Apple needs a $1500 singe duel core with good video system to compete with dell and others.
Amd 4X4 can use ddr2 non ecc and SLI with NVIDIA chip set / cross fire with a ATI one.
And when quad-core core come out amd 4x4 will let you have 2 of them and they will not be choked up like Intel ones that are 2 duel cores on one cpu linked by a FSB.
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.
2) It looks like you're an AMD fan. Like it or not, Conroe dominates an X2 or FX one-on-one. But two X2s could (probably) beat a single Woodcrest (which is basically a Conroe at heart) at 2.0 or 2.33 GHz. Thus my point is that Apple wants to position the Mac Pros as competing with Precision workstations (which they'll destroy on price) and 4x4 (which they'll destroy performance-wise, for now), not XPS & co. on the low end (which'll pretty much rule the single-socket roost).
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 think I just covered all the options for the low-end Mac Pro, and based on all the Conroe and Woodcrest chips, the only set-up that my analysis (which may be flawed) shows as being practical and impressive spec-wise is 2.0 Quad. It has a margin of about 15% at that price, but my analysis works for any price in the ballpark of $1800-$2200.
Note that for the purpose of this discussion, I've been defining "low-end Mac Pro" as the replacement for the current low-end Powermac ($2000), not a possible $1400-1800 "Mac Cube"/"xMac"/"Mac".
.......That's not a complete solution, and it nearly triples the cost of the system (several thousand for XSR vs. sub-$1k for several drives). XServeRAID is enterprise-class, not workstation-class, leaving a major cost and storage gap in between two drives and seven.
Cool. But so if we are looking at 3-5 drives some of the drives HAVE to be external for backup purposes. If you dump all 3-5 drives in one workstation what are you going to backup to? Unless you choose Option 2 below:
Option 1
We have 2 drives on RAID 0 in the workstation and then 2-3 drives as external backup, possibly in RAID 0, RAID 1, RAID 0+1, RAID 5, etc. configuration. The only downside to this is the FW400 connection out to the external drives/ RAID. Which is where eSATA would come in very handy.
Option 2
In your scenario fair enough, you dump 3-5 drives in one workstation and run a decent RAID on it. But as you mention, this requires an OS update to handle it well (maybe Leopard), but better still a dedicated hardware RAID controller on the chipset (eg shipping nVidia chipsets in PC land). I don't know in OS X now how it handles rebuilding the RAID if you're running RAID 1 and one of your two drives goes down. Supporting RAID 5 via OSX or OSX+chipset would also be interesting to see how setting up and rebuilding the raid and reporting lost data if any, will go. It is interesting to note that RAID 5 can have slow write speeds (http://www.xbitlabs.com/articles/chi...-intel_11.html)
Re: Chucker and others: "workstation" above can mean "desktop" as well
Cool. But so if we are looking at 3-5 drives some of the drives HAVE to be external for backup purposes. If you dump all 3-5 drives in one workstation what are you going to backup to? Unless you choose Option 2 below:
-snip-
My planned method is to have a stripe RAID internally, with an occasional back-up to a different computer. The system partitions is and will be backed up using SuperDuper to an external FW400 drive as a cold backup, only powered during backup. If something happens, I could turn the external on and boot from it. The data partition is backed up to a hardware RAID-5 in a Compaq W8000.
My planned method is to have a stripe RAID internally, with an occasional back-up to a different computer. The system partitions is and will be backed up using SuperDuper to an external FW400 drive as a cold backup, only powered during backup. If something happens, I could turn the external on and boot from it. The data partition is backed up to a hardware RAID-5 in a Compaq W8000.
So your stripe RAID internally would be just 2 drives? Or would you like 3 drives?
So your stripe RAID internally would be just 2 drives? Or would you like 3 drives? [/B]
I would like as many three drives on the stripe volume, and a fourth drive that's a system/boot drive. Whether that gets done depends on whether I care to go through the effort of making a hard drive mount bracket. It's not too much work, but I'm sometimes lazy like that.
But the Mac Pro may only support 2 hd sata drive bays. \ So anyway, no worries, thanks for the discussion. I was just trying to see why people were clamouring for 3-4 hd sata drive bays.
I think I am starting to understand, about the primary storage needs for the "workstation-class" computer. I am assuming that in most cases you would need a 1:1 backup solution. That means the bigger your primary storage, the backup solution needs to be equally big. But some people don't need a 1:1, they need primary boot/system drive/partition backed up 1:1 and then a data partition backup for finalised files only, not *all* the files used during the work.
Heh. Maybe it's just fun. Run a RAID 0 off four HITACHI Deskstar 500gb 7K500's for 2TB of internal storage Whereas two drive bays in the Mac Pro would only give you 1TB of internal storage with these drives.
Okay, cool, not being sarcastic here, just delving into the world of HDs, RAID, Backup solutions and the big question mark over the specs of the new Mac Pros. I think we'll be kept guessing till the very end... Hurry up and start WWDC already!!!
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
Wow, that's like a whole 'nother thread in and of itself. Let's start with: What do you do with your Mac? Do you derive income from it? What would you like to do with your Mac that you're not doing now? What software do you use mostly? What Mac do you own currently?
Your answers will be useful in answering your question
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
It would probably be faster, simply because the Core architecture will beat out the G5 architecture. Remember back in January, when the first Yonah iMacs came out? They were competitive with a low-end PowerMac (dual-2.0 GHz vs. dual-2.0 GHz) at the time. And the Core architecture chips (Merom/Conroe/Woodcrest) are supposed to be 20% to 40% faster than Yonahs clock-for-clock (according to Intel's numbers, which I can't find ATM). Therefore, a Woodcrest should beat a G5 clock-for-clock.
Is it worth $1000-$1500 (once you eBay your G5 Quad)? If you have the money, and if you use mostly UB apps, and if the Woodcrest Quads come with nice other features (Airport, BT, etc), it'd prolly be worth it, if you need the performance.
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
Wow, that's like a whole 'nother thread in and of itself. Let's start with: What do you do with your Mac? Do you derive income from it? What would you like to do with your Mac that you're not doing now? What software do you use mostly? What Mac do you own currently?
Your answers will be useful in answering your question
To cost-justify it, I'd suggest that you'd definitely need to derive significant benefit from time savings on compute-intensive tasks, assuming those tasks are a limiting factor in your productivity. Or be able to derive a time or productivity benefit from having Windows readily available for the occasional program or device that doesn't have an adequate (or affordable) Mac equivalent.
The thing should either be able to save you a lot of work time or increase productivity sufficiently for the maybe extra $1k (assumptions here, but it's a nice, round number) it might take to sell the old unit and buy a new one.
I don't have a quad, but I might get some form of Mac Pro for my work desk so I can add OS X apps and capabilities. I can't do without Windows because it will be running a fairly expensive (to me) type of commercial machinery that's not compatible with OS X, but I'm hoping will run within Parallels. If it weren't for Parallels, a Mac Pro wouldn't even be considered.
Or if you just want it, then there's no reasoning about that.
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
I'd hazard a guess that a 2S 2.66 GHz Woodcrest would be on average the equivalent of a 3.2 GHz G5 Quad assuming native apps and such.
Hands-down, Woodcrest will dominate in integer apps, possibly by as much as 2 times, with an integer performance/watt advantage that could be 3x to 4x better than a 2.5 GHz G5 Quad.
For floating point apps, it won't be such a huge benifit. Maybe 25% to 40% faster.
The Mac Pro lineup will have to be something like:
1S 2.66 GHz Conroe for $1999
2S 2.66 GHz Woodcrest for $2799
2S 3.00 GHz Woodcrest for $3599
I could buy into either a Conroe Extreme in the mid-range, and at lower odds, a 2S 2.33 GHz Woodcrest on the low end for something like $2399 or $2499 which would push the mid-range price to $2999
If Apple comes out with a headless Conroe system in the $1000 to $200 range with a modicum of upgradeability (1 PCIe x16 slot, 1 PCIe x2 slot, 1 external bay, 1 internal bay, 2 memory slots), it'll great for buyers who don't need all the pro features in the Pro desktop, but I'm fairly convinced that Apple thinks this will cannibalize "Mac Pro" sales producing a net negative on revenues.
But that's not the only system being discussed. Mac Pro isn't a consumer computer, and will likely use a 5000 series Intel chipset. Dell is claiming they have SLI on their Precision 690, which they say is running the 5000X chipset.
Personally, I'd prefer it if the Mac Pro came with 4 HD bays. That way, 4 identical drives on two SATA2 buses could be striped to RAID level 50. That could be a nice and fast 2TB array which would be dandy for HD video editing.
And by all means, let's have the Greencreek memory controller onboard with FB-DIMMs.
also i was saying that you can run SLI on an intel chipset (5000X), but there is no evidence saying that Apple is going to use an intel chipset... they could very easily use an nvidia.
I know it works/worked on certain ULI-chipsets, and now apparently on the Blackford too, so it certainly can be done, but I for one would not run uncertified and unsupported hardware on mission critical systems.
But that's not the only system being discussed. Mac Pro isn't a consumer computer, and will likely use a 5000 series Intel chipset. Dell is claiming they have SLI on their Precision 690, which they say is running the 5000X chipset.
I configured a Dell 690 and saw the SLI as an option as well. Since the 690 has the 5000X chipset (which is the only chipset that I know that supports the Woodcrest) that means that Dell is saying that it supports SLI????
I posted a link earlier to a game site review of a Tyan motherboard with the 5000X chipset. Tyan was also claiming to support SLI but the game website found that the Nvidia software gave an error message and did not actually support interlacing. I wonder if Dell has gotten 'special permission'/ programming from Nvidia to advertise SLI on a non-Nvidia chipset???
I wonder if Dell has gotten 'special permission'/ programming from Nvidia to advertise SLI on a non-Nvidia chipset???
I don't know, I'd like to see confirmation of it as well. I don't understand why Tyan is promoting the feature if it didn't work. Given that the cards in question that Dell says is configured for SLI ($2700 for the pair of them, more than an entire base 690 w/ 1kW PSU), I would think that it's in nVidia's best interests to support it regardless of whose chipset is used.
I could have sworn that there was an nVidia PCIe bridge chip that is added to a base PCIe chipset that enables two of their boards to operate in SLI mode, but I can't find that chip right now. It might be that this chip is used in conjunction with the 5000X.
All this is speculation though, especially as to whether Apple would support SLI assuming the chipsets they get from Intel are capable of running it.
While it's possible that the power connector could still be on the bottom (with a cable inside the case), when it came to drawing the mockups I decided to put the connector on top. The information received only specifed that the power supply was relocated to the top of the case.
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.
2) It looks like you're an AMD fan. Like it or not, Conroe dominates an X2 or FX one-on-one. But two X2s could (probably) beat a single Woodcrest (which is basically a Conroe at heart) at 2.0 or 2.33 GHz. Thus my point is that Apple wants to position the Mac Pros as competing with Precision workstations (which they'll destroy on price) and 4x4 (which they'll destroy performance-wise, for now), not XPS & co. on the low end (which'll pretty much rule the single-socket roost).
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.
Also intel new chips sets have raid bug even with a raid card.
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.
Also intel new chips sets have raid bug even with a raid card.
Comments
Originally posted by Joe_the_dragon
Woodcrest needs FB-DIMMs
Apple needs a $1500 singe duel core with good video system to compete with dell and others.
Amd 4X4 can use ddr2 non ecc and SLI with NVIDIA chip set / cross fire with a ATI one.
And when quad-core core come out amd 4x4 will let you have 2 of them and they will not be choked up like Intel ones that are 2 duel cores on one cpu linked by a FSB.
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.
2) It looks like you're an AMD fan. Like it or not, Conroe dominates an X2 or FX one-on-one. But two X2s could (probably) beat a single Woodcrest (which is basically a Conroe at heart) at 2.0 or 2.33 GHz. Thus my point is that Apple wants to position the Mac Pros as competing with Precision workstations (which they'll destroy on price) and 4x4 (which they'll destroy performance-wise, for now), not XPS & co. on the low end (which'll pretty much rule the single-socket roost).
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 think I just covered all the options for the low-end Mac Pro, and based on all the Conroe and Woodcrest chips, the only set-up that my analysis (which may be flawed) shows as being practical and impressive spec-wise is 2.0 Quad. It has a margin of about 15% at that price, but my analysis works for any price in the ballpark of $1800-$2200.
Note that for the purpose of this discussion, I've been defining "low-end Mac Pro" as the replacement for the current low-end Powermac ($2000), not a possible $1400-1800 "Mac Cube"/"xMac"/"Mac".
.......That's not a complete solution, and it nearly triples the cost of the system (several thousand for XSR vs. sub-$1k for several drives). XServeRAID is enterprise-class, not workstation-class, leaving a major cost and storage gap in between two drives and seven.
Cool. But so if we are looking at 3-5 drives some of the drives HAVE to be external for backup purposes. If you dump all 3-5 drives in one workstation what are you going to backup to? Unless you choose Option 2 below:
Option 1
We have 2 drives on RAID 0 in the workstation and then 2-3 drives as external backup, possibly in RAID 0, RAID 1, RAID 0+1, RAID 5, etc. configuration. The only downside to this is the FW400 connection out to the external drives/ RAID. Which is where eSATA would come in very handy.
Option 2
In your scenario fair enough, you dump 3-5 drives in one workstation and run a decent RAID on it. But as you mention, this requires an OS update to handle it well (maybe Leopard), but better still a dedicated hardware RAID controller on the chipset (eg shipping nVidia chipsets in PC land). I don't know in OS X now how it handles rebuilding the RAID if you're running RAID 1 and one of your two drives goes down. Supporting RAID 5 via OSX or OSX+chipset would also be interesting to see how setting up and rebuilding the raid and reporting lost data if any, will go. It is interesting to note that RAID 5 can have slow write speeds (http://www.xbitlabs.com/articles/chi...-intel_11.html)
Re: Chucker and others: "workstation" above can mean "desktop" as well
Quote:
Cool. But so if we are looking at 3-5 drives some of the drives HAVE to be external for backup purposes. If you dump all 3-5 drives in one workstation what are you going to backup to? Unless you choose Option 2 below:
-snip-
My planned method is to have a stripe RAID internally, with an occasional back-up to a different computer. The system partitions is and will be backed up using SuperDuper to an external FW400 drive as a cold backup, only powered during backup. If something happens, I could turn the external on and boot from it. The data partition is backed up to a hardware RAID-5 in a Compaq W8000.
My planned method is to have a stripe RAID internally, with an occasional back-up to a different computer. The system partitions is and will be backed up using SuperDuper to an external FW400 drive as a cold backup, only powered during backup. If something happens, I could turn the external on and boot from it. The data partition is backed up to a hardware RAID-5 in a Compaq W8000.
So your stripe RAID internally would be just 2 drives?
Originally posted by sunilraman
So your stripe RAID internally would be just 2 drives?
I would like as many three drives on the stripe volume, and a fourth drive that's a system/boot drive. Whether that gets done depends on whether I care to go through the effort of making a hard drive mount bracket. It's not too much work, but I'm sometimes lazy like that.
I think I am starting to understand, about the primary storage needs for the "workstation-class" computer. I am assuming that in most cases you would need a 1:1 backup solution. That means the bigger your primary storage, the backup solution needs to be equally big. But some people don't need a 1:1, they need primary boot/system drive/partition backed up 1:1 and then a data partition backup for finalised files only, not *all* the files used during the work.
Heh. Maybe it's just fun. Run a RAID 0 off four HITACHI Deskstar 500gb 7K500's for 2TB of internal storage
Okay, cool, not being sarcastic here, just delving into the world of HDs, RAID, Backup solutions and the big question mark over the specs of the new Mac Pros. I think we'll be kept guessing till the very end... Hurry up and start WWDC already!!!
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
Wow, that's like a whole 'nother thread in and of itself. Let's start with: What do you do with your Mac? Do you derive income from it? What would you like to do with your Mac that you're not doing now? What software do you use mostly? What Mac do you own currently?
Your answers will be useful in answering your question
Originally posted by ascii
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
It would probably be faster, simply because the Core architecture will beat out the G5 architecture. Remember back in January, when the first Yonah iMacs came out? They were competitive with a low-end PowerMac (dual-2.0 GHz vs. dual-2.0 GHz) at the time. And the Core architecture chips (Merom/Conroe/Woodcrest) are supposed to be 20% to 40% faster than Yonahs clock-for-clock (according to Intel's numbers, which I can't find ATM). Therefore, a Woodcrest should beat a G5 clock-for-clock.
Is it worth $1000-$1500 (once you eBay your G5 Quad)? If you have the money, and if you use mostly UB apps, and if the Woodcrest Quads come with nice other features (Airport, BT, etc), it'd prolly be worth it, if you need the performance.
Quote:
Originally posted by ascii
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
Wow, that's like a whole 'nother thread in and of itself. Let's start with: What do you do with your Mac? Do you derive income from it? What would you like to do with your Mac that you're not doing now? What software do you use mostly? What Mac do you own currently?
Your answers will be useful in answering your question
To cost-justify it, I'd suggest that you'd definitely need to derive significant benefit from time savings on compute-intensive tasks, assuming those tasks are a limiting factor in your productivity. Or be able to derive a time or productivity benefit from having Windows readily available for the occasional program or device that doesn't have an adequate (or affordable) Mac equivalent.
The thing should either be able to save you a lot of work time or increase productivity sufficiently for the maybe extra $1k (assumptions here, but it's a nice, round number) it might take to sell the old unit and buy a new one.
I don't have a quad, but I might get some form of Mac Pro for my work desk so I can add OS X apps and capabilities. I can't do without Windows because it will be running a fairly expensive (to me) type of commercial machinery that's not compatible with OS X, but I'm hoping will run within Parallels. If it weren't for Parallels, a Mac Pro wouldn't even be considered.
Or if you just want it, then there's no reasoning about that.
Originally posted by ascii
Would a quad 2.66 be much faster than a quad 2.5 G5? And with the same case, what would be the reason to upgrade?
I'd hazard a guess that a 2S 2.66 GHz Woodcrest would be on average the equivalent of a 3.2 GHz G5 Quad assuming native apps and such.
Hands-down, Woodcrest will dominate in integer apps, possibly by as much as 2 times, with an integer performance/watt advantage that could be 3x to 4x better than a 2.5 GHz G5 Quad.
For floating point apps, it won't be such a huge benifit. Maybe 25% to 40% faster.
The Mac Pro lineup will have to be something like:
1S 2.66 GHz Conroe for $1999
2S 2.66 GHz Woodcrest for $2799
2S 3.00 GHz Woodcrest for $3599
I could buy into either a Conroe Extreme in the mid-range, and at lower odds, a 2S 2.33 GHz Woodcrest on the low end for something like $2399 or $2499 which would push the mid-range price to $2999
If Apple comes out with a headless Conroe system in the $1000 to $200 range with a modicum of upgradeability (1 PCIe x16 slot, 1 PCIe x2 slot, 1 external bay, 1 internal bay, 2 memory slots), it'll great for buyers who don't need all the pro features in the Pro desktop, but I'm fairly convinced that Apple thinks this will cannibalize "Mac Pro" sales producing a net negative on revenues.
Originally posted by demenas
See this:
http://www.dell.com/content/products...=19&l=en&s=dhs
"The ultimate gaming solution includes the optional NVIDIA® Quad SLI..."
Steve
Via the nForce 590 chipset.
Originally posted by BenRoethig
Via the nForce 590 chipset.
But that's not the only system being discussed. Mac Pro isn't a consumer computer, and will likely use a 5000 series Intel chipset. Dell is claiming they have SLI on their Precision 690, which they say is running the 5000X chipset.
And by all means, let's have the Greencreek memory controller onboard with FB-DIMMs.
Originally posted by doh123
also i was saying that you can run SLI on an intel chipset (5000X), but there is no evidence saying that Apple is going to use an intel chipset... they could very easily use an nvidia.
I know it works/worked on certain ULI-chipsets, and now apparently on the Blackford too, so it certainly can be done, but I for one would not run uncertified and unsupported hardware on mission critical systems.
Originally posted by JeffDM
But that's not the only system being discussed. Mac Pro isn't a consumer computer, and will likely use a 5000 series Intel chipset. Dell is claiming they have SLI on their Precision 690, which they say is running the 5000X chipset.
I configured a Dell 690 and saw the SLI as an option as well. Since the 690 has the 5000X chipset (which is the only chipset that I know that supports the Woodcrest) that means that Dell is saying that it supports SLI????
I posted a link earlier to a game site review of a Tyan motherboard with the 5000X chipset. Tyan was also claiming to support SLI but the game website found that the Nvidia software gave an error message and did not actually support interlacing. I wonder if Dell has gotten 'special permission'/ programming from Nvidia to advertise SLI on a non-Nvidia chipset???
Originally posted by Silverdog
I wonder if Dell has gotten 'special permission'/ programming from Nvidia to advertise SLI on a non-Nvidia chipset???
I don't know, I'd like to see confirmation of it as well. I don't understand why Tyan is promoting the feature if it didn't work. Given that the cards in question that Dell says is configured for SLI ($2700 for the pair of them, more than an entire base 690 w/ 1kW PSU), I would think that it's in nVidia's best interests to support it regardless of whose chipset is used.
I could have sworn that there was an nVidia PCIe bridge chip that is added to a base PCIe chipset that enables two of their boards to operate in SLI mode, but I can't find that chip right now. It might be that this chip is used in conjunction with the 5000X.
All this is speculation though, especially as to whether Apple would support SLI assuming the chipsets they get from Intel are capable of running it.
Originally posted by Apparatus
Power Supply on top makes no sense.
Maybe not, but that's what the source says.
While it's possible that the power connector could still be on the bottom (with a cable inside the case), when it came to drawing the mockups I decided to put the connector on top. The information received only specifed that the power supply was relocated to the top of the case.
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.
2) It looks like you're an AMD fan. Like it or not, Conroe dominates an X2 or FX one-on-one. But two X2s could (probably) beat a single Woodcrest (which is basically a Conroe at heart) at 2.0 or 2.33 GHz. Thus my point is that Apple wants to position the Mac Pros as competing with Precision workstations (which they'll destroy on price) and 4x4 (which they'll destroy performance-wise, for now), not XPS & co. on the low end (which'll pretty much rule the single-socket roost).
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.
Also intel new chips sets have raid bug even with a raid card.
http://hardware.slashdot.org/article.../07/06/1831221
http://www.nvnews.net/vbulletin/showthread.php?p=931556
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.
Also intel new chips sets have raid bug even with a raid card.
http://hardware.slashdot.org/article.../07/06/1831221
http://www.nvnews.net/vbulletin/showthread.php?p=931556
I'm sorry, the source of that news is the Inquirer, with no other confirmation. Inquirer is tabloid tech journalism at best.