In a nut shell if you believe that poor quality is the result of the products being built in China it might be a good idea to ask who owns the companies building those products. If the company transferred the production from the US or Europe and quality went down who is really to blame?
I try and avoid companies (within limits) that trade US jobs for overseas profits. Computers and TVs I can't really avoid.
Getting back more on topic, has there been anything on the new mini or has news died down?
I try and avoid companies (within limits) that trade US jobs for overseas profits. Computers and TVs I can't really avoid.
It has happened to a number of electronics industry segments and frankly the companies aren't to blame. The consumers are the ones demanding ever cheaper PCs, televisions and the like. Look at stereo equipment, a number of American companies try to hang on but at best create niches for themselves. You don't see their products on the mass market retailer shelves.
Getting back more on topic, has there been anything on the new mini or has news died down?
I've heard nothing and frankly find that frustrating. Based upon rational delays after Intel started shipping hardware we should have seen something by now. The Mini is getting a little old for the price being asked.
On slightly different news Intel did announce today a new high performance XEON, 15 core chip. This chip was delayed by several months so Intel appears to be getting a free ride in the media even if they have screwed up royally. This chip would be ideal as the basis for a new chip in the Mac Pro. I say basis because the core and cache architecture is very nice the rest of the chip is too server specific. Intel screwed up the current implementation by not providing enough PCI Express lanes for the Mac Pro. Lately I've been under the impression that Intels engineering and marketing teams are on two different planets.
In any event that side tracks your question, I've heard nothing at all.
It's time for Apple to start the preparation to migrate a portion of Macs away from Intel.
I wouldn't be surprised if efforts are already underway!
ARM 64-bit will make for a nice lower cost mini in 2-3 years IMO.
It might not do that bad at all right now, put a fat heat sink on the chip and bump the clock rate up a GHZ or two. Even now the chip at the current clock rate does as well as a machine form 2010. Global foundries and others have had ARM cores running at 2.5 GHZ years ago so it is possible to run Apples chip faster. That is if the low power optimizations haven't slowed the chip down. I'd go for the machine if it came out in the $300 to $400 range and gave me respectable performance.
I wouldn't be surprised if efforts are already underway!
It might not do that bad at all right now, put a fat heat sink on the chip and bump the clock rate up a GHZ or two. Even now the chip at the current clock rate does as well as a machine form 2010. Global foundries and others have had ARM cores running at 2.5 GHZ years ago so it is possible to run Apples chip faster. That is if the low power optimizations haven't slowed the chip down. I'd go for the machine if it came out in the $300 to $400 range and gave me respectable performance.
Right. And destroy the only headless desktop other than the very expensive Mac Pro since it would have virtually no software and insufficient horse power to run desktop apps at a reasonable speed anyway.
Even worse if you try to do something like Rosetta.
On slightly different news Intel did announce today a new high performance XEON, 15 core chip. This chip was delayed by several months so Intel appears to be getting a free ride in the media even if they have screwed up royally. This chip would be ideal as the basis for a new chip in the Mac Pro. I say basis because the core and cache architecture is very nice the rest of the chip is too server specific. Intel screwed up the current implementation by not providing enough PCI Express lanes for the Mac Pro.
In what way did Intel screw up the Xeon in the Mac Pro. Of course there's a server focus because it's a server chip and the vast majority of Xeons will go into servers and not workstations with far less need for extra PCIe lanes.
And since when was it late? Did intel promise a launch date before?
"In October of last year we reported on new Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2 CPUs, expected to launch this quarter. The processors will utilize Ivy Bridge architecture, and will have 50% more cores and 25% more cache than the first generation of Xeon E7 products, that were built on Nehalem architecture. In total, there will be 21 different models. Using multiple sources, like this one (PD file), we were able to come up with specifications for most of these processors."
Disappointing but 14nm is hard and frankly with AMD not executing and providing much competition this go around intel has the ability to release conservatively as opposed to pushing the envelope. Even the ARM threat into servers and ultra books has receded somewhat.
So they are pushing their mobile chips harder to see if they can't get take some mobile design wins away from ARM (windows tablets don't count in that regard).
Right. And destroy the only headless desktop other than the very expensive Mac Pro since it would have virtually no software and insufficient horse power to run desktop apps at a reasonable speed anyway.
Even worse if you try to do something like Rosetta.
OS X on ARM is really close. Most API run on both ARM (iOS) & OS X and Xcode targets them just fine. Rosetta is dead
Right. And destroy the only headless desktop other than the very expensive Mac Pro since it would have virtually no software and insufficient horse power to run desktop apps at a reasonable speed anyway.
Apple can address the no software issue with a simple deadline. In XCode a simple recompile would take care of that. Frankly I think most developers would be happy to support Apple with this sort of initiative.
As to performance iPad Air already performs to the level of a 2010 computer. Up the clock rate a bit and Apple can be very closer to today's best chips and a lot better than many of the discount computers selling in that price range. At this point they might as well wait for A8 too.
Even worse if you try to do something like Rosetta.
Well of course, but let's be honest Apples developer tools are very good at targeting multiple architectures.
Obviously success here depends upon how far Apple can up the clock rate of the current A7 and future A8. That ability depends a bit upon how much they are restricting clock rate simply to manage power usage. There is enough evidence, based on 64 bit hardware announced by AMD, that ARMs 64 bit architecture can be clocked much higher and give respectable performance.
In what way did Intel screw up the Xeon in the Mac Pro. Of course there's a server focus because it's a server chip and the vast majority of Xeons will go into servers and not workstations with far less need for extra PCIe lanes.
There is an article someplace on the net where an Intel engineer / manager was talking to reporters at launch where the question of PCI Express lanes came up. She acknowledged that several customers had expressed disappointment with the number of PCI Express lanes. With the shift to high speed SSD storage those PCI Express lanes are becoming very important even in server class machines. In effect they screwed up and customers have let them know this.
And since when was it late? Did intel promise a launch date before?
It was suppose to come at the end of summer into the fall of last year.
"In October of last year we reported on new Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2 CPUs, expected to launch this quarter. The processors will utilize Ivy Bridge architecture, and will have 50% more cores and 25% more cache than the first generation of Xeon E7 products, that were built on Nehalem architecture. In total, there will be 21 different models. Using multiple sources, like this one (PD file), we were able to come up with specifications for most of these processors."
No mention of them being months late but rather appearing as expected.
Maybe plans have changed but there was an expectation of an earlier launch.
It should be noted that Intel has high expectations for this processor and a correspondingly high price to go with it. What would be interesting to see, by the fall of this year for a Mac Pro refresh is a variant that brings back the PCI express lanes yet leaves all of the improvements in place. There is potential here as the 15 core chip has 5 core complexes so a ten core machine is doable.
The market for lower end server CPU is rapidly dwindling. Virtualization has won. You're only going to make money selling the big iron stuff if you're Intel. Expect the scale of low to mid level chips to be reduced in the next few years.
Even Mobile CPU will face a challenge. Tablets make for an already nice compromise. How much R&D is Intel going to put in their mobile lineup when the fat of the market becomes tablet class architectures?
Barring a stumble I expect Apple to continue to hone their skills developing the Ax series processors. Eventually they'll make a go of a desktop class ARM architecture SoC. The advantages are clear.
1. They decide when to rollout new hardware rather than be tied to Intel's release cycle.
2. They design the processor to meet their software needs
3. Cost - ARM is going to remain the cheaper option for now.
Cons
No Thunderbolt, PCI-Express, virtualization
More low level OS X stuff needs to be supported on ARM.
OS X on ARM is really close. Most API run on both ARM (iOS) & OS X and Xcode targets them just fine. Rosetta is dead
OSX on ARM works just fine today somewhere on Apple's campus but you still need to recompile. And then we'll end up with fat binaries again. And it'll be dog slow given the A7 is equivalent to a Core 2 Duo or a 2.5W Bay Trail Atom.
There's zero advantage in running OSX on ARM vs Intel on the desktop. Not even on price given the Bay Trails are supposed to go into relatively inexpensive tablets.
Apple can address the no software issue with a simple deadline. In XCode a simple recompile would take care of that. Frankly I think most developers would be happy to support Apple with this sort of initiative.
As long as those developers are still actively working on that software. Or thinks compiling for just the mini, retesting and handling another SKU worth the effort.
And no, Apple won't simply issue an ultimatum. Time and time again we see that Apple is very patient with developers and provides great support for transitions.
Quote:
As to performance iPad Air already performs to the level of a 2010 computer. Up the clock rate a bit and Apple can be very closer to today's best chips and a lot better than many of the discount computers selling in that price range. At this point they might as well wait for A8 too.
No, it performs at the level of a 2008 computer. And that's fine for light tasks but not for desktop use.
Quote:
Obviously success here depends upon how far Apple can up the clock rate of the current A7 and future A8. That ability depends a bit upon how much they are restricting clock rate simply to manage power usage. There is enough evidence, based on 64 bit hardware announced by AMD, that ARMs 64 bit architecture can be clocked much higher and give respectable performance.
And none of those compare favorably with the Core i5 much less the Core i7 in the current minis. You're talking about neutering Apple's best bang for the buck computer.
There is an article someplace on the net where an Intel engineer / manager was talking to reporters at launch where the question of PCI Express lanes came up. She acknowledged that several customers had expressed disappointment with the number of PCI Express lanes. With the shift to high speed SSD storage those PCI Express lanes are becoming very important even in server class machines. In effect they screwed up and customers have let them know this.
When you find this article let me know. There are 40 PCIe 3.0 lanes in comparison to 16+4 lanes in a Core i7. And given that it is socket compatible with Sandy Bridge Xeons there's no way to get more lanes anyway so how the hell anyone is disappointed is beyond me.
And that's 40 lanes per socket with dual QPI links between CPUs so if Apple REALLY felt like there wasn't enough PCIe lanes they'd have gone with a dual CPU configuration and 80 PCIe 3.0 lanes.
And there's a limit to how many PCIe SSDs can be deployed on one server currently anyway. Even the new servers designed with PCIe SSDs in mind have limited PCIe SSD capabilities. For example one of the new Dells has 4 front panel PCIe SSD slots + 7 more internal PCIe slots. You could do some kind of dense pack like IBM but with 40 lanes of PCIe 3.0 you can support up to 20 PCIe SSDs minus whatever you are using for the GPUs. This is done by using 8 lanes of PCIe 3.0 into a PCIe bridge to generate a x4 PCIe 2.0 channel for each of four PCIe SSD controllers.
A little short of 2GB/s but on the Mac Pro the SSD uses 4 PCI 2.0 lanes off the C600 PCH for 2GB/s but the PCH is talking to the CPU via a 2GB/s DMI. Meaning in the Mac Pro that SSD is competing with ethernet/wireless and USB 3.0 throughput on the 2GB/s DMI link. So splitting 7.8GB/s into 1.95 GB/s for each SSD controller should be okay.
Quote:
It should be noted that Intel has high expectations for this processor and a correspondingly high price to go with it. What would be interesting to see, by the fall of this year for a Mac Pro refresh is a variant that brings back the PCI express lanes yet leaves all of the improvements in place. There is potential here as the 15 core chip has 5 core complexes so a ten core machine is doable.
"Brings back the PCI express lanes" from where? The E7 parts added more QPI lanes but still only have 40 lanes of PCIe per socket. Of course if you go 4 way then you have up to 160 lanes available + 16 DMI lanes.
The market for lower end server CPU is rapidly dwindling. Virtualization has won. You're only going to make money selling the big iron stuff if you're Intel. Expect the scale of low to mid level chips to be reduced in the next few years.
Why? Because they are going to use ARM or something? Low end servers aren't necessarily dwindling, they're moving from the IT shops and into cloud providers. And obviously Lenovo thinks there's a market there.
Brawny cores still beat wimpy cores, most of the time
Urs Hölzle
Google
Slower but energy efficient “wimpy” cores only win for general workloads if their single-core speed is reasonably close to that of mid-range “brawny” cores.
"We are skeptical that chips [from] outside of Intel and Advanced Micro Devices’ x86 [architecture] will find much uptake outside of niche use cases. Even if you could drop the power consumption of x86 by 25%, it would not make that much of a difference. Chip architectures don’t change because of 25% power improvements. You have got to have 10x power improvements to make that happen,” said Pat Gelsinger at EMC and VMware strategic forum for institutional investors."
With Intel delaying Broadwell, what does this mean for the future?
Intel CEO Brian Krzanich said Next-gen Broadwell chip: The company expects Broadwell -- the successor to the current Haswell used in PCs -- production to start later in the first quarter. Broadwell's release: "still second half of 2014. Squarely on target.".
I'm not sure what's going on these days. We haven't gotten much out of Apple. I'm stuck with Lion until they manage to get out new hardware. I've never been two OS's behind before. It's so uncomfortable hitting that wall. There's absolutely no way I can buy a current version Mac Mini now after it's received absolutely no attention from the Engineers at Apple for over a year. They ought to be ashamed to sell something as "new" when they know full well it's outdated. I think the most upsetting thing about the Mac Mini is that it's the bastard red-headed step-child that Johnny Ive kicks in the corner every chance he gets while making sure to never provide people the equivalent of an iMac in a headless configuration. Might as well throw in a 600MHz G3 processor in there with 2GB of RAM. He certainly doesn't care about the Mac Mini. None of them there do. It's flagrantly obvious when they bring out a new machine that doesn't deserve to be considered "current". I'm proud of him and his team for the Mac Pro, but they need to realize that we're not all frickin' rich. We don't want all-in-one's either. We want a computer we can be proud of, not one we want to hide away so that no one sees that we're running something that was sluggish the day it was released. What ever happened to raising the bar, rather than letting it hit your feet Ive? Going to the Apple Store and seeing it offered as "new" is a slap in the face to everyone.
It's time the Knight get off his butt and provide a Mac Mini that doesn't suck. And the talk above about Mac's switching to ARM processors is designed to make people vomit right? That's just the underpowered crap I'm talking about. I'd love there to be a company out there that takes iMac's and guts them. Puts them in a nice small case (and no I don't care if there's a fan or not), and sells them. I'd buy one right away. I don't want an all-in-one. I don't want an underpowered piece of crap from the Knight. You might think there'd be some Mac lovers that still work at Apple. They certainly aren't around when the Mini is being created. I've waited far too long for this to continue being kind about this issue. It genuinely pisses me off that the people at Apple don't care unless it's the biggest seller.
Intel CEO Brian Krzanich said <strong style="border:0px;color:rgb(0,0,0);margin:0px;padding:0px;vertical-align:baseline;">Next-gen Broadwell chip:</strong> The company expects Broadwell -- the successor to the current Haswell used in PCs -- production to start later in the first quarter. Broadwell's release: "still second half of 2014. Squarely on target.".
There's absolutely no way I can buy a current version Mac Mini now after it's received absolutely no attention from the Engineers at Apple for over a year.
It's because hardly anyone buys them even when they update them. I reckon they should try ditching the entry model and just build one with the entry 15" MBP chip with Iris Pro. This pushes up the average selling price at least.
And the talk above about Mac's switching to ARM processors is designed to make people vomit right That's just the underpowered crap I'm talking about.
That's a misconception with ARM, they wouldn't use 2W chips in a machine designed to handle 45W. They could use 4x A7s and it would rival a 2011 MBP quad-i7 with a much lower TDP and likely cost over $200 less, although they'd have to add a laptop GPU on top.
It would be incompatible with a lot of software but they could do binary translation like they did with PPC and it wouldn't be near as noticeable a performance hit. No VMWare or native Windows (except maybe RT) but it would be a very low power server and run cooler.
They can perhaps just make the server model have the ARM chip and make the desktop one use the i7 with the server one at $599. The server one would even get away with using a PowerVR GPU.
Comments
I try and avoid companies (within limits) that trade US jobs for overseas profits. Computers and TVs I can't really avoid.
Getting back more on topic, has there been anything on the new mini or has news died down?
No news on mac mini yet.
http://www.fudzilla.com/home/item/33968-broadwell-delay-all-but-confirmed
It's time for Apple to start the preparation to migrate a portion of Macs away from Intel.
ARM 64-bit will make for a nice lower cost mini in 2-3 years IMO.
I've heard nothing and frankly find that frustrating. Based upon rational delays after Intel started shipping hardware we should have seen something by now. The Mini is getting a little old for the price being asked.
On slightly different news Intel did announce today a new high performance XEON, 15 core chip. This chip was delayed by several months so Intel appears to be getting a free ride in the media even if they have screwed up royally. This chip would be ideal as the basis for a new chip in the Mac Pro. I say basis because the core and cache architecture is very nice the rest of the chip is too server specific. Intel screwed up the current implementation by not providing enough PCI Express lanes for the Mac Pro. Lately I've been under the impression that Intels engineering and marketing teams are on two different planets.
In any event that side tracks your question, I've heard nothing at all.
It might not do that bad at all right now, put a fat heat sink on the chip and bump the clock rate up a GHZ or two. Even now the chip at the current clock rate does as well as a machine form 2010. Global foundries and others have had ARM cores running at 2.5 GHZ years ago so it is possible to run Apples chip faster. That is if the low power optimizations haven't slowed the chip down. I'd go for the machine if it came out in the $300 to $400 range and gave me respectable performance.
Right. And destroy the only headless desktop other than the very expensive Mac Pro since it would have virtually no software and insufficient horse power to run desktop apps at a reasonable speed anyway.
Even worse if you try to do something like Rosetta.
In what way did Intel screw up the Xeon in the Mac Pro. Of course there's a server focus because it's a server chip and the vast majority of Xeons will go into servers and not workstations with far less need for extra PCIe lanes.
And since when was it late? Did intel promise a launch date before?
"In October of last year we reported on new Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2 CPUs, expected to launch this quarter. The processors will utilize Ivy Bridge architecture, and will have 50% more cores and 25% more cache than the first generation of Xeon E7 products, that were built on Nehalem architecture. In total, there will be 21 different models. Using multiple sources, like this one (PD file), we were able to come up with specifications for most of these processors."
http://www.cpu-world.com/news_2014/2014020201_Specifications_of_Xeon_E7_v2_processors.html
No mention of them being months late but rather appearing as expected.
Disappointing but 14nm is hard and frankly with AMD not executing and providing much competition this go around intel has the ability to release conservatively as opposed to pushing the envelope. Even the ARM threat into servers and ultra books has receded somewhat.
So they are pushing their mobile chips harder to see if they can't get take some mobile design wins away from ARM (windows tablets don't count in that regard).
Right. And destroy the only headless desktop other than the very expensive Mac Pro since it would have virtually no software and insufficient horse power to run desktop apps at a reasonable speed anyway.
Even worse if you try to do something like Rosetta.
OS X on ARM is really close. Most API run on both ARM (iOS) & OS X and Xcode targets them just fine. Rosetta is dead
As to performance iPad Air already performs to the level of a 2010 computer. Up the clock rate a bit and Apple can be very closer to today's best chips and a lot better than many of the discount computers selling in that price range. At this point they might as well wait for A8 too.
Well of course, but let's be honest Apples developer tools are very good at targeting multiple architectures.
Obviously success here depends upon how far Apple can up the clock rate of the current A7 and future A8. That ability depends a bit upon how much they are restricting clock rate simply to manage power usage. There is enough evidence, based on 64 bit hardware announced by AMD, that ARMs 64 bit architecture can be clocked much higher and give respectable performance.
Maybe plans have changed but there was an expectation of an earlier launch.
It should be noted that Intel has high expectations for this processor and a correspondingly high price to go with it. What would be interesting to see, by the fall of this year for a Mac Pro refresh is a variant that brings back the PCI express lanes yet leaves all of the improvements in place. There is potential here as the 15 core chip has 5 core complexes so a ten core machine is doable.
Intel sees the writing on the wall.
Lenovo buys IBM's server biz for 2.3 Billion
The market for lower end server CPU is rapidly dwindling. Virtualization has won. You're only going to make money selling the big iron stuff if you're Intel. Expect the scale of low to mid level chips to be reduced in the next few years.
Even Mobile CPU will face a challenge. Tablets make for an already nice compromise. How much R&D is Intel going to put in their mobile lineup when the fat of the market becomes tablet class architectures?
Barring a stumble I expect Apple to continue to hone their skills developing the Ax series processors. Eventually they'll make a go of a desktop class ARM architecture SoC. The advantages are clear.
1. They decide when to rollout new hardware rather than be tied to Intel's release cycle.
2. They design the processor to meet their software needs
3. Cost - ARM is going to remain the cheaper option for now.
Cons
No Thunderbolt, PCI-Express, virtualization
More low level OS X stuff needs to be supported on ARM.
OS X on ARM is really close. Most API run on both ARM (iOS) & OS X and Xcode targets them just fine. Rosetta is dead
OSX on ARM works just fine today somewhere on Apple's campus but you still need to recompile. And then we'll end up with fat binaries again. And it'll be dog slow given the A7 is equivalent to a Core 2 Duo or a 2.5W Bay Trail Atom.
There's zero advantage in running OSX on ARM vs Intel on the desktop. Not even on price given the Bay Trails are supposed to go into relatively inexpensive tablets.
Apple can address the no software issue with a simple deadline. In XCode a simple recompile would take care of that. Frankly I think most developers would be happy to support Apple with this sort of initiative.
As long as those developers are still actively working on that software. Or thinks compiling for just the mini, retesting and handling another SKU worth the effort.
And no, Apple won't simply issue an ultimatum. Time and time again we see that Apple is very patient with developers and provides great support for transitions.
No, it performs at the level of a 2008 computer. And that's fine for light tasks but not for desktop use.
And none of those compare favorably with the Core i5 much less the Core i7 in the current minis. You're talking about neutering Apple's best bang for the buck computer.
There is an article someplace on the net where an Intel engineer / manager was talking to reporters at launch where the question of PCI Express lanes came up. She acknowledged that several customers had expressed disappointment with the number of PCI Express lanes. With the shift to high speed SSD storage those PCI Express lanes are becoming very important even in server class machines. In effect they screwed up and customers have let them know this.
When you find this article let me know. There are 40 PCIe 3.0 lanes in comparison to 16+4 lanes in a Core i7. And given that it is socket compatible with Sandy Bridge Xeons there's no way to get more lanes anyway so how the hell anyone is disappointed is beyond me.
And that's 40 lanes per socket with dual QPI links between CPUs so if Apple REALLY felt like there wasn't enough PCIe lanes they'd have gone with a dual CPU configuration and 80 PCIe 3.0 lanes.
And there's a limit to how many PCIe SSDs can be deployed on one server currently anyway. Even the new servers designed with PCIe SSDs in mind have limited PCIe SSD capabilities. For example one of the new Dells has 4 front panel PCIe SSD slots + 7 more internal PCIe slots. You could do some kind of dense pack like IBM but with 40 lanes of PCIe 3.0 you can support up to 20 PCIe SSDs minus whatever you are using for the GPUs. This is done by using 8 lanes of PCIe 3.0 into a PCIe bridge to generate a x4 PCIe 2.0 channel for each of four PCIe SSD controllers.
A little short of 2GB/s but on the Mac Pro the SSD uses 4 PCI 2.0 lanes off the C600 PCH for 2GB/s but the PCH is talking to the CPU via a 2GB/s DMI. Meaning in the Mac Pro that SSD is competing with ethernet/wireless and USB 3.0 throughput on the 2GB/s DMI link. So splitting 7.8GB/s into 1.95 GB/s for each SSD controller should be okay.
"Brings back the PCI express lanes" from where? The E7 parts added more QPI lanes but still only have 40 lanes of PCIe per socket. Of course if you go 4 way then you have up to 160 lanes available + 16 DMI lanes.
Intel sees the writing on the wall.
Lenovo buys IBM's server biz for 2.3 Billion
The market for lower end server CPU is rapidly dwindling. Virtualization has won. You're only going to make money selling the big iron stuff if you're Intel. Expect the scale of low to mid level chips to be reduced in the next few years.
Why? Because they are going to use ARM or something? Low end servers aren't necessarily dwindling, they're moving from the IT shops and into cloud providers. And obviously Lenovo thinks there's a market there.
Brawny cores still beat wimpy cores, most of the time
Urs Hölzle
Google
Slower but energy efficient “wimpy” cores only win for general workloads if their single-core speed is reasonably close to that of mid-range “brawny” cores.
http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/pubs/archive/36448.pdf
And VMWare seems to think the same:
"We are skeptical that chips [from] outside of Intel and Advanced Micro Devices’ x86 [architecture] will find much uptake outside of niche use cases. Even if you could drop the power consumption of x86 by 25%, it would not make that much of a difference. Chip architectures don’t change because of 25% power improvements. You have got to have 10x power improvements to make that happen,” said Pat Gelsinger at EMC and VMware strategic forum for institutional investors."
http://www.xbitlabs.com/news/other/display/20130313235700_Chief_Exec_of_VMware_ARM_Chips_Will_Play_a_Niche_Role_on_Server_Market.html
/shrug
I think that server CPUs at all levels have a place in these cloud compute farms and ISP colos.
With Intel delaying Broadwell, what does this mean for the future?
Intel CEO Brian Krzanich said Next-gen Broadwell chip: The company expects Broadwell -- the successor to the current Haswell used in PCs -- production to start later in the first quarter. Broadwell's release: "still second half of 2014. Squarely on target.".
http://news.cnet.com/8301-1001_3-57617372-92/intel-ceo-talks-delayed-factory-broadwell-production-start/
I'm not sure what's going on these days. We haven't gotten much out of Apple. I'm stuck with Lion until they manage to get out new hardware. I've never been two OS's behind before. It's so uncomfortable hitting that wall. There's absolutely no way I can buy a current version Mac Mini now after it's received absolutely no attention from the Engineers at Apple for over a year. They ought to be ashamed to sell something as "new" when they know full well it's outdated. I think the most upsetting thing about the Mac Mini is that it's the bastard red-headed step-child that Johnny Ive kicks in the corner every chance he gets while making sure to never provide people the equivalent of an iMac in a headless configuration. Might as well throw in a 600MHz G3 processor in there with 2GB of RAM. He certainly doesn't care about the Mac Mini. None of them there do. It's flagrantly obvious when they bring out a new machine that doesn't deserve to be considered "current". I'm proud of him and his team for the Mac Pro, but they need to realize that we're not all frickin' rich. We don't want all-in-one's either. We want a computer we can be proud of, not one we want to hide away so that no one sees that we're running something that was sluggish the day it was released. What ever happened to raising the bar, rather than letting it hit your feet Ive? Going to the Apple Store and seeing it offered as "new" is a slap in the face to everyone.
It's time the Knight get off his butt and provide a Mac Mini that doesn't suck. And the talk above about Mac's switching to ARM processors is designed to make people vomit right? That's just the underpowered crap I'm talking about. I'd love there to be a company out there that takes iMac's and guts them. Puts them in a nice small case (and no I don't care if there's a fan or not), and sells them. I'd buy one right away. I don't want an all-in-one. I don't want an underpowered piece of crap from the Knight. You might think there'd be some Mac lovers that still work at Apple. They certainly aren't around when the Mini is being created. I've waited far too long for this to continue being kind about this issue. It genuinely pisses me off that the people at Apple don't care unless it's the biggest seller.
The following article suggests later 2014:
http://semiaccurate.com/2014/02/19/sky-falling-intels-14nm-broadwell/
There's a Haswell refresh coming first:
http://www.escapistmagazine.com/news/view/132295-Rumor-Intel-Moves-Haswell-CPU-Refresh-From-May-to-April
They might be better off waiting until April.
It's because hardly anyone buys them even when they update them. I reckon they should try ditching the entry model and just build one with the entry 15" MBP chip with Iris Pro. This pushes up the average selling price at least.
That's a misconception with ARM, they wouldn't use 2W chips in a machine designed to handle 45W. They could use 4x A7s and it would rival a 2011 MBP quad-i7 with a much lower TDP and likely cost over $200 less, although they'd have to add a laptop GPU on top.
It would be incompatible with a lot of software but they could do binary translation like they did with PPC and it wouldn't be near as noticeable a performance hit. No VMWare or native Windows (except maybe RT) but it would be a very low power server and run cooler.
They can perhaps just make the server model have the ARM chip and make the desktop one use the i7 with the server one at $599. The server one would even get away with using a PowerVR GPU.