.) There is virtually no difference. A memory controller update isn't going to make that big of a difference when you are hitting the CPU with, as the example you used, such things as iTunes encoding.

 

[backdraft]

Let's not forget the fact that the FSB on Yonah is 4x faster then the G4. That will make a HUGE difference.

LMAO of course I want top of the line, but you won't be getting it with G4. Don't worry buddy, you'll be using x86 soon enough...

yeah thats why I want a G5 in the powerbook.

 

[Randall]

yeah thats why I want a G5 in the powerbook.

They have that, but it's not called a Powerbook, it's called a Space Heater.

 

[SiliconAddict]

IBM can Apple has to cough up $$$ though. Oh, switch to Intel means bye bye Hypertransport. mobile dual core G5 + Hypertransport would smoke anything...

Ahh the mythical mobile dual core G5. That must be the one that was to be released the month after they released the single core mobile G5. Do you have any idea how much R&D for a single customer costs? I don't. But it has to be a [bleep] load. More then Apple could afford I'm willing to bet. It's been documented that IBM was not all that interested in the mobile G5 but was more interested in gaming consoles. A market that runs rings around Mac sales. https://www.macrumors.com/pages/2005/12/20051207073320.shtml

Face facts. The glorious IBM stopped being relevant after they couldn't deliver a 3Ghz G5. Not that I'm in the camp of people who were/are bitching and moaning about Jobs lying to them. But when you look at the advances from year to year. IBM wasn't exactly progressing at a blindingly fast rate.


PS- Hypertransport == SSE2\SSE3 (Maybe not as fast but its in the ballpark.)

 

[AidenShaw]

You forgot about all the spyware running in the background stealing all the CPU time and memory.

Hyperbole....

Grab the free antispyware tool from MS (http://www.microsoft.com/downloads/...2-6a57-4c57-a8bd-dbf62eda9671&displaylang=en), and spyware be gone.

Furthermore, that access time is in Seconds, and I see a HUGE performace hit in that chart with McAfee.

If you would read the link, or even what I wrote - you'll see that this test did open/close as fast as possible on 200 individual files.

Even for McAfee, that's 67 msec per file....

 

[shawnce]

From osx86project.org, posted 11/21/2005:
If this is true (no reason to doubt it's not, considering the source is not a rumor site), then all the worries of AltiVec-enhanced applications "sucking" under Rosetta is no longer an issue. Even under emulation, the increased bus speed and processing power of a dual-core Yonah over a single-core G4 should provide more than enough horsepower for older PPC applications during the transition period.

Umm based on your beliefs of the magical power of Yonah? Seriously folks Rosetta _is_ going to _suck_ for any computationally intense application.

-Shawn