Mac Pro advice, 2.93 or 2.26 octo (go easy, unusual request)

[MCHR]

It's come to that time to (finally) upgrade. I've been looking at the quad 2.93 as well as the 2.26 octo. Sure both are formidable upgrades from my 2005 G5 2.5 water cooled tower.

Here's what I currently run on the machine.

Photoshop. Most files under 50MB, less than 20 layers when doing photo editing or sketching.

Illustrator. Nothing beyond three-fold page layouts, or editing vector art to be exported to Photoshop.

Web design stuff. Previously used GoLive, but we shall see what I have to learn with the newer Adobe package.

I am running a Wacom Cintiq tablet for sketching images (in Photoshop above), which seems to be my mainstay for secondary income. I also leave Mail, iTunes, Preview, possibly other apps open for reference when doing a sketch (.ppt, keynote, etc).

For future growth, I may entertain getting and using a 3D program such as Alias Autostudio if it becomes a demand. At my day job, I'm not using any .wire or ,iges files over 65MB, and I don't often use Autostudio for ultra ultra glossy renders. Mostly for package verification and rough screen shots for perspective reference.

Now for the question.

Based on my current workflow, which MP would make the best sense? The 2.93 quad, or the DP 2.26?

Last, would it make any sense to get the DP 2.26 based on the "possibility" of future 3D work, especially since I have neither the demand nor the software for it currently? However, I see it being a possible workflow advantage in many ways if I go that route, especially if I have that capability to some clients.

Keep in mind I would likely get a refurb, to minimise overall costs on an MP.

Thanks for any observations or comments.

 

[thermodynamic]

2.66 Single Quad owner here - I turned down the 2.93 CPU upgrade in favor of RAM and a better video card. 270MHz difference is not worth $500. Or even $450 (student cost).

I've included some links that include benchmarking tests - might be worth your time to read through (scroll down through them):

http://www.macworld.com/article/139507/2009/03/macpro2009.html


http://www.macworld.com/article/139919/2009/04/cto_macpro.html

My favorite:
http://www.barefeats.com/nehal04.html

Now while the bars look nice, don't forget the upfront costs. Cost-per-performance, there's a midpoint where you can't spend any more to justify smaller returns. (though it's amusing the 4x2.93 or 4x2.66 aren't tested with more than 6GB RAM).

The 2.93x4 vs 2.66x4 in AfterEffects is 12 seconds apart. No worth $500. Between 2.66x4 and 2.26x8 is 26 seconds. Again, not worth it -- never mind that most apps will be hindered at 2.26. If you ran 8 separate applications that all needed CPU time, at once, that'll do some good. But even I, and I do a lot of what you do and I can use all 8GB quickly, I have yet to get my Mac Pro (2.66 w/8GB) to crawl. Maybe I'll upgrade to 12 or 16GB, but that's about it.)

When the time comes, I'll do what others have and plonk in a w3580 (3.33GHz, $1000).

For $500, going from a 2.66 to a 2.93 offers negligible improvement and you're not going to see any benefits by going dual qual-core CPUs. And extrapolating what's shown in those benchmark sites (especially the first two from Macworld), even doing a single quad at 3.33 may not be cost-effective unless that time were essential and we compressed video for DVDs for a living...

I do some occasional AfterEffects and Cinema4D stuff too. They're the only apps that really take advantage of multiple cores (so do SOME functions of Photoshop, but not many) .

 

[tri3limited]

If you can afford the Octo, why not go for it?

My theory being that it'll be more useful for longer. I purchased an Octo yesterday with the aim of using it for 3 years before selling it and using the income to fund the next upgrade. Macs keep their value very very well as I found out recently. Sold a 1 year old, 24" 3.06 iMac three weeks ago for £1475 on eBay. Purchased it for approximately £1,320 originally.

Making a profit a whole year later is pretty good going!!

 

[MCHR]

If you can afford the Octo, why not go for it?

My theory being that it'll be more useful for longer.

Making a profit a whole year later is pretty good going!!

That's part of the equation I'm contemplating. My 2005 dual processor 2.5 is going on five years old, and I'd like to keep the system for a length of time. I know any decision isn't "future proof", but this time I am looking realistically at what I use and how often I'd ustilise the heavy lifting of the octo.

Part of me tells me that a 2.93 would make quicker work of graphics files.

Part of me is saying that an octo would be best for future growth.

The third part of me chimes in saying that I could always shorten the time I would take to upgrade to the next machine. . . .

 

[tri3limited]

I look at upgrading every 3 years as a target but looking at my recent habits this is my third upgrade this year and I can't imagine not upgrading any of my macs for 3 years. Look at sale prices of G4s and G5s on eBay, still alot of money for what they are in comparison to the PC market.

I imagine my MacBook Pro will be a year and a half to two, and I'll just be cutting my losses each time by eBaying the current model before hand.

Mac Mini probably will be 3-5 years as I don't really use it enough to justify a regular update, if one at all.

My Mac Pro will be almost certainly 3 years. I need to keep up with current tech for my profession so would require an upgrade at that time.

Do bear in mind swirling rumours of Mac Pro updates in Jan-March time. I went ahead and purchased anyway because I can't see that any of the rumoured updates would cost me significantly in the long run and now is the best time for me to upgrade from my financial position.

 

[MCHR]

One of the larger considerations is that most of the software and upgrades are actually slowing down the IBM-based PowerMac I have right now. The graphics apps work completely well enough, but if I find the need to get into a 3D program like Alias Autostudio, I will need to get into an Intel machine.

The upshot is that most of what I do isn't terribly processor intensive (mostly ram, usually in Photoshop), so I have some latitude on choice. Speed of 1 processor vs power of two. Forecasting 3-4 years out, I could always opt for a third party processor to boost speed, but that's no guarantee that a newer architecture wouldn't extract faster speeds at that time anyway.

 

[VirtualRain]

What about the new i7 iMac? For what you are doing, the screen is probably the most important thing to consider.

 

[goodcow]

What about the new i7 iMac? For what you are doing, the screen is probably the most important thing to consider.

Wouldn't that specifically disqualify him from an iMac then, because the glossy screen sucks for graphics work?

 

[thermodynamic]

Wouldn't that specifically disqualify him from an iMac then, because the glossy screen sucks for graphics work?

Not necessarily - the IPS panel is of top quality, but the gloss can be annoying for some... it was for me, but it is ultimately personal taste...

Not to mention, if the i7 processor used is the Mobile i7, that takes a performance hit to help keep heat down.

 

[MCHR]

The provision for the screen is definitely subjective, but I have a 24"HP that I really like. Matte finish works great for image mastering.

I had thought about the iMac lineup, but I have a number of peripherals (scanner, printer, extenal HD, camera I/O). Plus I'd have to believe the working lifespan of the iMac with my workflowwould be much shorter.

 

[VirtualRain]

The provision for the screen is definitely subjective, but I have a 24"HP that I really like. Matte finish works great for image mastering.

I had thought about the iMac lineup, but I have a number of peripherals (scanner, printer, extenal HD, camera I/O). Plus I'd have to believe the working lifespan of the iMac with my workflowwould be much shorter.

If you are definitely leaning towards a MP... I would get the highest clock speeds you can afford. It' not worth trading cores for clocks... especially with what you are doing. The ideal solution is probably to buy the 2.66 Quad now and upgrade the CPU to a 3.33 down the road when the price comes down.

 

[goodcow]

If you are definitely leaning towards a MP... I would get the highest clock speeds you can afford. It' not worth trading cores for clocks... especially with what you are doing. The ideal solution is probably to buy the 2.66 Quad now and upgrade the CPU to a 3.33 down the road when the price comes down.

This is probably better off for another thread, but I don't understand why people would buy a MacPro to just hack it with things like a new processor, thus voiding the warranty. In that case, just build your own Hackintosh from scratch. The only reason I pay the "Apple Tax" is because of the one-stop support from Apple.

 

[MCHR]

For the sake of keeping this discussion simple, let's assume the two MPs (2.93 quad, 2.26 DP) are the only ones I'm considering. Future processor swaps are IMHO a separate discussion, even if a viable option.

 

[VirtualRain]

This is probably better off for another thread, but I don't understand why people would buy a MacPro to just hack it with things like a new processor, thus voiding the warranty. In that case, just build your own Hackintosh from scratch. The only reason I pay the "Apple Tax" is because of the one-stop support from Apple.

Who says upgrading the processor is a "hack"? You guys kill me... There's threads on this forum swearing up and down that the Mac Pro is worth a huge premium over the iMac because it's upgradable, but then you see posts like this where upgrading it is considered a hack?!

For the sake of keeping this discussion simple, let's assume the two MPs (2.93 quad, 2.26 DP) are the only ones I'm considering. Future processor swaps are IMHO a separate discussion, even if a viable option.

Well, I guess you know my suggestion then... the 2.93 quad is probably a better investment. There are very limited circumstances that can leverage 8 cores and 16 threads whereas almost every task can benefit from the added clock speed. The fact is that both systems will be overkill for what you are planning to do, so don't worry too much and Mac Pro's hold their value well so if you need to sell and re-invest because your workload changes dramatically, you won't be out much.

I should add that if at all possible, try to budget for an SSD or two for your new rig... this is probably the most significant upgrade you can make assuming you equip with enough RAM.

 

[nanofrog]

For the sake of keeping this discussion simple, let's assume the two MPs (2.93 quad, 2.26 DP) are the only ones I'm considering. Future processor swaps are IMHO a separate discussion, even if a viable option.

Out of this pair, and what you've listed, the clockspeed would be important. More so than the additional cores IMO. So the 2.93GHz Quad would make more sense.

Who says upgrading the processor is a "hack"? You guys kill me... There's threads on this forum swearing up and down that the Mac Pro is worth a huge premium over the iMac because it's upgradable, but then you see posts like this where upgrading it is considered a hack?!

I don't see it as a "hack" either. Everyone's familiar and fairly comfortable with 3rd party upgrades (drives, memory,...) to save money.

So how is the CPU that different?
It just takes a little more effort is all, but quite doable. And it's been proven by others that it is indeed the case. One only needs follow the threads on it (namely the the height differences between naked and non-naked (IHS) on the CPU's used, so maybe a conductive gap material is needed for the Voltage Regs (mosfets)).

Useful with either machine (and the Quads won't need any filler material IIRC for the Voltage regs, as they had the IHS). Worth checking out the details though, as a fried daughterboard (VR's) isn't desirable at all.

Well, I guess you know my suggestion then... the 2.93 quad is probably a better investment. There are very limited circumstances that can leverage 8 cores and 16 threads whereas almost every task can benefit from the added clock speed. The fact is that both systems will be overkill for what you are planning to do, so don't worry too much and Mac Pro's hold their value well so if you need to sell and re-invest because your workload changes dramatically, you won't be out much.

I should add that if at all possible, try to budget for an SSD or two for your new rig... this is probably the most significant upgrade you can make assuming you equip with enough RAM.

The only reasoning I can see by choosing an Octad over the Quad, is either the applications that can utilize the cores will be needed shortly, or the possible issues with DIMM slots (memory requirement exceeding what's financially suitable with smaller sticks).

RAID of some sort may also be an option, especially if capacity is needed. SSD's still rather expensive in terms of cost per GB.

 

[Boomhowler]

I'd go for the quad one and put my money on a better graphicscard instead since that will both help open CL and make the computer work better overall (and buy a SSD systemdisc so you get a really snappy computer)

 

[MCHR]

Thanks for all the replies and suggestions.

It seems that with the (mostly) 2D softwares that I would be using, the 2.93 quad would make more sense. I wish I could forecast when I might be getting into 3D / Alias work, which would justify the 2.26 octo. Who knows, by the time I may need an upgrade, an octo 2.93 (or higher) may be reasonably priced.

Good suggestions on the graphics cards / ram to make the apps run faster. can anyone recommend an SSD? I probably would be looking at a low-capacity drive, since I have an external LaCie for storage.

Right now I'm booting up from a 74GB Raptor and I can tell the difference from the OE drive.

 

[nanofrog]

Can anyone recommend an SSD? I probably would be looking at a low-capacity drive, since I have an external LaCie for storage.

If you want performance, look at the Intel's. Say the 80GB, as it's the smallest. Others, such as OCZ, still have issues that need ironed out (i.e. reliability).

 

[VirtualRain]

If you want performance, look at the Intel's. Say the 80GB, as it's the smallest. Others, such as OCZ, still have issues that need ironed out (i.e. reliability).

+1... The Intel G2 drives are hard to beat.

 

[sboerup]

I'm also considering a 2.93 quad. Never heard about upgrading to a 3.33, that sounds interesting though . . .

I currently have a 2.8x8 (late 2008) with 16GB RAM. My main applications are also the same (Lightroom 40%, Photoshop 30%, Indesign, Illustrator, Dreamweaver) and I feel like I could be doing a little better with a 2.93 Quad. I'm also running Intel SSDs for operating system. Also running an ATi 4870, so I feel like I have maxed out my current system.

The big debate for me, cost isn't the biggest factor, is either go with the 2.93 Quad, or the 3.2 Octo (2008). Similar costs, near exact almost.

Would I be crazy to dump the 2.8x8 and go to a 2.93 Quad?

 

[alphaod]

Get the 2.93GHz Quad. The 2.26 Octad is kind of useless. It's not very fast in single threaded applications and it's also not very fast in multi-threaded applications either. RAM will be pricier, because you have less slots.