Is this a Good Machine

[DoFoT9]

But the way I see it from my point of view, until the day comes that I can walk into a Best Buy and buy any computer, and I mean any.. and be able to install mac os x on it, I'd rather wait until hackintosh installations become standard that way should I ever retire from a real mac, but wish to take os x with me, at least I can do it say on a 350 dollars laptop from HP or Acer for instance..

i dont think that this is ever going to happen "officially" but given the current status of hackintoshs and their ever expanding compatibility with a lot of PC hardware, its NEARLY possible. if you purchase identical hardware then you dont need to change anything! there are some installs of OSX that have been modified and will allow for virtually no hacking for ANY computer hardware, this brings limitations of course - but its nearly there.

What will be even more interesting is when the 2010 mac pro DOES come out, how one would be able to extract the microcode needed to run B1 stepping gulftown processors on the 2009's? That would certainly prolong any 2009 for many more years.. but, unlike the PC world.. this is not BIOS and its not some easy modification.. though my theory would be is to construct a efi flash utility to flash the microcode over to the existing efi firmware of the 2009s - enabling B1 to be recognized.. thus a 980x or its xeon equivalent would work on it.

A plan in theory, though I doubt it could ever be brought into fruitation.

i guess time will tell, give it a go when the 2010 Macs come out

 

[nanofrog]

you'd think that apple would consider such technologies - a simple upgrade such as eSATA or USB3.0 (i know its not on Intels cards yet, but im sure there are ways around this) might bring in lots of potential switchers onboard.

I doubt it, as they're into high margins, nor have they offered such features to date in the Intel based systems, despite additional features on competitor's enterprise systems.

 

[jav6454]

I doubt it, as they're into high margins, nor have they offered such features to date in the Intel based systems, despite additional features on competitor's enterprise systems.

Intel isn't supporting USB 3.0 and SATA 6Gb/s until mid 2011, so for all 2010 and early 2011, we are stuck with slow USB 2.0, unless you buy a PCI or PCIe based adapter or buy a motherboard with an integrated USB 3.0 chipset. Same goes with SATA 6Gb/s ports.

 

[DoFoT9]

I doubt it, as they're into high margins, nor have they offered such features to date in the Intel based systems, despite additional features on competitor's enterprise systems.

do you really think adding eSATA would take away the ~30% profit that apple could take away on each machine?

 

[nanofrog]

do you really think adding eSATA would take away the ~30% profit that apple could take away on each machine?

Definitely not, but they're using the chipset and a FW chip (only add-on semi) for features. As Intel's not making eSATA available in the chipset (or USB 3.0 until 2012), it's not going to happen.

I'd think they'd skip it in favor of LightPeak as it's faster, and potentially cheaper if other ports are eliminated (seems they're interested at least, given OS X was used during the first demo).

 

[Hellhammer]

Intel isn't supporting USB 3.0 and SATA 6Gb/s until mid 2011, so for all 2010 and early 2011, we are stuck with slow USB 2.0, unless you buy a PCI or PCIe based adapter or buy a motherboard with an integrated USB 3.0 chipset. Same goes with SATA 6Gb/s ports.

As far as I know, Cougar Point (LGA 1155 chipset) does support SATA 6Gb/s and it is expected to be launched in early 2011 along with LGA 1155 CPUs. USB 3.0 isn't supported though

 

[deconstruct60]

First off comparing a $600 used CPU to a brand new CPU sold through Apple as a means of "apples to apples" price comparison difference is whacked. A brand new 975 is much closer to $999 than $600.

http://www.newegg.com/Product/Produ...2&cm_re=i7_975_extreme-_-19-115-212-_-Product

You end up with around a $200 difference in price. Not these "many hundreds" and "thousand" differences that keep getting pushed forward.

It is $600 "cheap" because it is used. Has little to do with being more consumer oriented.

Hey Nano, you wrote:

"You don't need absolute parity though, as your usage doesn't require ECC memory. If you were doing things like heavy floating point calculations for scientific simulations/computing, it would be another story"

My response: Who on here would use such programs other than those who are scientists and or physicists?

ECC benefits come into play when your data is valuable and you have lots of data in memory. That is not necessarily floating point simulations.

So if you have 10's of GB of memory you have an increased need for ECC.

Simulations are sensitive because data from one iteration feeds into subsequent iterations. So the error starts to propagate. Similar issue with double versus single precision floats. If not sensitive to errors then can sometime use single precision.

Graphics is often one of those situations where it doesn't matter if data is "off". The display screens are finite (and relatively small). Approaches like ray tracing every single frame won't propagate between frames. If you run the frames in rapid sequence in an animation errors many not be noticeable. Few will care if running Nuke Dukem at 45fps and one of them botched.

Likewise if the response to a data corruption error is to press the reset/reboot bottom on the machine. Again don't really need ECC is don't want to find root causes of errors or good diagnostics. Used to be cheaper hard drives without SMART too. That doesn't mean there isn't utility in SMART being present on hard drives.

As for studies... the "studies" that come from folks not running ECC are suspect. Can't particularly study something if not measuring it.