Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
^^^^That is an excellent thread, it helped me make some decisions on the direction I wanted to head in upgrading my machine. However, the poster has a dual CPU 2009 machine, the OP has a single CPU 2012 machine. Some the sections are handy, while others may not be apropos.

Lou
 
I recently bought the Mac Pro Quad Core 3.2 (2012) and I want to ensure this tower lasts me as long as possible. I am not one of those that views the 'new Mac Pro' as something I see myself getting. So with this in mind, what would be the most important upgrade paths that would ensure the longevity of this machine? I am thinking of buying the 6-core Westmere 3.33 / 3.46 GHz to upgrade the processor in 2-3 years, also thinking of buying the ATI Radeon 5780 to upgrade the graphics card. As far as RAM goes and SSD I know those can be added later, so would anyone have suggestions if these 2 components are the most important things to buy now? I heard the processors are on the way out in the next year or so, so this is my reasoning. Please feel free to add your suggestions / recommendations. Thanks :)

Congrats on your new machine. It will provide you with many years of service. Some additional suggestions. Get a good UPS regulator and a surge protector to protect your Mac Pro from electric power spikes or fluctuations. My friend's PC desktop motherboard got fried when he was using the internet during a thunderstorm.. His modem was built in to his motherboard. Though this was a PC hardware, better to take precautions with Macs.

Try keeping your interior clean from dust from time to time as also suggested by other members. I also try to avoid installing videocards that exceed the Mac Pro's power supply requirements.

You can also put in additional SSDs and HDs at the second optical bay area That will give you around 6-7 storage drives including a flash SSD at your PCIe.

----------

^^^^Boy, you guys. He bought a Mac Pro for a Reason, the same as the rest of us. No matter, how we use it, it's upgradeable, and if we can afford to, why not not upgrade the stinkin' thing. If we weren't interested in performance we'd own a Mac Mini or iMac or even wait for the MiniPro. They'd get the job done, but be a lot less fun.

My wife and I live in a 3200 square foot home, do I really need all that room. Nope, I certainly don't, but I certainly enjoy it. I drive a 450 horsepower car. Do I really need all that power, again Nope, but, again, I enjoy it and top performance, within reason, is important to me. Both the car and my Mac are modded and I wouldn't have it any other way.

What I do with my car and my Mac is really my business. And modifying them with information gleaned from sources like this very forum is important to me. Getting Top Performance out of my equipment is what it's all about.

I probably spend 5 to 6 hours a day on my Mac, I do a lot in Photoshop (I'm an amateur Photographer). I'm on the internet a lot and do a little gaming. I'm also 74 years old and retired.

So, when someone asks for advice on how to upgrade his personal machine, offer advice or remain silent for gosh sakes. It's his machine, his money and his choice on how he spends it. Why be so negative:confused:

Lou

+1000
I was always hoping the forum would be a fun and enjoyable venue to exchange ideas learn a few and enjoy the discussion. After a day's hard work visiting this forum would be a source of relaxation in a way. :)
 
FYI, here's a recent news article that is still apparently vaporware, but worth keeping an eye on:


http://vr-zone.com/articles/thunderbolts-great-pcie-hope/50677.html


Nothing that hasn't been discussed as one of Apple's Mac Pro design options ... but it does look like Intel may actually be encouraging this now.


-hh

That's really interesting, but this part concerns me:

All add in cards and motherboards must be certified together and must contain prominent “Thunderbolt ready” identifiers according to Intel.

I don't see old Mac Pro motherboards getting certified and they certainly don't have a "Thunderbolt ready" identifier on them.

This sounds like the existing Asus PCIe card solution, which they actually mention in the article, that only works in tandem with a motherboard designed to support it. In which case, I don't understand why the Thunderbolt port isn't simply on the motherboard itself.
 
That's really interesting, but this part concerns me:

All add in cards and motherboards must be certified together and must contain prominent “Thunderbolt ready” identifiers according to Intel.

I don't see old Mac Pro motherboards getting certified and they certainly don't have a "Thunderbolt ready" identifier on them.

Yes, that caveat concerned me too.

This sounds like the existing Asus PCIe card solution, which they actually mention in the article, that only works in tandem with a motherboard designed to support it. In which case, I don't understand why the Thunderbolt port isn't simply on the motherboard itself.

True, but then the motherboard would cost slightly more, and since the PC hardware market is effectively a commodities market with razor-thin margins, it would be disadvantageous to them until such time that consumers differentiate on TB such that they're willing to pay more for it.

So by making it a BTO option as a PCIe card, it splits this apart - the consumers who differentiate on TB and are willing to pay more do so, and those who don't care about TB don't have it on their motherboard where it would reduce the hardware manufacturer's profit margin (since they're not able to differentiate pricing).

In any case, it does illustrate that Intel has concerns regarding the adoption of TB and is willing to do something about it. Granted, this doeesn't mean that Apple will be interested in getting a Mac Pro card certified - - and pragmatically, there's really no way that Apple would officially do such a thing -- but it does provide some tiny glimmer of hope for 3rd Party activities to be energized to possibly (maybe!) bring it about, such as what's happened with flashed firmware for non-Mac graphics cards and USB3.


-hh
 
True, but then the motherboard would cost slightly more,

Really? Did you skip over what was also in the article

"... The card will require a PCIe x4 or wider slot from the motherboard’s platform controller hub, a supporting BIOS that can manage SMBus as well as hot plug events, and an internal display port output connector. ..."

I doubt Intel is going to decouple their GPUs from this solution and that internal, on the board, output connector will be required. Highly doubtful this is going to be pervasive on non iGPU variant motherboards, just the the current motherboard implementer who are weaving in TB.


and since the PC hardware market is effectively a commodities market with razor-thin margins, it would be disadvantageous to them until such time that consumers differentiate on TB such that they're willing to pay more for it.

Even this "maybe later" option is going to cost a little more. All of these vendors are likely going to have variants of any of these compliant boards with this "maybe later" option stripped out and at a lower price. The folks who don't see any value right now in Thunderbolt are largely just going to be by-pass the option. [ Maybe take a shot at the next board in a future upgrade. ]


In any case, it does illustrate that Intel has concerns regarding the adoption of TB and is willing to do something about it.

TB adoption or just getting the folks who are itching for Rube Goldberg solutions to just stop complaining?

Details on PCI-e v2 interface requirements for USB 3.1 are pretty vague right now but this is one area where USB3.1 and Thunderbolt are going to compete. That is for the limited subset of x8 PCI-e v2 lanes on the IOHub chipset. ( 3.1 likely needs at least a x2 connection to get anywhere near over 5Gb/s. ). A slot is going to be pitched as a either a or b configuration mechanism.

That isn't so much adoption as survival change in direction. This has much more the flavor of a "finger in the dike" move to offset losses on ultrabook/portable design wins with backfill with a submarket that isn't strategic, but may keep momentum from completely stalling.


but it does provide some tiny glimmer of hope for 3rd Party activities to be energized to possibly (maybe!) bring it about, such as what's happened with flashed firmware for non-Mac graphics cards and USB3.

3rd parties are going to put a DP connector on a Mac Pro motherboard?
Not liklely at all. Someone comes up with some certified hack to mimic GPIO and some external loop back dispy do with external DisplayPort ... maybe. Folks come up with Rube Goldberg / Frankenmac contraptions all the time. That isn't particularly going to drive overall TB growth though significantly.
 
...
That isn't so much adoption as survival change in direction. This has much more the flavor of a "finger in the dike" move to offset losses on ultrabook/portable design wins with backfill with a submarket that isn't strategic, but may keep momentum from completely stalling.

Right.

3rd parties are going to put a DP connector on a Mac Pro motherboard? Not liklely at all. Someone comes up with some certified hack to mimic GPIO and some external loop back dispy do with external DisplayPort ... maybe.

The latter, of course. But the point is that until someone coughs up the board with the widget...no hack can even be contemplated, let alone actually attempted. Similarly, there's also certain things that a hardware producer can elect to do (with a nod and a wink) that can either aid or hinder such "off label" forks.


That isn't particularly going to drive overall TB growth though significantly.

Probably not, but more so because the overall percentage of personal computer sales which are still desktops (to possess PCIe slots) - instead of laptops - keeps dropping: if you look at it from the 'FrankenMacPro' perspective, its potential relevance can only grow, since the fishpond that they're swimming in keeps on getting smaller and smaller.


-hh
 

But, non-strategic solutions are not the root cause of the adoption problem. Intel not addressing the root cause problems is the only thing that will significantly turn around the adoption rate.

The gates on peripheral vendors probably is going to open up a bit with the volume arrival of TB v2. By end of CES 2014 should see if that happens or not. Given TB v2 isn't going to arrive before that time anyway to the general market TB is somewhat drifting.

The 2013 controller updates only really impacted hosts and only on Display Port v1.2.


The latter, of course. But the point is that until someone coughs up the board with the widget...no hack can even be contemplated, let alone actually attempted.

Actually I meant to type uncertified hack, but yes a hack.

Similarly, there's also certain things that a hardware producer can elect to do (with a nod and a wink) that can either aid or hinder such "off label" forks.

Folks who think they are clever and will promote uncertified distribution of cards are probably going to be cut off by Intel. It isn't going to be that hard for them to figure out when a significant number of TB controllers start to fall off the back of the truck for renegade hacks. Narrow band of tinker in the basement hacks? Sure, let a few folks whack away in the basement. Real volume to impact TB's volume problem? Probably not.

There is no good long term outcome is Thunderbolt getting wound up in a race to the bottom war. That is only going to drag it into deeper overlap with USB. If that becomes the primary battlefront, it will loose. I don't see back channel, cheap-for-cheapness sake being allowed by Intel while they still have 100% control on production. If that is the only way forward for some vendor just shut it down.

There is a slippery slope of enough competitors to keep the TB devices competitive on pricing and letting cheap chop shops in that will throw quality under the bus for a quick buck.


Probably not, but more so because the overall percentage of personal computer sales which are still desktops (to possess PCIe slots) - instead of laptops - keeps dropping: if you look at it from the 'FrankenMacPro' perspective, its potential relevance can only grow, since the fishpond that they're swimming in keeps on getting smaller and smaller.

So is the money in the fishpond. Intel isn't out to be King of Mayberry RFD. The only growth in significance here is to ignore a larger and larger segment of reality.

As a slow evolutionary tool to get pool shrinking resident amphibians up walking on land more (i.e., starting to transition to more Thunderbolt mix); perhaps.

The only thing about Intel and this FrankenMacPro thing that is a bit fuzzy is what is Intel's reference design that weaves in Xeon E5 going to be that incorporates Thunderbolt. The DP port on the board thing is a problem if Intel isn't going to bring iGPU to a least a subset of the series over the next couple of iterations.

If that is a future move, then I can see were these kinds of cards better fit into the puzzle in the FrankenMacPro pond.
 
I recently bought the Mac Pro Quad Core 3.2 (2012) and I want to ensure this tower lasts me as long as possible.

Please feel free to add your suggestions / recommendations. Thanks :)

A-B Expanding Foam Encasement would probably do the trick!


I think it lasts for about 40 years.
 
A-B Expanding Foam Encasement would probably do the trick!

i order from that company on occasion.. last time it was sleeved fiberglass mesh.. they have some cool stuff there but i can pretty much guarantee it's all super nasty to use.. they should include a free respirator with all purchases :)
 
With comments like this, this thread has now run it's course!

Lou

@Lou for Police Chief!!!!

Forum%20Police.gif
 
I'd recommend upgrading when you feel the need to, with a few exceptions.

Why wait? Things get cheaper over time, if you're not having any issues with the hardware you have, don't upgrade. You will be surprised.

Random things:

ActionableMango mentioned SSDs: I couldn't agree more. I would actually recommend setting up a "fusion drive" with it. I have something similar on my PC and couldn't possibly be happier (unless all my data were on the SSD in the first place... but that's expensive). It sounds like you have some big files so you'll have to keep some platter drives (unless you have tons of money, but let's assume you don't), but since you likely only use a few files in any given week, the fusion drive will seriously speed up your workflow. You will notice a huge difference.

As far as video cards, if you don't play games, don't upgrade. However, if you do, there are a ton of options at your disposal. Many last generation NVidia cards work great in the Mac Pro out of the box. Unfortunately, current generation cards are requiring more and more power. Some require 8 pin power inputs while the Mac Pro only offers 6. DO NOT USE ADAPTERS to fix this--you will need a second power supply. You have your choice between internal and external power supply. Internal power supplies (that fit into your extra optical bay) are currently cheaper than the price to make them, as in BUY NOW. Be aware that rigging an internal PSU will require a bit of time and some case modding. External PSUs are easier.

----------

I doubt Intel is going to decouple their GPUs from this solution and that internal, on the board, output connector will be required. Highly doubtful this is going to be pervasive on non iGPU variant motherboards, just the the current motherboard implementer who are weaving in TB.

Why is this doubtful? Why not just offer a thunderbolt port without the display? If you're not running a thunderbolt monitor, who would care?

There's also the ASUS option, which externally connects the display port to the card. You mention it like it wont happen, but Asus says this will be out "soon" (I know, it's been a while, intel's been poo-pooing it).

If intel is desperate for TB adoption, being more flexible would definitely help.
 
Last edited:
Why is this doubtful?

Intel is not going to actively promote the value/utility of its own GPUs. Seriously? Yes, they are. That's what companies do for their products.

Why not just offer a thunderbolt port without the display?

a. because it isn't Thunderbolt anymore. (there is already an external PCI-e standard)
b. it is not uniform. Normal user expectation is that if it has a standard Display Port plug if plug in a Display Port device will get video out. If there was a USB port and nothing happened when plug in a USB device would that be a good thing? No. Nothing is particularly different in the Thunderbolt context.


If you're not running a thunderbolt monitor, who would care?

All those orders of magnitude larger number of folks who have display port and DVI monitors which are compatible (with minor adapter) to the Thunderbolt socket. You don't need a pure TB monitor to get video out of a port provisioned by a TB controller.

Thunderbolt's purpose is not simply to replace PCI-e. It isn't really replacing PCI-e at all. It is a relocation of controllers who use PCI-e to communicate to the host far more than any "replacement".



There's also the ASUS option, which externally connects the display port to the card. You mention it like it wont happen, but Asus says this will be out "soon" (I know, it's been a while, intel's been poo-pooing it).

A card will be out 'soon". From these reported Intel requirements, it won't be that specific card from a year ago. Something that will fit in the same slot but with some physical changes.

That external cable thing is just largely silly. The vast majority of video consuming devices are perfectly capable of consuming the video stream that is already external on a cable from that cable. If already have gone through the drama of pushing it out through a port... to go back inside just to come out again is rarely going to be a value add beyond "just because I can and want to do it". There are some long distance transport scenarios that make sense if there was affordable fiber, but short range desktop contexts... it is just another cable to go to another cable. The latter is pure Rube Goldberg.

"but the Intel card has another cable inside, so it is cable to a cable too". That isn't thunderbolt driving that additional complexity. That is the forced modularity being overlaid.

If intel is desperate for TB adoption, being more flexible would definitely help.

Desperate is likely the wrong connotation. Intel has minimally sold 10's of millions of these controllers. I doubt they ever expected it to be a USB "killer" nor expected it to have USB like adoption rate. The industry's glacial adoption of EFI to replace BIOS is far more indicative of how slow motion the overall industry moves unless hard bundled into a required chipset package.

Flexibility to not deploy doesn't help really help promote adoption all that much. Extensive BIOS compatibility mode kept Windows not widespread adopting EFI for an extended period of time.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.