The performace of Merom is good but only modestly(generally on the order of 5-10%) better than current Yonah chips.
The lack of a massive performance increase relative to the part that's being replaced is due in part to the performance of the part that's being replaced relative to the P4 that's being edged out in favor of Conroe, coupled with the existence of bottlenecks elsewhere in the performance equation that mask CPU performance increases (especially with games, which is the same problem Conroe and Woodcrest face when compared with markedly inferior AMD X2 chips). Many current games are GPU limited at most resolutions using any of the new CPUs, so you can easily be disappointed by benchmarks if your expectations are poorly set.
Even still, the expected 20%+ performance increase clock for clock over Yonah is there to be had, with the right sort of application that lets the CPU stretch its legs. It's even visible in the charts you linked yourself.
As I stated, clock for clock there is little significant performance difference between any of the new chips. There is a clock speed difference of 14% for the high end Core 2 Duo E6700 CPU. This is still negligible compared to the differences imposed by the architecture of the system itself (GPU performance capabilities chief among them). In other words, the limitations inherent in the design of the iMac will weigh more heavily than a minor clock speed increase between the various Core 2 Duo parts. I'm much more concerned about RAM capacities and GPU selection.
Finally, the benefits of a truly quiet system are appreciated by a great many people. The current iMac is fantastic in this regard, and losing that just to go with a hotter chip for a miniscule performance delta would represent a terrific failure. If you don't mind a system that produces more noise than the near-silent iMac, the Mac Pro is a great alternative with plenty of room to grow.
I actually wish my soon to be retired G4 tower was an AIO. If it were an AIO that had a screen "that had to be thrown away" the machine would easily make it's way into the kitchen or some other room where I don't currently have a computer.
Instead, I'm stuck with a reusable screen which isn't that great by today's standards. I'm stuck with a computer that isn't unobtrusive enough to put in a room that only sort of needs a computer.
These "throw-away" screens in AIO machines are actually _more_ reusable and have a longer life than the separates, at least for me.
The Xeon 51xx has a third faster FSB and more L2 cache. 533Mhz FB-DIMMs which are error correcting and more reliable than non-ECC 667Mhz DDR2 whallop the old technology in many tests.
Performance-wise:
Quad channel = good
FB-DIMM = bad
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
The ports of the iMac located on the back of the screen instead of the front.
Wouldn't it be a option to locate some extra ports out of sight under the computer case.
It will look better than ugly holes in the glossy front of the iMac and be easier accessible than on the back.
Would Apple introduce a Black One?
Out of sight is a good solution. I use a reasonably good-looking USB hub with mine tucked beneath and a little behind the display, mostly for capacity but accessibility is nice too. I wouldn't really want holes (and the plugs / cords while in use) on the front of my iMac. Or the side.
The lack of a massive performance increase relative to the part that's being replaced is due in part to the performance of the part that's being replaced relative to the P4 that's being edged out in favor of Conroe, coupled with the existence of bottlenecks elsewhere in the performance equation that mask CPU performance increases (especially with games, which is the same problem Conroe and Woodcrest face when compared with markedly inferior AMD X2 chips). Many current games are GPU limited at most resolutions using any of the new CPUs, so you can easily be disappointed by benchmarks if your expectations are poorly set.
Even still, the expected 20%+ performance increase clock for clock over Yonah is there to be had, with the right sort of application that lets the CPU stretch its legs. It's even visible in the charts you linked yourself.
As I stated, clock for clock there is little significant performance difference between any of the new chips. There is a clock speed difference of 14% for the high end Core 2 Duo E6700 CPU. This is still negligible compared to the differences imposed by the architecture of the system itself (GPU performance capabilities chief among them). In other words, the limitations inherent in the design of the iMac will weigh more heavily than a minor clock speed increase between the various Core 2 Duo parts. I'm much more concerned about RAM capacities and GPU selection.
Finally, the benefits of a truly quiet system are appreciated by a great many people. The current iMac is fantastic in this regard, and losing that just to go with a hotter chip for a miniscule performance delta would represent a terrific failure. If you don't mind a system that produces more noise than the near-silent iMac, the Mac Pro is a great alternative with plenty of room to grow.
The yonah to Merom boost may be larger on OSX than windows because of out superior SIMD support.
I actually wish my soon to be retired G4 tower was an AIO. If it were an AIO that had a screen "that had to be thrown away" the machine would easily make it's way into the kitchen or some other room where I don't currently have a computer.
Instead, I'm stuck with a reusable screen which isn't that great by today's standards. I'm stuck with a computer that isn't unobtrusive enough to put in a room that only sort of needs a computer.
These "throw-away" screens in AIO machines are actually _more_ reusable and have a longer life than the separates, at least for me.
Well, if you buy Apple monitors, that's definitely true. But, I'd take a Mac mini or a Mac Pro with a nice HDTV, HDCP capable, VESA mountable monitor any day over an iMac. Especially when you're talking about one with $800+ throw-away 23" display built-in.
The G5's were in the 65 watt range. Remember that IBM doesn't rate the chips as to MAX power use as Intel does. They rate them for average power, so the numbers look better upon first glance, and are more confusing.
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
i just want to know how easy, feasible, and affordable it would be to replace the Yonah processor in my iMac with a Merom (I'm more interested in the 64-bit, especially with Leopard's imminent release, than with any performance gains).
i just want to know how easy, feasible, and affordable it would be to replace the Yonah processor in my iMac with a Merom (I'm more interested in the 64-bit, especially with Leopard's imminent release, than with any performance gains).
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
I was under the impression that FB-DIMM is higher-latency but also higher-bandwidth.
I actually wish my soon to be retired G4 tower was an AIO. If it were an AIO that had a screen "that had to be thrown away" the machine would easily make it's way into the kitchen or some other room where I don't currently have a computer.
Yeah, like the bathroom! It's a great way to increase productivity.
39watts max dissipation, not power draw to run the thing.
Irrelevant. We're talking about thermal characteristics. The 2.0Ghz G5 dissipates MAX 39W according to IBM in heat. I'd always presumed it was more like 65W MAX too like Melgross so I went and looked.
The Intel Conroe has a TDP (ie. the THERMAL output) of 65W and up to 80 for the Extreme model.
The Merom/Yonah has a TDP of 35W - ie. just less than the G5.
That's why it's unlikely we'd see Conroe in an iMac or Mini as Apple would need a much much noisier cooling solution or a larger case.
Conroe does not fit in Apple's line. They have no qualms about using portable chips in iMacs. Said long ago that Apple wouldn't use Conroe. I love being right.
Conroe does not fit in Apple's line. They have no qualms about using portable chips in iMacs. Said long ago that Apple wouldn't use Conroe. I love being right.
Don't count on it. It could be true, but I think this is just speculative deductions from AI because they can't figure out why Apple hasn't updated iMac with new chips. The most simple reason for that is 1.) Apple still has some inventory of iMacs, and 2.) It's a bit too recent to suddenly spring an updated Conroe model on users at the same price that 1000's of users just paid a few months ago for one. Apple is on a roll and this would cause some serious upset to a lot of new, and old Mac users alike. No need to update it until they feel sufficient time has passed. Waiting does not effect Apple. Dozens of you were trying to tell me Apple had to have the woodcrest Mac Pro announced at the same time PC manufacturers did. I alone said Apple would hold off until WWDC, but no one believed me. Everybody said Apple couldn't wait the 3 months because of some loss of PC sales BS. Well...
x86-64 offers more than just more memory addressing over standard IA32. This was already covered in this thread.
Quote:
Originally Posted by doh123
39watts max dissipation, not power draw to run the thing.
From the chip perspective, they are very close to the same number. In equilibrium, power in equals power out, the chip is very good at converting electrical power into heat. Any other heat sources or sinks are negligible.
Quote:
Originally Posted by onlooker
It's a bit too recent to suddenly spring an updated Conroe model on users at the same price that 1000's of users just paid a few months ago for one. Apple is on a roll and this would cause some serious upset to a lot of new, and old Mac users alike.
You may be right (though I now somewhat doubt it will be Conroe, Merom is pretty likely), but I don't think updates will be held back for that reason. I think that's silly reasoning. That's the nature of technology, no tech company I am aware of holds back their updates because it would upset current owners. Time goes on and technology marches on, we have to deal with it. By your reasoning, the iSight G5 and late 2005 Powerbook owners should have been in a big uproar when they were replaced only three months later with iMac Core Duo, having only been available for three months. The iMac Core Duo has been out for about eight months already. The iMac is overdue for an update, the average time between iMac updates is six about months. I do agree that Apple will update when they are ready, from a supply and marketing standpoint. I just don't think that fearing anger from current owners factor into it.
Quote:
Dozens of you were trying to tell me Apple had to have the woodcrest Mac Pro announced at the same time PC manufacturers did. I alone said Apple would hold off until WWDC, but no one believed me. Everybody said Apple couldn't wait the 3 months because of some loss of PC sales BS. Well...
I thought I and several others said that Woodcrest Mac Pro was likely going to be announced at WWDC. I don't think it was an everyone-against-you deal.
Comments
The performace of Merom is good but only modestly(generally on the order of 5-10%) better than current Yonah chips.
The lack of a massive performance increase relative to the part that's being replaced is due in part to the performance of the part that's being replaced relative to the P4 that's being edged out in favor of Conroe, coupled with the existence of bottlenecks elsewhere in the performance equation that mask CPU performance increases (especially with games, which is the same problem Conroe and Woodcrest face when compared with markedly inferior AMD X2 chips). Many current games are GPU limited at most resolutions using any of the new CPUs, so you can easily be disappointed by benchmarks if your expectations are poorly set.
Even still, the expected 20%+ performance increase clock for clock over Yonah is there to be had, with the right sort of application that lets the CPU stretch its legs. It's even visible in the charts you linked yourself.
As I stated, clock for clock there is little significant performance difference between any of the new chips. There is a clock speed difference of 14% for the high end Core 2 Duo E6700 CPU. This is still negligible compared to the differences imposed by the architecture of the system itself (GPU performance capabilities chief among them). In other words, the limitations inherent in the design of the iMac will weigh more heavily than a minor clock speed increase between the various Core 2 Duo parts. I'm much more concerned about RAM capacities and GPU selection.
Finally, the benefits of a truly quiet system are appreciated by a great many people. The current iMac is fantastic in this regard, and losing that just to go with a hotter chip for a miniscule performance delta would represent a terrific failure. If you don't mind a system that produces more noise than the near-silent iMac, the Mac Pro is a great alternative with plenty of room to grow.
Shows how much people want a headless Mac instead of an iMac.
No it doesn't.
It just shows that some people don't leave an option unused to whine about their beloved headless Mac.
Back on one of the topics in this thread:
The ports of the iMac located on the back of the screen instead of the front.
Wouldn't it be a option to locate some extra ports out of sight under the computer case.
It will look better than ugly holes in the glossy front of the iMac and be easier accessible than on the back.
Would Apple introduce a Black One?
Instead, I'm stuck with a reusable screen which isn't that great by today's standards. I'm stuck with a computer that isn't unobtrusive enough to put in a room that only sort of needs a computer.
These "throw-away" screens in AIO machines are actually _more_ reusable and have a longer life than the separates, at least for me.
The Xeon 51xx has a third faster FSB and more L2 cache. 533Mhz FB-DIMMs which are error correcting and more reliable than non-ECC 667Mhz DDR2 whallop the old technology in many tests.
Performance-wise:
Quad channel = good
FB-DIMM = bad
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
Back on one of the topics in this thread:
The ports of the iMac located on the back of the screen instead of the front.
Wouldn't it be a option to locate some extra ports out of sight under the computer case.
It will look better than ugly holes in the glossy front of the iMac and be easier accessible than on the back.
Would Apple introduce a Black One?
Out of sight is a good solution. I use a reasonably good-looking USB hub with mine tucked beneath and a little behind the display, mostly for capacity but accessibility is nice too. I wouldn't really want holes (and the plugs / cords while in use) on the front of my iMac. Or the side.
macosrumors...ha, ha, ha!
Mind you, this is MacOSXRumors we talk about, NOT MacOSRumors.
Mind you, this is MacOSXRumors we talk about, NOT MacOSRumors.
You can really determine the value of something to society based on the number of rumor sites about it.
The lack of a massive performance increase relative to the part that's being replaced is due in part to the performance of the part that's being replaced relative to the P4 that's being edged out in favor of Conroe, coupled with the existence of bottlenecks elsewhere in the performance equation that mask CPU performance increases (especially with games, which is the same problem Conroe and Woodcrest face when compared with markedly inferior AMD X2 chips). Many current games are GPU limited at most resolutions using any of the new CPUs, so you can easily be disappointed by benchmarks if your expectations are poorly set.
Even still, the expected 20%+ performance increase clock for clock over Yonah is there to be had, with the right sort of application that lets the CPU stretch its legs. It's even visible in the charts you linked yourself.
As I stated, clock for clock there is little significant performance difference between any of the new chips. There is a clock speed difference of 14% for the high end Core 2 Duo E6700 CPU. This is still negligible compared to the differences imposed by the architecture of the system itself (GPU performance capabilities chief among them). In other words, the limitations inherent in the design of the iMac will weigh more heavily than a minor clock speed increase between the various Core 2 Duo parts. I'm much more concerned about RAM capacities and GPU selection.
Finally, the benefits of a truly quiet system are appreciated by a great many people. The current iMac is fantastic in this regard, and losing that just to go with a hotter chip for a miniscule performance delta would represent a terrific failure. If you don't mind a system that produces more noise than the near-silent iMac, the Mac Pro is a great alternative with plenty of room to grow.
The yonah to Merom boost may be larger on OSX than windows because of out superior SIMD support.
I actually wish my soon to be retired G4 tower was an AIO. If it were an AIO that had a screen "that had to be thrown away" the machine would easily make it's way into the kitchen or some other room where I don't currently have a computer.
Instead, I'm stuck with a reusable screen which isn't that great by today's standards. I'm stuck with a computer that isn't unobtrusive enough to put in a room that only sort of needs a computer.
These "throw-away" screens in AIO machines are actually _more_ reusable and have a longer life than the separates, at least for me.
Well, if you buy Apple monitors, that's definitely true. But, I'd take a Mac mini or a Mac Pro with a nice HDTV, HDCP capable, VESA mountable monitor any day over an iMac. Especially when you're talking about one with $800+ throw-away 23" display built-in.
The G5's were in the 65 watt range. Remember that IBM doesn't rate the chips as to MAX power use as Intel does. They rate them for average power, so the numbers look better upon first glance, and are more confusing.
The 2.0Ghz 970FX is rated at 39W MAX by IBM. See http://www-306.ibm.com/chips/techlib...256E4B005E43EC
Performance-wise:
Quad channel = good
FB-DIMM = bad
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
Citation?
I read the opposite about FB-DIMM...
http://www.phoronix.com/scan.php?pag...item=495&num=1
i just want to know how easy, feasible, and affordable it would be to replace the Yonah processor in my iMac with a Merom (I'm more interested in the 64-bit, especially with Leopard's imminent release, than with any performance gains).
Why? you can't fit more than 2GB RAM anyway?
Performance-wise:
Quad channel = good
FB-DIMM = bad
Current FB-DIMM implementations (especially as seen in the Mac Pro) represent a performance decrease in every measurable category compared with DDR2. Higher latency than even the MacBook and iMac, and lower bandwidth than desktop DDR2 systems across the board to boot. Measured memory bandwidth of the new Mac Pro in quad channel mode is almost exactly the same as the dual-channel (and slower clocked) MacBook and iMac, which is disappointing and blunts the bus speed advantages suggested by the Xeon.
I was under the impression that FB-DIMM is higher-latency but also higher-bandwidth.
I actually wish my soon to be retired G4 tower was an AIO. If it were an AIO that had a screen "that had to be thrown away" the machine would easily make it's way into the kitchen or some other room where I don't currently have a computer.
Yeah, like the bathroom! It's a great way to increase productivity.
The 2.0Ghz 970FX is rated at 39W MAX by IBM. See http://www-306.ibm.com/chips/techlib...256E4B005E43EC
39watts max dissipation, not power draw to run the thing.
39watts max dissipation, not power draw to run the thing.
Irrelevant. We're talking about thermal characteristics. The 2.0Ghz G5 dissipates MAX 39W according to IBM in heat. I'd always presumed it was more like 65W MAX too like Melgross so I went and looked.
The Intel Conroe has a TDP (ie. the THERMAL output) of 65W and up to 80 for the Extreme model.
The Merom/Yonah has a TDP of 35W - ie. just less than the G5.
That's why it's unlikely we'd see Conroe in an iMac or Mini as Apple would need a much much noisier cooling solution or a larger case.
Conroe does not fit in Apple's line. They have no qualms about using portable chips in iMacs. Said long ago that Apple wouldn't use Conroe. I love being right.
Don't count on it. It could be true, but I think this is just speculative deductions from AI because they can't figure out why Apple hasn't updated iMac with new chips. The most simple reason for that is 1.) Apple still has some inventory of iMacs, and 2.) It's a bit too recent to suddenly spring an updated Conroe model on users at the same price that 1000's of users just paid a few months ago for one. Apple is on a roll and this would cause some serious upset to a lot of new, and old Mac users alike. No need to update it until they feel sufficient time has passed. Waiting does not effect Apple. Dozens of you were trying to tell me Apple had to have the woodcrest Mac Pro announced at the same time PC manufacturers did. I alone said Apple would hold off until WWDC, but no one believed me. Everybody said Apple couldn't wait the 3 months because of some loss of PC sales BS. Well...
Why? you can't fit more than 2GB RAM anyway?
x86-64 offers more than just more memory addressing over standard IA32. This was already covered in this thread.
39watts max dissipation, not power draw to run the thing.
From the chip perspective, they are very close to the same number. In equilibrium, power in equals power out, the chip is very good at converting electrical power into heat. Any other heat sources or sinks are negligible.
It's a bit too recent to suddenly spring an updated Conroe model on users at the same price that 1000's of users just paid a few months ago for one. Apple is on a roll and this would cause some serious upset to a lot of new, and old Mac users alike.
You may be right (though I now somewhat doubt it will be Conroe, Merom is pretty likely), but I don't think updates will be held back for that reason. I think that's silly reasoning. That's the nature of technology, no tech company I am aware of holds back their updates because it would upset current owners. Time goes on and technology marches on, we have to deal with it. By your reasoning, the iSight G5 and late 2005 Powerbook owners should have been in a big uproar when they were replaced only three months later with iMac Core Duo, having only been available for three months. The iMac Core Duo has been out for about eight months already. The iMac is overdue for an update, the average time between iMac updates is six about months. I do agree that Apple will update when they are ready, from a supply and marketing standpoint. I just don't think that fearing anger from current owners factor into it.
Dozens of you were trying to tell me Apple had to have the woodcrest Mac Pro announced at the same time PC manufacturers did. I alone said Apple would hold off until WWDC, but no one believed me. Everybody said Apple couldn't wait the 3 months because of some loss of PC sales BS. Well...
I thought I and several others said that Woodcrest Mac Pro was likely going to be announced at WWDC. I don't think it was an everyone-against-you deal.