Merom apparently works without a hitch in current Intel Mac's

[(L)]

What most of us would enjoy thoroughly

Upgrade potential.

Laptops can and should have the ability to upgrade processors and VRAM just like you can with RAM. Yes, that means considerably less hardware sales of people buying faster laptops...but there is a demand for it and they can buy through Apple if they price it right, plus a lot of people like me would just direct that sort of upgrade money to a new one anyway, so Apple doesn't need to be all scared about losing money. You don't lose money by giving your customers upgradability. Plus, if Macs do get to be far more popular, upgradability adds to reliability and dependability, since you can rely on the Mac to be better longer, and might replace broken parts like old HDD's that just spun their lifespans.

Having said that, when I bought my 12" PB, I had intended to upgrade the RAM eventually. I don't think I will - it just doesn't really need it, and I'd rather spend the money otherwise or even count that for a new laptop. By the time the average consumer is considering that their machines aren't quick enough, they'll probably find the latest models far more attractive than a cheap upgrade, if they can afford it. (If they can't, they aint buying a new Mac anyways.) As for storage...well, that will probably go a lot to external media anyway, which in some ways are more reliable and compatible.

 

[dudemac]

I was reading that the virtualization software (from parallel i think the company was) does not work on the Mac Mini as well as it does on the MPB and iMac. Will this processor upgrade help that at all or is there another issue at hand here?

All I have to say is that it may be faster on an iMac with a faster processor. But my core duo 1.66 mini runs the parallels software just fine. In fact I am using it now to post this. Currently running linspire 5.0. I would also have to note that I am running with matched 2gigs of ram which helps everything out. Also just as a speed comparison I have a old AMD 700 Duron running the LInspire and win xp pro, the Mini is faster overall with the parallels software.

 

[(L)]

There was a bigger boost from the G4 to the Yonah. But it is nice to know I can squeeze another 20% if I had a Mini. Too bad it wasn't a 4x improvement again.

Now more than ever, 20% of already-rather-fast or fast-enough isn't so attractive!

 

[MrCrowbar]

"Rosetta:

Yonah 38K @ 2.16Ghz
Merom 42K @ 2Ghz"

not too much of a speed bump, but considering the Merom has a slower clock, that's very good. Merom works cooler, right? I guess the ext MBP will have that chip since anything else would be out of date

And if Fugger can get a Merom, Apple should be able to do this too. Maybe we'll see the 12" MBP with that CPU in near futer since battery life is improved and Apple can make it as thin and lightweight as planned (batteries are still big and heavy).

 

[dr_lha]

That's strange. Because the Intel roadmap specifies the Core Solo as not having VT at all! However, I read on the same thread linked in this article that the "Solo" CPU's sold by Apple in the Mac Mini are in fact Duo's with the second core disabled. However it turned out not to be possible to enable the second core.

You experience would prove that theory. Very interesting!

If you look at the CPU features of the Core Solo (sysctl -A), you'll see that VMX, which is the code for Vanderpool, VT-x, etc. is present on the Core Solo.

I think you are right, Core Solos are just Core Duos with a disabled core. Most likely because of a manufacturing defect I'd guess.

 

[Photorun]

It should be noted that you could upgrade the Wallstreet PowerBook, if you were daring, with entire motherboard kits complete with processor. If some 3rd party comes along and offers that it's the portable's only hope.

 

[Ranma13]

I think you are right, Core Solos are just Core Duos with a disabled core. Most likely because of a manufacturing defect I'd guess.

Actually, not a manufacturing defect but rather the way processors are created in general. See, processors are produced on wafers, all with the goal of achieving the maximum speed possible. So say one wafer can produce 100 processors (theoretical number). Maybe only 5 of those can run at 3 GHz reliably (the max speed), and 20 of those at 2.5 GHz reliably, and the rest at 2.0 reliably. That's the reason why overclocking is very easy to do, because sometimes these processors are sold as cheaper processors because they can only run at a certain speed reliably as tested in factory, but with a little voltage boost and some modifications, it can run at higher speeds no problem.

The Core Solos are nothing more than Core Duos that didn't make the cut for a Core Duo. Hence, the second core is disabled and the chip is sold as a Core Solo. Physically, the chip is exactly the same as a Core Duo, but its real-life performance can only perform as a Core Solo reliably.

 

[bah-bah'd]

Yeah! Customizability seems to be the reason why a lot hard core computer users such as gamers go for machines that run windows... because they can choose all of the important specs.

I hope this is a slippery slope!

that & lots of people don't have all the money at once to put towards their dream system...

 

[firestarter]

Actually, not a manufacturing defect but rather the way processors are created in general. See, processors are produced on wafers, all with the goal of achieving the maximum speed possible. So say one wafer can produce 100 processors (theoretical number). Maybe only 5 of those can run at 3 GHz reliably (the max speed), and 20 of those at 2.5 GHz reliably, and the rest at 2.0 reliably. That's the reason why overclocking is very easy to do, because sometimes these processors are sold as cheaper processors because they can only run at a certain speed reliably as tested in factory, but with a little voltage boost and some modifications, it can run at higher speeds no problem.

The Core Solos are nothing more than Core Duos that didn't make the cut for a Core Duo. Hence, the second core is disabled and the chip is sold as a Core Solo. Physically, the chip is exactly the same as a Core Duo, but its real-life performance can only perform as a Core Solo reliably.

Hmm. The speed of a chip is most likely to affect the whole wafer, not just a single core of an individual processor. More likely that a core is turned off through a contamination defect (which do happen). The ability to resurrect a chip with one core out of action makes for much higher yields.

 

[timswim78]

this swapability could be a foreshawdow of some great new flexibility in apple hardware.

It also means that we'll finally be able to upgrade the CPU's on Macs without spending obscene amounts of money.

 

[smokeyboi]

The way you talked in the other Macbook thread its as if everything will be 64-bit by year's end and Yonah 32-bit systems will be obsolete, but now you are saying Apple doesn't even have a roadmap for 64-bit and we don't really know when we will see them implement 64-bit, not to mention all software developers need to get on board...

AidenShaw seems to be suffering from multiple personality disorder. He's always contradicting himself.

 

[wPod]

dude! i am going to totally upgrade my mini (core solo) when this comes out!!! i got the core solo figuring id upgrade to the duo once i wanted more performance. but man, upgrading to 64 bit with 25% ish increase in performance would be awesome!!! (just wish i could upgrade the video card too! ha!)

 

[AidenShaw]

I'm assuming true 64-bit in 10.5

The way you talked in the other Macbook thread its as if everything will be 64-bit by year's end and Yonah 32-bit systems will be obsolete, but now you are saying Apple doesn't even have a roadmap for 64-bit and we don't really know when we will see them implement 64-bit, not to mention all software developers need to get on board...

AidenShaw seems to be suffering from multiple personality disorder. He's always contradicting himself.

I'm assuming that Apple will announce at WWDC that 10.5 will be true 64-bit on Intel (I don't see any other reasonable option - they can't continue to ignore 64-bit, and to port 10.4's lame 64-bit support to the new chip would be a mistake).

In that scenario, when 10.5 ships with true 64-bit support there will be a lot of MacIntels (Merom/Conroe/Woodcrest) able to switch to 64-bit OS.

If WWDC does what it should, then vendors will have some 64-bit applications ready from day 1 of the 64-bit era. (Just as the Intel DTKs released last June gave developers a head start on fat-binary apps.)

Each of those steps (64-bit hardware, 64-bit OS, 64-bit applications) increases the "obsolete-y-ness" of the Yonah systems.

It's not an instant transition - you won't go to sleep one night with a "state-of-the-art" MacBook and wake up the next morning with an obsolete boat anchor.

It will happen a lot faster than typical for Apple computers, though.

 

[dr_lha]

Actually, not a manufacturing defect but rather the way processors are created in general. See, processors are produced on wafers, all with the goal of achieving the maximum speed possible. So say one wafer can produce 100 processors (theoretical number). Maybe only 5 of those can run at 3 GHz reliably (the max speed), and 20 of those at 2.5 GHz reliably, and the rest at 2.0 reliably. That's the reason why overclocking is very easy to do, because sometimes these processors are sold as cheaper processors because they can only run at a certain speed reliably as tested in factory, but with a little voltage boost and some modifications, it can run at higher speeds no problem.

The Core Solos are nothing more than Core Duos that didn't make the cut for a Core Duo. Hence, the second core is disabled and the chip is sold as a Core Solo. Physically, the chip is exactly the same as a Core Duo, but its real-life performance can only perform as a Core Solo reliably.

Yes, that is what I meant.

 

[dr_lha]

I'm assuming that Apple will announce at WWDC that 10.5 will be true 64-bit on Intel (I don't see any other reasonable option - they can't continue to ignore 64-bit, and to port 10.4's lame 64-bit support to the new chip would be a mistake).

Agreed, they will be forced to produce a 64-bit version of OS X. However I think the reason for this is due to the fact that PPC64 chips allow you to run 64-bit apps under a 32-bit OS (like Tiger does right now for PPC64 chips like the G5), whereas Intel's 64-bit chips do not support this.

I may be wrong, but I remember reading this.

 

[daneoni]

Yes, if you have the nerve to crack open your iMac, you'll be able to replace the processor. The 2GB RAM limit is, to the best of my knowledge, a function of the fact that 1GB chips are the largest you can get at the moment. If a 2GB chip becomes available, which it surely will, you'll be able to upgrade.

I know I'll be doing this with my iMac once the current Yonah feels too slow to run the newest software. But I wont be voiding the warranty by cracking it open just yet.

I could do it mid-warranty and if something happens, i'll simply dump the yonah chip back in and no one wil be the wiser...well theoretically

 

[AidenShaw]

Mostly correct....

Agreed, they will be forced to produce a 64-bit version of OS X. However I think the reason for this is due to the fact that PPC64 chips allow you to run 64-bit apps under a 32-bit OS (like Tiger does right now for PPC64 chips like the G5), whereas Intel's 64-bit chips do not support this.

I may be wrong, but I remember reading this.

This is basically true for normal OS - the conventions for 32-bit mode and 64-bit mode make it difficult for a 32-bit OS to support a 64-bit application - which the 32-bit OSX 10.4 is able to do on the PPC.

It is not impossible, though - VMware is able to run 64-bit guests from 32-bit host operating systems. That's a very special case, of course, since the guest is strictly isolated from the host by the VMM.

IMO it wouldn't make sense to rewrite a 32-bit OS to support individual 64-bit applications - spend the engineering effort to make a true 64-bit OS.

And, don't forget that x64 64-bit is faster than 32-bit, so if you do a full 64-bit OS your OS will be faster.

 

[plinden]

I was reading that the virtualization software (from parallel i think the company was) does not work on the Mac Mini as well as it does on the MPB and iMac. Will this processor upgrade help that at all or is there another issue at hand here?

There appears to have been a bug in the firmware of some of the Mac Minis that disabled the virtualization support and this affected performance. From this thread it seems there is no real pattern to which versions were affected.

No, unfortunately the VT-capability was disabled in the firmware of the mini, even for the dual-core one. So upgrading to a Merom won't enable VT support. You can still run parallels, but with a lower performance.

I found this on the Parallels site somewhere, and on the forum thread about it here. Sorry, I don't have any links..

It's a shame, because I was waiting to see if that was possible before I'd buy a mini. I might get an iMac now. They're better value anyway.

Edit: Sorry, isgoed already mentioned the same thing. DIdn't see that before I posted, sorry.

Nope, read the thread I linked to above. (Well it was disabled in some, but perhaps not intentionally)

 

[840quadra]

And, don't forget that x64 64-bit is faster than 32-bit, so if you do a full 64-bit OS your OS will be faster.

Not to mention the nice benefit of extra memory Bandwidth . Sadly the Mini will not be able to utilize this feature with the 2gb ceiling .

 

[YoNeX]

Actually, not a manufacturing defect but rather the way processors are created in general. See, processors are produced on wafers, all with the goal of achieving the maximum speed possible. So say one wafer can produce 100 processors (theoretical number). Maybe only 5 of those can run at 3 GHz reliably (the max speed), and 20 of those at 2.5 GHz reliably, and the rest at 2.0 reliably. That's the reason why overclocking is very easy to do, because sometimes these processors are sold as cheaper processors because they can only run at a certain speed reliably as tested in factory, but with a little voltage boost and some modifications, it can run at higher speeds no problem.

The Core Solos are nothing more than Core Duos that didn't make the cut for a Core Duo. Hence, the second core is disabled and the chip is sold as a Core Solo. Physically, the chip is exactly the same as a Core Duo, but its real-life performance can only perform as a Core Solo reliably.

So if I understood you right, in theory you could unlock the a processor's other core to give you a big performance boost? We already know overclocking is possible, just unlocking the cores are possibe. If it can be done, someone will do it eventually.

As someone said before, yes replacing the processor might yield a performance gain, but might be bottlenecked by the rest of the components, RAM, HD, and others. So, if you replace the CPU, ideally you should replace other components, so one part doesn't bottleneck your CPU.

 

[Lord Blackadder]

Merom apparently works without a hitch in current Intel Mac's

<does CPU victory dance>

Apple, please keep putting sockets in those Intel Macs!

 

[^squirrel^]

Not to mention the nice benefit of extra memory Bandwidth . Sadly the Mini will not be able to utilize this feature with the 2gb ceiling .

SO let me get this clear....

You can only run 64bit if you have a min of 4gb ram?

 

[shawnce]

If WWDC does what it should, then vendors will have some 64-bit applications ready from day 1 of the 64-bit era. (Just as the Intel DTKs released last June gave developers a head start on fat-binary apps.)

I believe we will see a solid 64b story from Apple at WWDC 2006 (in August, I plan to attend).

The nice things is that the Apple's tool chain (Xcode Tools) is -- partly a result of having to deal with the Intel transition and PPC64 before that -- rather well prepared to allow the creation of PowerPC, IA32, and IA32-64 (aka x86-64) universal libraries with a simple project configuration change and recompile. Apple even could in theory provide a 64b SDK for developers to compile/link against (not run) while they are still whipping up a developer preview of a IA32-64 Mac OS X. If developers are smart they are getting their applications internally 64b clean now (I know we are trying) so it can be as easy as a recompile against a 64b SDK.

I also fully expect Apple to allow existing 32b applications to run on a 64b configured Mac OS X... so folks don't panic that you will have to upgrade your existing Intel native application yet again once 64b comes around.

Providing 64b version of Mac OS X and its frameworks is non-trivial for Apple (especially when you factor in legacy Mac OS that is still reflected in parts of Carbon, etc.)... but I believe they have been working on it since at least 10.3 (of course the Intel transition likely disrupted the timeline).

 

[daneoni]

This i think suggests the 15" MBP will launch merom with 1.83 & 2.0 GHz clock speeds with the option of upgrading to 2.16GHz just as we have now, whilst the 17" MBP will launch with a 2.16GHz base model and 2.33Ghz (which according to rumors has an optional 800MHz FSB) as the high end version and will hopefully have the Dual Layer 8X SuperDrive.

The future iMacs will probably launch with Conroe instead although i think those will be reserved for the PowerMacs (MacPro/ProMac?) and the Xserves will launch on Woodcrests.

So all in all now is a good time to buy the iMac because the only changes i see in the nearest future are the Yonah chip substitutions for Meroms, the bus will remain the same at 667MHz with the only difference being you can upgrade past the 2GB RAM limit (due to 64 bit abillities of the chip) once those 2GB memory sticks start being manufactured....which i think they are because dell offers 4GB (2x 2GB) RAM configurations in their latitude notebook lineup for about $900-1000+ i think.

The graphics cards (ATI will release the RV570/R560 cards...successor to the X1600/X1800 series) will also change but the performance difference would be negligible for an average iMac user like me.

Sigh....the question now becomes buy an iMac now or buy a merom based MacBook Pro

 

[shawnce]

SO let me get this clear....

You can only run 64bit if you have a min of 4gb ram?

No. The amount of physical RAM in a system does not affect the size of virtual addresses that the virtual memory system could provide to an application. The later depends on the capabilities of the VM system, the ABI of the OS (including framework support), and addressing support in the CPU & MMU. The former depends on the number of RAM slots, addressing pins, etc. supported by the mother board chipset.

The amount of physical RAM only affects how much data you can quickly work with (data in RAM is radically faster then having to access it from VM swap file or a file on disk).

I believe the OP was talking about no being able to fully leverage a large address space since you only have the ability to install 2 GiB of RAM in a mini, not that it couldn't support 64b addressing (assuming it had the needed software and hardware support, which current mini do not have -- possibly with a CPU upgrade but...).