Intel quad-core will end up the same way or worse then the first Intel dual cores. Linked to each other by FSB where AMD is useing true quad-cores with Shared L3.
You're thinking of quad-core as 4 cores per processor. What I'm talking about is 2 dual-core chips. That'd function identically to how the Dual-G4 Powermacs functioned. There are two sockets, each connected to the north and south bridges by their own FSBs.
You are correct that Intel's Clovertown processor will be two Woodcrests connected by FSB in the same package. But the next generation (mid-2007 or late 2007) will be "true quad-core".
When I'm talking about what AMD is planning, they're planning something that involves 2 dual-core X2s or FXs. That is not "true quad-core". It's a dual-dual setup. Except with desktop chips, not server/workstation chips.
Intel quad-core will end up the same way or worse then the first Intel duel cores. Linked to each other by FSB where AMD is useing true quad-cores with Shared L3.
No doubt, AMD's HyperTransport is a better architecture. Intel Core 2 Duos and Woodcrest do quite well as the 2 cores share the cache without involving the FSB. All the performance numbers in the 1 and 2 socket are in Intel's favor. Socket 4 and up is a different story. Intel's Tulsa (not Core 2 arch) will not stand upto AMD Opteron even with its large L3 cache. Clovertown may be the same as its just two Woodcrests on the same packaging. AMD might be able to wrestle back the performance crown again. The 65nm process may not work for Intel competitively this time as AMD will have moved over to it as well for the Quad core chip. The only solution for Intel is to either develop the Common System Interconnect soon or (if possible) just license the HyperTransport.
No doubt, AMD's HyperTransport is a better architecture. Intel Core 2 Duos and Woodcrest do quite well as the 2 cores share the cache without involving the FSB. All the performance numbers in the 1 and 2 socket are in Intel's favor. Socket 4 and up is a different story. Intel's Tulsa (not Core 2 arch) will not stand upto AMD Opteron even with its large L3 cache. Clovertown may be the same as its just two Woodcrests on the same packaging. AMD might be able to wrestle back the performance crown again. The 65nm process may not work for Intel competitively this time as AMD will have moved over to it as well for the Quad core chip. The only solution for Intel is to either develop the Common System Interconnect soon or (if possible) just license the HyperTransport.
Intel quad-core seem like the hack job of there duel core thay may be the first to have it but a real quad-core kick it ass and if intel went to HyperTransport thay will not make it better then then a true quad-core.
HyperTransport based accelerators may force intel to license HyperTransport and thay sound like a cool think that apple may want.
Here's the real reason the low-end Mac Pros will be Quad: AMD's 4x4 initiative. In theory, AMD might cut low-end X2 or even FX prices by 1/3 to 1/2, which'll leave you with some bargain prices. They are then releasing a motherboard that can have two X2s or FXs (these are Athlons, not Opterons) on a prosumer desktop board.
You'll have to provide a link for AMD's 4x4 being able to use an Athlon X2. As far as I know, it's an "enthusiast" 2S platform using Athlon FX CPUs. Enthusiast doesn't equate to being affordable, and a 4x4 system will likely be in the $2500 to $3000 range.
As far as the low end, Apple's got a dilemma with 2S systems. They are expensive and the floor will be around $2.5k. If they can't sell a 2S 2.33 GHz Woodcrest for $2499, there won't be much point because a 2.67 GHz Conroe will be faster than a 2S 2 GHz Woodcrest by a big margin for many many apps.
The lineup has to be a variant of 1x2.67, 2x2.67, 2x3 or 1x2.67, 1x2.93, 2x3 or perhaps 2x2.33, 2x2.67 or 2x3. An all 2S platform will be very very pricey. If they go all 2S in the Mac Pro, Apple will be forced into developing a Mac Cube to fill the gigantic hole between $1500 and $2500.
Quote:
Additionally, Dell/everyone will have 2.4 or 2.67 GHz Conroes shipping for $1500.
Apple has had no entry in the mid-range headless desktop market for years now. I'd guess that they would be happy with selling Mac minis and iMacs and leave a gigantic gulf between them and the Mac Pros.
1) In theory, Apple could (if they wanted to) hack together a "normal" memory controller for the Mac Pros and leave the FB-DIMMs to the XServe. It'd be annoying, but cheaper and not killer performance-wise.
If Apple gets a custom northbridge that support dual-channel DDR2 memory, they should jump on that immediately. I mean immediately. They don't need to have 32 GB memory for the Mac Pro. 16 GB memory is fine. DDR2 memory should be slightly faster than FB-DIMM for most of Apple's market (content creation and media). Being cheaper is also a big win too.
Intel quad-core seem like the hack job of there duel core thay may be the first to have it but a real quad-core kick it ass and if intel went to HyperTransport thay will not make it better then then a true quad-core.
I am assuming that Intel will move to a more "uniform" architecture once it has CSI or HyperTransport. Clovertown (with two Woodcrests) on the same MCM is an interim solution. Not necessarily a bad strategy since it can be offered with little effort mid-cycle between the Woodcrest and the true Quad core.
I expect AMD, once they have the 65nm process implemented, to come out strong with a Quad core and easily challenge Intel.
Of course, our favourite fruit company would never toy with the rumour mill for their own benefit.
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Am I seriously to believe that Steve Jobs is going to debut an all-new Pro machine with a brand new motherboard that will feature practically ZERO "drive-the-world-forward" technologies? This isn't going to happen. Not in this universe.
It's Think Secret vs. Appleinsider. My money says Steve Jobs is just not a status quo kind of guy.
AMD's going to need more than just a quad core chip and L3 to wrestle back the crown. Current benchmarks show that Core 2 product just isn't a little faster than AMDs current stuff but a LOT faster and it overclocks well.
Having two dual core CPU send some info across the FSB isn't that bad of a situation for now. You have dual independent busses and they don't seem to be taxed and thus limiting in any benchmarks i've seen.
As far as the low end, Apple's got a dilemma with 2S systems. They are expensive and the floor will be around $2.5k. If they can't sell a 2S 2.33 GHz Woodcrest for $2499, there won't be much point because a 2.67 GHz Conroe will be faster than a 2S 2 GHz Woodcrest by a big margin for many many apps.
Looks like no-one's 100% sure if it's X2 or just FX.
Anyways, I'd argue that Quad 2.0 is better than a 2.67 GHz Conroe. There are two markets at issue here: high-end "gaming" desktop, and workstation.
A "gaming" desktop would have a Conroe (most of it's programs are not multi-threaded), and a workstation would try to go Quad (most programs are better about SMP).
Powermacs have mostly been workstations, and the ones priced over $2000 have been for sure. The Mac Pro line doesn't want to sort of awkwardly straddle the two markets. Apple will either ignore the high-end desktop market (likely) or make an "xMac" (not a real name) to cover it. Splitting a product line's purpose is not something Apple seems likely to do.
Also, I just want to re-iterate that the custom northbridge was speculation, based on what Apple could be changing in a motherboard and aiming for the "low-hanging fruit" in terms of saving money.
If Apple gets a custom northbridge that support dual-channel DDR2 memory, they should jump on that immediately. I mean immediately. They don't need to have 32 GB memory for the Mac Pro. 16 GB memory is fine. DDR2 memory should be slightly faster than FB-DIMM for most of Apple's market (content creation and media). Being cheaper is also a big win too.
I'm not sure if Apple would want to give up the high end like that, they've tried to be as close to the cutting edge of workstation computing as they could in the past, even if only a small portion of their user base would fill up that capacity. Chucker mentioned SAS and I think they'll have to at least offer that sort of thing in order to stay in the league to justify their pricing.
Besides, making a custom chipset is not cheap. Maybe nVidia does have something coming out soon for the Xeon chips. Broadcom's ServerWorks chipsets are the primary alternative to Intel chipsets in Xeon workstations, and they have been more expensive yet when they were current. Broadcom appears to be behind at the moment with no chipset for the most recent models.
I suppose the best we can do is just wait and see. Apple is a bit behind at the moment though, so I hope they do pull a rabbit out of their hats, their asses if they have to, at WWDC.
AMD's going to need more than just a quad core chip and L3 to wrestle back the crown. Current benchmarks show that Core 2 product just isn't a little faster than AMDs current stuff but a LOT faster and it overclocks well.
Having two dual core CPU send some info across the FSB isn't that bad of a situation for now. You have dual independent busses and they don't seem to be taxed and thus limiting in any benchmarks i've seen.
Single Clovertown against Quad-Core Opteron will do fine. That's similar to what we have today with 2 Woodcrests vs 2 Dual Core Opterons. But I am not sure if this will still hold true of multiple socket configurations. As we see today for >2 sockets, Opteron is a better choice.
Yeah, but Apple can keep the same mobo for Clovertown and Kentsfield, and all the other Woodcrest-successors for a while. All it'll have to do is bump the FSB once or twice. A custom mobo/chipset would be reused, that R&D wouldn't be spent only on the Rev A Mac Pros.
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Why not? I would be surprised if the changes amounted to anything more significant than what is shown in the article's mock-ups. In transitioning from the PPC units, there have not been a significant change to the aesthetic design of any of the products so far. The iMac and mini had practically no enclosure changes aside from a few connector changes. The Macbook Pros largely retain their appearance aside from fairly minor feature additions. Macbook is a little more significant of a change but it doesn't really depart that much from the previous iBook appearance other than widescreen and the addition of a color option. If the enclosure gets a major change in aesthetic design then it would be a departure.
So Jonathan Ives is getting paid *how much* to move a power socket from the bottom to the top?
I can see the Mac Pro "Design" video - "We thought about it long and hard and considered the history and vitality of the Professional range of Macs through time. The team I led worked dilligently in exploring a huge amount of options and our solution is extremely compelling... Move the power socket to the top, and leave the Cheese Grater in all it's glory the same! We're all very proud of it..."
Yeah, I'm just jealous of anyone that works at Apple. \
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Ex-fracking-actly. WTF.
Come to think of it, Apple has been brilliant in it's budgeting - slash outside redesign, focus on internals (unavoidable) and market the Intel Inside heavily. Quite a turnaround from the days when they were not that loud about PowerPC Inside and instead marketed the "look how beautiful it is" heavily.
Do note that the MacBook Pro's case design is virtually the same as the Aluminum PowerBook's, which is from January 2003, so the Mac Pro enclosure wouldn't exactly be the oldest design.
IMO since the PowerBook G3 to Titanium G4 there has not been astonishing developments in the design of the PowerBook. But then again I'm biased because the Titanium G4 PowerBook was my first PowerBook. It was fracking cool. The balance of edge and curve was just puurfect. The Aluminium PowerBooks/ MacBookPros are too "curvy" to really say, this is *professional* stuff don't f*** with me..!!11!!1one
Single Clovertown against Quad-Core Opteron will do fine. That's similar to what we have today with 2 Woodcrests vs 2 Dual Core Opterons. But I am not sure if this will still hold true of multiple socket configurations. As we see today for >2 sockets, Opteron is a better choice.
The Opteron is better in the context of memory bandwidth. And that was going up against netburst architecture. Beating a 4s Woodcrest system is going to be a lot harder. Right from the jump you're giving up IPC because for the foreseeable future the K8 looks to be keeping its 3-issue core vs 4-issue in Core 2. Benchmarks today bear this out as lower clocked Conroe are creaming AMDs clock.
Should be fun to watch. We'll know soon enough what AMD has to do to beat some beefy Woodcrest setups.
No, it's a terrible idea. This "Mac" would not only be a hugely confusing brand name ("What do you have?" ?_"A Mac" ?_"Oh, you mean like, the platform? Or the model? Or??"); it would also severely eat into the iMac's sales, because it would inevitably be priced at the exact same levels. Apple's mid-range desktop is an all-in-one. Deal with it.
I agree that it needs a better name than just Mac. I haven't heard any suggestions I really like. "Mac Medium"? Maybe follow the software naming and go with "Mac Express"? But it would be a GREAT addition to the line. I don't buy "deal with it" as an argument...it's the same thing people said before the mini was announced, and look how wrong they were.
Who cares if it eats into iMac sales, as long as it's just as profitable and increases mac sales overall?
Quote:
Originally posted by Chucker
Well, for one, they can sell the entire thing (the iMac) at a higher prices than they could a headless, because there's less competition in that segment, so consumers will have less comparison, so they'll have less of a chance of pointing out (and/or choosing) an alternative product that's more affordable ?_there likely isn't any. (In fact, the same is the case for the Mac mini. The only halfway serious Mac mini competitor is the creatively named MiniPC, which is actually more expensive and has lower specs.)
But more importantly, an AIO provides for higher revenues because it is purchased more frequently. Hard drive, screen, optical drive or CPU not good enough for the customer any more? Rather than upgrade one or multiple components, they're more likely to simply get a complete new computer.
Apple chooses their prices and their margins. A midtower may not compare as well to a dell, but it will still make some mac buyers very happy. I don't think more frequent purchases are much of a factor, most people are ready for a new monitor at about the same time they're ready for a new computer. Look at all the PC bundles where people already have a monitor but get a new one anyway just because prices have dropped so much since their last purchase.
Quote:
Originally posted by Apparatus
Power Supply on top makes no sense.
Why not? What's the problem?
Quote:
Originally posted by ZachPruckowski
If a Mac Pro at $2000 gets you a 2.0 GHz Quad, it'll beat a 2.4GHz 4x4 (based on Conroe vs. FX benchmarks) for the same price. It'll be the cheapest Quad on the workstation market (by a mile), hurting Dell (who has 30-40 percent workstation mark-ups to cover for single-digit margins on $400 boxes).
If a Mac Pro at $2000 gets you a 2.66 GHz Conroe, it'll look over-priced next to the "same-spec" XPS at $1500-1600 (which'll also have SLI or Crossfire BTO). It'll also mean a different Motherboard (which is a SKU issue for Apple, a relatively low-volume company).
If a Mac Pro at $2000 gets you any sort of single Woodcrest, it'll be underperforming compared to XPS and 4x4 (2.0 GHz or 2.33 GHz WC) or it'll be more expensive than a 2.0 GHz Quad (2.66 GHz WC).
I don't think apple can do quad for $2K. But I think they can do conroe for $1500, maybe even less if they go minitower on the low end. Apple needs speed on the high end. On the low end, price is way more important than speed - people will complain more about a base model being too expensive than it not being fast enough.
Comments
Originally posted by Joe_the_dragon
Intel quad-core will end up the same way or worse then the first Intel dual cores. Linked to each other by FSB where AMD is useing true quad-cores with Shared L3.
You're thinking of quad-core as 4 cores per processor. What I'm talking about is 2 dual-core chips. That'd function identically to how the Dual-G4 Powermacs functioned. There are two sockets, each connected to the north and south bridges by their own FSBs.
You are correct that Intel's Clovertown processor will be two Woodcrests connected by FSB in the same package. But the next generation (mid-2007 or late 2007) will be "true quad-core".
When I'm talking about what AMD is planning, they're planning something that involves 2 dual-core X2s or FXs. That is not "true quad-core". It's a dual-dual setup. Except with desktop chips, not server/workstation chips.
Originally posted by Joe_the_dragon
Intel quad-core will end up the same way or worse then the first Intel duel cores. Linked to each other by FSB where AMD is useing true quad-cores with Shared L3.
No doubt, AMD's HyperTransport is a better architecture. Intel Core 2 Duos and Woodcrest do quite well as the 2 cores share the cache without involving the FSB. All the performance numbers in the 1 and 2 socket are in Intel's favor. Socket 4 and up is a different story. Intel's Tulsa (not Core 2 arch) will not stand upto AMD Opteron even with its large L3 cache. Clovertown may be the same as its just two Woodcrests on the same packaging. AMD might be able to wrestle back the performance crown again. The 65nm process may not work for Intel competitively this time as AMD will have moved over to it as well for the Quad core chip. The only solution for Intel is to either develop the Common System Interconnect soon or (if possible) just license the HyperTransport.
Originally posted by mwswami
No doubt, AMD's HyperTransport is a better architecture. Intel Core 2 Duos and Woodcrest do quite well as the 2 cores share the cache without involving the FSB. All the performance numbers in the 1 and 2 socket are in Intel's favor. Socket 4 and up is a different story. Intel's Tulsa (not Core 2 arch) will not stand upto AMD Opteron even with its large L3 cache. Clovertown may be the same as its just two Woodcrests on the same packaging. AMD might be able to wrestle back the performance crown again. The 65nm process may not work for Intel competitively this time as AMD will have moved over to it as well for the Quad core chip. The only solution for Intel is to either develop the Common System Interconnect soon or (if possible) just license the HyperTransport.
Intel quad-core seem like the hack job of there duel core thay may be the first to have it but a real quad-core kick it ass and if intel went to HyperTransport thay will not make it better then then a true quad-core.
HyperTransport based accelerators may force intel to license HyperTransport and thay sound like a cool think that apple may want.
Originally posted by ZachPruckowski
Here's the real reason the low-end Mac Pros will be Quad: AMD's 4x4 initiative. In theory, AMD might cut low-end X2 or even FX prices by 1/3 to 1/2, which'll leave you with some bargain prices. They are then releasing a motherboard that can have two X2s or FXs (these are Athlons, not Opterons) on a prosumer desktop board.
You'll have to provide a link for AMD's 4x4 being able to use an Athlon X2. As far as I know, it's an "enthusiast" 2S platform using Athlon FX CPUs. Enthusiast doesn't equate to being affordable, and a 4x4 system will likely be in the $2500 to $3000 range.
As far as the low end, Apple's got a dilemma with 2S systems. They are expensive and the floor will be around $2.5k. If they can't sell a 2S 2.33 GHz Woodcrest for $2499, there won't be much point because a 2.67 GHz Conroe will be faster than a 2S 2 GHz Woodcrest by a big margin for many many apps.
The lineup has to be a variant of 1x2.67, 2x2.67, 2x3 or 1x2.67, 1x2.93, 2x3 or perhaps 2x2.33, 2x2.67 or 2x3. An all 2S platform will be very very pricey. If they go all 2S in the Mac Pro, Apple will be forced into developing a Mac Cube to fill the gigantic hole between $1500 and $2500.
Additionally, Dell/everyone will have 2.4 or 2.67 GHz Conroes shipping for $1500.
Apple has had no entry in the mid-range headless desktop market for years now. I'd guess that they would be happy with selling Mac minis and iMacs and leave a gigantic gulf between them and the Mac Pros.
Originally posted by ZachPruckowski
1) In theory, Apple could (if they wanted to) hack together a "normal" memory controller for the Mac Pros and leave the FB-DIMMs to the XServe. It'd be annoying, but cheaper and not killer performance-wise.
If Apple gets a custom northbridge that support dual-channel DDR2 memory, they should jump on that immediately. I mean immediately. They don't need to have 32 GB memory for the Mac Pro. 16 GB memory is fine. DDR2 memory should be slightly faster than FB-DIMM for most of Apple's market (content creation and media). Being cheaper is also a big win too.
Originally posted by Joe_the_dragon
Intel quad-core seem like the hack job of there duel core thay may be the first to have it but a real quad-core kick it ass and if intel went to HyperTransport thay will not make it better then then a true quad-core.
I am assuming that Intel will move to a more "uniform" architecture once it has CSI or HyperTransport. Clovertown (with two Woodcrests) on the same MCM is an interim solution. Not necessarily a bad strategy since it can be offered with little effort mid-cycle between the Woodcrest and the true Quad core.
I expect AMD, once they have the 65nm process implemented, to come out strong with a Quad core and easily challenge Intel.
Of course, our favourite fruit company would never toy with the rumour mill for their own benefit.
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Am I seriously to believe that Steve Jobs is going to debut an all-new Pro machine with a brand new motherboard that will feature practically ZERO "drive-the-world-forward" technologies? This isn't going to happen. Not in this universe.
It's Think Secret vs. Appleinsider. My money says Steve Jobs is just not a status quo kind of guy.
Having two dual core CPU send some info across the FSB isn't that bad of a situation for now. You have dual independent busses and they don't seem to be taxed and thus limiting in any benchmarks i've seen.
Originally posted by THT
As far as the low end, Apple's got a dilemma with 2S systems. They are expensive and the floor will be around $2.5k. If they can't sell a 2S 2.33 GHz Woodcrest for $2499, there won't be much point because a 2.67 GHz Conroe will be faster than a 2S 2 GHz Woodcrest by a big margin for many many apps.
Link 1 for X2 4x4 Link 2
Looks like no-one's 100% sure if it's X2 or just FX.
Anyways, I'd argue that Quad 2.0 is better than a 2.67 GHz Conroe. There are two markets at issue here: high-end "gaming" desktop, and workstation.
A "gaming" desktop would have a Conroe (most of it's programs are not multi-threaded), and a workstation would try to go Quad (most programs are better about SMP).
Powermacs have mostly been workstations, and the ones priced over $2000 have been for sure. The Mac Pro line doesn't want to sort of awkwardly straddle the two markets. Apple will either ignore the high-end desktop market (likely) or make an "xMac" (not a real name) to cover it. Splitting a product line's purpose is not something Apple seems likely to do.
Also, I just want to re-iterate that the custom northbridge was speculation, based on what Apple could be changing in a motherboard and aiming for the "low-hanging fruit" in terms of saving money.
Originally posted by THT
If Apple gets a custom northbridge that support dual-channel DDR2 memory, they should jump on that immediately. I mean immediately. They don't need to have 32 GB memory for the Mac Pro. 16 GB memory is fine. DDR2 memory should be slightly faster than FB-DIMM for most of Apple's market (content creation and media). Being cheaper is also a big win too.
I'm not sure if Apple would want to give up the high end like that, they've tried to be as close to the cutting edge of workstation computing as they could in the past, even if only a small portion of their user base would fill up that capacity. Chucker mentioned SAS and I think they'll have to at least offer that sort of thing in order to stay in the league to justify their pricing.
Besides, making a custom chipset is not cheap. Maybe nVidia does have something coming out soon for the Xeon chips. Broadcom's ServerWorks chipsets are the primary alternative to Intel chipsets in Xeon workstations, and they have been more expensive yet when they were current. Broadcom appears to be behind at the moment with no chipset for the most recent models.
I suppose the best we can do is just wait and see. Apple is a bit behind at the moment though, so I hope they do pull a rabbit out of their hats, their asses if they have to, at WWDC.
Originally posted by hmurchison
AMD's going to need more than just a quad core chip and L3 to wrestle back the crown. Current benchmarks show that Core 2 product just isn't a little faster than AMDs current stuff but a LOT faster and it overclocks well.
Having two dual core CPU send some info across the FSB isn't that bad of a situation for now. You have dual independent busses and they don't seem to be taxed and thus limiting in any benchmarks i've seen.
Single Clovertown against Quad-Core Opteron will do fine. That's similar to what we have today with 2 Woodcrests vs 2 Dual Core Opterons. But I am not sure if this will still hold true of multiple socket configurations. As we see today for >2 sockets, Opteron is a better choice.
Originally posted by JeffDM
Besides, making a custom chipset is not cheap.
Yeah, but Apple can keep the same mobo for Clovertown and Kentsfield, and all the other Woodcrest-successors for a while. All it'll have to do is bump the FSB once or twice. A custom mobo/chipset would be reused, that R&D wouldn't be spent only on the Rev A Mac Pros.
Originally posted by Frank777
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Why not? I would be surprised if the changes amounted to anything more significant than what is shown in the article's mock-ups. In transitioning from the PPC units, there have not been a significant change to the aesthetic design of any of the products so far. The iMac and mini had practically no enclosure changes aside from a few connector changes. The Macbook Pros largely retain their appearance aside from fairly minor feature additions. Macbook is a little more significant of a change but it doesn't really depart that much from the previous iBook appearance other than widescreen and the addition of a color option. If the enclosure gets a major change in aesthetic design then it would be a departure.
I can see the Mac Pro "Design" video - "We thought about it long and hard and considered the history and vitality of the Professional range of Macs through time. The team I led worked dilligently in exploring a huge amount of options and our solution is extremely compelling... Move the power socket to the top, and leave the Cheese Grater in all it's glory the same! We're all very proud of it..."
Yeah, I'm just jealous of anyone that works at Apple.
No new case? That's unbelievable. Intel's done a lot of work on the guts of this machine. What has Apple's team been doing since October of 2005?, vacationing in Hawaii?
Ex-fracking-actly. WTF.
Come to think of it, Apple has been brilliant in it's budgeting - slash outside redesign, focus on internals (unavoidable) and market the Intel Inside heavily. Quite a turnaround from the days when they were not that loud about PowerPC Inside and instead marketed the "look how beautiful it is" heavily.
According to this article the Woodcrest/Xeon 5100 series will use FB-DIMMs, not DDR2: http://arstechnica.com/articles/paed.../promacs.ars/2
Furthermore the prices cannot be correct since the processors alone would account for 1380,- US$: http://arstechnica.com/articles/paed.../promacs.ars/4
Originally posted by mwswami
Single Clovertown against Quad-Core Opteron will do fine. That's similar to what we have today with 2 Woodcrests vs 2 Dual Core Opterons. But I am not sure if this will still hold true of multiple socket configurations. As we see today for >2 sockets, Opteron is a better choice.
The Opteron is better in the context of memory bandwidth. And that was going up against netburst architecture. Beating a 4s Woodcrest system is going to be a lot harder. Right from the jump you're giving up IPC because for the foreseeable future the K8 looks to be keeping its 3-issue core vs 4-issue in Core 2. Benchmarks today bear this out as lower clocked Conroe are creaming AMDs clock.
Should be fun to watch. We'll know soon enough what AMD has to do to beat some beefy Woodcrest setups.
Originally posted by Chucker
No, it's a terrible idea. This "Mac" would not only be a hugely confusing brand name ("What do you have?" ?_"A Mac" ?_"Oh, you mean like, the platform? Or the model? Or??"); it would also severely eat into the iMac's sales, because it would inevitably be priced at the exact same levels. Apple's mid-range desktop is an all-in-one. Deal with it.
I agree that it needs a better name than just Mac. I haven't heard any suggestions I really like. "Mac Medium"? Maybe follow the software naming and go with "Mac Express"? But it would be a GREAT addition to the line. I don't buy "deal with it" as an argument...it's the same thing people said before the mini was announced, and look how wrong they were.
Who cares if it eats into iMac sales, as long as it's just as profitable and increases mac sales overall?
Originally posted by Chucker
Well, for one, they can sell the entire thing (the iMac) at a higher prices than they could a headless, because there's less competition in that segment, so consumers will have less comparison, so they'll have less of a chance of pointing out (and/or choosing) an alternative product that's more affordable ?_there likely isn't any. (In fact, the same is the case for the Mac mini. The only halfway serious Mac mini competitor is the creatively named MiniPC, which is actually more expensive and has lower specs.)
But more importantly, an AIO provides for higher revenues because it is purchased more frequently. Hard drive, screen, optical drive or CPU not good enough for the customer any more? Rather than upgrade one or multiple components, they're more likely to simply get a complete new computer.
Apple chooses their prices and their margins. A midtower may not compare as well to a dell, but it will still make some mac buyers very happy. I don't think more frequent purchases are much of a factor, most people are ready for a new monitor at about the same time they're ready for a new computer. Look at all the PC bundles where people already have a monitor but get a new one anyway just because prices have dropped so much since their last purchase.
Originally posted by Apparatus
Power Supply on top makes no sense.
Why not? What's the problem?
Originally posted by ZachPruckowski
If a Mac Pro at $2000 gets you a 2.0 GHz Quad, it'll beat a 2.4GHz 4x4 (based on Conroe vs. FX benchmarks) for the same price. It'll be the cheapest Quad on the workstation market (by a mile), hurting Dell (who has 30-40 percent workstation mark-ups to cover for single-digit margins on $400 boxes).
If a Mac Pro at $2000 gets you a 2.66 GHz Conroe, it'll look over-priced next to the "same-spec" XPS at $1500-1600 (which'll also have SLI or Crossfire BTO). It'll also mean a different Motherboard (which is a SKU issue for Apple, a relatively low-volume company).
If a Mac Pro at $2000 gets you any sort of single Woodcrest, it'll be underperforming compared to XPS and 4x4 (2.0 GHz or 2.33 GHz WC) or it'll be more expensive than a 2.0 GHz Quad (2.66 GHz WC).
I don't think apple can do quad for $2K. But I think they can do conroe for $1500, maybe even less if they go minitower on the low end. Apple needs speed on the high end. On the low end, price is way more important than speed - people will complain more about a base model being too expensive than it not being fast enough.