Mac Pro to Get 6-Core Xeon Gulftown Processor in 2010?

[Eidorian]

I remember them. I remember it well. I remember despite the huge costs of co-developing the PowerPC chip and actually developing the motherboard and other components, the top Macs were cheaper. I remember Macs were refreshed every 6-9 months, not every 12-16. I remember the promise of CHEAPER Macs as part of the transition to the Intel "mass market". I remember the current PowerMac/MacPro enclosure design is about 7 years old. I remember a time when Macs offered more options than PCs, not far fewer. I remember...

For all that extra money you get a dated video card, no eSATA, no HDMI, no Blu-Ray,...What a joke.

Core 2 and the 9400M G have pretty much flattened the majority of Apple's line. It feels a lot like the late PowerPC era where a clock speed multiplier and plastic vs. metal was it.

Arrandale and Clarkdale are going to make things interesting on the processor side but there's no wow in the products right now.

 

[MorphingDragon]

Core 2 and the 9400M G have pretty much flattened the majority of Apple's line. It feels a lot like the late PowerPC era where a clock speed multiplier and plastic vs. metal was it.

TBH, a lot of the X86 world is a bit boring ATM. I was kinda hoping AMD would be allowed to go Fabless.

 

[*LTD*]

Not to mention we'd love to be able to easily upgrade our computer, have multiple internal hard drives, etc.

I hate imacs, and I'll never buy one.

www.winsupersite.com.

 

[DanielSw]

Having spoke with a friend who is a chief engineer at Apple. He told me there are no plans for any new mac pros, so these rumors are INVALID and don't make any sense. 2010 will not bring any new mac pros, sorry.. but I know this from inside information.

A "chief engineer" at Apple wouldn't be telling you ANYTHING about company plans. Idiot.

 

[guzhogi]

Hopefully it will support more than the pitiful 32 GB of memory like the current model. Life begins at 128 GB, so I sure hope the new model supports that. With 12 cores / 24 threads, the machine will be a complete joke if the memory ceiling is still the same 32 GB.

To get at least 128 GB, you'd need 16 memory slots & fill them with 8 GB sticks. I'll admit I don't know how many computer makers sell computers w/ 16 RAM slots, much less more than that. Haven't heard of any 16 GB sticks either. But then again, I don't watch every electronics manufacturer constantly.

 

[Rocketman]

For all that extra money you get a dated video card, no eSATA, no HDMI, no Blu-Ray,...What a joke.

If those ports are important to you, why not get a cheap PC with them and network it to the Mac? Being here I presume there are Mac features and benefits you need as well.

 

[Bubba Satori]

You'll be disappointed when you go to buy an HP, Dell, or Lenovo workstation and it costs as much or more than the Mac Pro. But have fun with that!

They can be had with dozens of better pro video cards than the GT130 or 4870 and other unlimited hardware choices that The Mac Pro can't.

But why go to HP, Dell or Lenovo? Go to Newegg and configure a workstation for 50% less with any possible hardware configuration you'd like.

Nice try.

 

[NSMonkey]

Just found this on Arstechnica...


I'm sure there are many benchmarks that show that HT is useful for some apps, and I do turn it on for my production servers and render-farm type systems - but I'll keep HT disabled on my desktop machines unless I'm going to spend the weekend stealing DVD content with Handbrake.

You cherry-picked the single benchmark that showed a performance degradation with HT. Every other benchmark in the article shows HT performing better than non-HT:

As another example, this benchmark shows Snow Leopard achieving up to a 30% performance increase on the i7 vs the i5. Since the two processors are essentially identical except for a 5% speed difference and the presence of HT, that leaves HT responsible for up to a ~25% improvement.

Even the article Eidorian linked to (Is Hyper-Threading worth it?), which shows no significant gains with HT vs without HT, concludes that "enabling Hyper-Threading is worthwhile".

 

[NSMonkey]

I didn't realize that you could buy an Imac without paying for a monitor that you might not need or want....

You must have misunderstood the comment. guzhogi was merely pointing out that iMacs support multiple monitors.

 

[afrowq]

www.winsupersite.com.

I'm not getting your point here?

 

[Maserati7200]

Having spoke with a friend who is a chief engineer at Apple. He told me there are no plans for any new mac pros, so these rumors are INVALID and don't make any sense. 2010 will not bring any new mac pros, sorry.. but I know this from inside information.

Sure Nadav! *pats you on the head* Run along now!

 

[guzhogi]

You must have misunderstood the comment. guzhogi was merely pointing out that iMacs support multiple monitors.

Yup, that's what I meant.

 

[Terminal.app]

Arrandale and Clarkdale are going to make things interesting on the processor side but there's no wow in the products right now.

We often forget that POWER and SPARC are still superior architectures to x86/x64.

Not to mention, they can very often destroy Intel/AMD on the clock speed front. 6 GHz native speed, anyone? I'd like to see Intel or AMD do that (especially AMD) without setting fire to everything within a 10-mile radius.

 

[MorphingDragon]

We often forget that POWER and SPARC are still superior architectures to x86/x64.

Not to mention, they can very often destroy Intel/AMD on the clock speed front. 6 GHz native speed, anyone? I'd like to see Intel or AMD do that (especially AMD) without setting fire to everything within a 10-mile radius.

Nah, CBE > X68. The Cell has room to improve massively if IBM would just push their own tech once in a while. Isn't Fujitsu making a 32-Core SPARC Proc?

---

Theres more to a processor than just clock speed, if IBM worked on improving the SPE and changing from their silly little SDK to something like OpenCL I could see massive potential.

 

[Terminal.app]

Cell = POWER. I think.

 

[cmaier]

We often forget that POWER and SPARC are still superior architectures to x86/x64.

Not to mention, they can very often destroy Intel/AMD on the clock speed front. 6 GHz native speed, anyone? I'd like to see Intel or AMD do that (especially AMD) without setting fire to everything within a 10-mile radius.

Nonsense. I've designed sparcs at sun, powerpcs at exponential, and x86 at AMD. In the real world clock speed doesn't win, and x86 is the best.

 

[Bubba Satori]

Nah, CBE > X68. The Cell has room to improve massively if IBM would just push their own tech once in a while. Isn't Fujitsu making a 32-Core SPARC Proc?

---

Theres more to a processor than just clock speed, if IBM worked on improving the SPE and changing from their silly little SDK to something like OpenCL I could see massive potential.

I believe IBM has cancelled development of cell

 

[Bubba Satori]

Cell = POWER. I think.

No.

 

[cmaier]

I believe IBM has cancelled development of cell

Yep.

 

[Bubba Satori]

We often forget that POWER and SPARC are still superior architectures to x86/x64.

Not to mention, they can very often destroy Intel/AMD on the clock speed front. 6 GHz native speed, anyone? I'd like to see Intel or AMD do that (especially AMD) without setting fire to everything within a 10-mile radius.

Those are large server application CPUs. Totally inappropriate for consumer PCs. Sparc is dead anyway, for all intents and purposes. Fujitsu will hang on for a few more years. But it will go the way of the DEC Alpha. To the graveyard of great server cpus that couldn't survive in the marketplace.

 

[cmaier]

Those are large server application CPUs. Totally inappropriate for consumer PCs. Sparc is dead anyway, for all intents and purposes. Fujitsu will hang on for a few more years. But it will go the way of the DEC Alpha. To the graveyard of great server cpus that couldn't survive in the marketplace.

Yep. Register windows. LOL.

 

[Pigumon]

12 cores should not be considered massively multicore anymore. According to the following URL, Intel have stated that developers should start thinking about how they would develop for *thousands* of cores.
http://news.cnet.com/8301-13924_3-9981760-64.html

I agree. This is only a 50% upgrade from the 8-core machines. That's wonderful, certainly, but it's also expected and completely in line with how processors are advancing.

Now if it had gone from 8 to 24-core, that would be more "massive".

What I don't understand is what "24 Logical" means. When I render on an 8-core machine, there are 8 little boxes calculating away on screen, not 16. So what good is "24 Logical" if it only amounts to 12?

 

[cmaier]

I agree. This is only a 50% upgrade from the 8-core machines. That's wonderful, certainly, but it's also expected and completely in line with how processors are advancing.

Now if it had gone from 8 to 24-core, that would be more "massive".

What I don't understand is what "24 Logical" means. When I render on an 8-core machine, there are 8 little boxes calculating away on screen, not 16. So what good is "24 Logical" if it only amounts to 12?

Since each core contains multiple physical units (multiple integer adders, an integer shifter, load/store unit, floating point unit, integer multiplier, etc.) that are almost never all simultaneously busy, treating a core as multiple cores allows the OS to schedule multiple threads on the core simultaneously, which raises the likelihood of all the hardware units being kept busy. (This is an oversimplification, but it works). It's not nearly as good as a real extra core, but it can buy you 10% in some conditions.

 

[Pigumon]

Since each core contains multiple physical units (multiple integer adders, an integer shifter, load/store unit, floating point unit, integer multiplier, etc.) that are almost never all simultaneously busy, treating a core as multiple cores allows the OS to schedule multiple threads on the core simultaneously, which raises the likelihood of all the hardware units being kept busy. (This is an oversimplification, but it works). It's not nearly as good as a real extra core, but it can by you 10% in some conditions.

Well, that's awesome! thanks!

 

[Dr.Pants]

Cell = POWER. I think.

Almost. The processor is divided up into the control unit (Power processing element) and the other cores (Synergistic Processing Elements). It was quite a bit of innovation, but now they're mainly being produced for Sony.

TBH, a lot of the X86 world is a bit boring ATM. I was kinda hoping AMD would be allowed to go Fabless.

How's this for potential? Although I think its more vaporware then anything else...