Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
6-Core Version of Apple's New Mac Pro Appears in Benchmarks
- Thread starter MacRumors
- Start date
- Sort by reaction score
Monitors
I don't see much in these Mac Pro forums about what sort of monitors folks plan on hooking up to this new coffee machine. Without a newer refresh on the Apple monitors, what sort of displays are out there without a glossy screen for pros to do retouching?
As for the discussions on speed, as a commercial photographer, a typical day is about 64 GB of files that need crunching into web proof pages. Right now with my 1st gen Intel Mac Pro (no, I don't upgrade my gear very often) I am chewing on these files for hours at a time... I can only take a lunch break once a day really... so this new speed will be greatly appreciated and needed.
I'm just a bit worried about buying a whole new rig but with old monitors, figured there must be something out there that will hook up well to this new gear to come in December. I wish that Apple would release new monitors in December with these!
Quote:
Originally Posted by WestonHarvey1 View Post
I don't understand the overkill argument. I've never met a computer that was fast enough for anything. Everything can always be faster.
It's called the law of diminishing returns. You're spending $1000 to load a webpage .1 seconds faster, or have your iPhoto library boot up 1 second faster. That's called overkill.
----------
Quote:
Originally Posted by friedmud View Post
The price of this machine is a drop in the bucket compared to the amount I cost my company. If it saves me even 2% of my time over _one_ year it will already pay for itself. Each year after that is gravy.
2% is nothing. Nobody is going to notice that, and nobody is going to fund that.
__________________
Apple TV Display - 4K h.265 - Coming Summer 2014
I don't see much in these Mac Pro forums about what sort of monitors folks plan on hooking up to this new coffee machine. Without a newer refresh on the Apple monitors, what sort of displays are out there without a glossy screen for pros to do retouching?
As for the discussions on speed, as a commercial photographer, a typical day is about 64 GB of files that need crunching into web proof pages. Right now with my 1st gen Intel Mac Pro (no, I don't upgrade my gear very often) I am chewing on these files for hours at a time... I can only take a lunch break once a day really... so this new speed will be greatly appreciated and needed.
I'm just a bit worried about buying a whole new rig but with old monitors, figured there must be something out there that will hook up well to this new gear to come in December. I wish that Apple would release new monitors in December with these!
Quote:
Originally Posted by WestonHarvey1 View Post
I don't understand the overkill argument. I've never met a computer that was fast enough for anything. Everything can always be faster.
It's called the law of diminishing returns. You're spending $1000 to load a webpage .1 seconds faster, or have your iPhoto library boot up 1 second faster. That's called overkill.
----------
Quote:
Originally Posted by friedmud View Post
The price of this machine is a drop in the bucket compared to the amount I cost my company. If it saves me even 2% of my time over _one_ year it will already pay for itself. Each year after that is gravy.
2% is nothing. Nobody is going to notice that, and nobody is going to fund that.
__________________
Apple TV Display - 4K h.265 - Coming Summer 2014
no you didn't because I left out the word "black" to copy your joke. wow, right over your head.
----------
you didn't. I left out the word black.
Oh I definitely hear you there. That is a real thing. John Carmack analyzed a Mac with a high speed camera and measured a pretty significant latency in the mouse input that isn't there in other OSes. It's been around since Tiger at least.
I don't know, a business of course wouldn't build a hackintosh, but someone who takes their work home might. Nowadays there are companies that build motherboards that run OS X out of the box; no need to customize any drivers or modify the installer in any way.
Neither poor performance of the Xeon or 32bit vs 64bit is to blame. Both CPU families used have a turbo boost feature so for single core tasks, they will be much closer in performance than clock speed and CPU type suggests. Comparing a quad i7 to a quad Xeon is even closer.
----------
But that means anyone can run it on their existing system without paying for it and their results will be completely comparable. This is important because it lets people know in real world terms whether the upgrade is worth it.
Years ago, it was the RAM that made all the difference. I too would be curious to know about threads and processor speed and multi-core.
That is what I would like to know, too!
----------
America is a continent, not a country...as in "South America", "Central America" and "North America" - unless you still believe in the Monroe Doctrine, of course
People have already explained, but in stylistics it is called totum pro parte, an inversion of pars pro toto, which is a type of synecdoche.
One of my friends thinks his couple-year-old hackintosh is going to be considerably faster. He says there is nothing really all that special about the hardware being used. It's going to be fast compared to the current Mac Pro, but not to non-Mac hardware currently available.
Then there is the whole problem of, faster for what? Without OpenCL versions of many of the pro apps, they might end up being orders of magnitude slower than a hackintosh with good nVidia cards.
His fear is that this is going to be a really cool semi-pro machine, but not what the real pros need... in which case it's too expensive. I guess we'll see where things go. I guess I'm hopeful Apple knows something we don't... like they've worked hard with all the major software devs to bring out killer OpenCL updates.
I think that's still a relatively small market in the big picture though. And, Apple would probably have a hard time winning them anyway. But, yea, I'd like to see them try too... as I'd love a something like a Mac Pro, but at a bit lower entry cost. I'm currently an iMac user (formerly mini) for that reason. If business goes well, maybe a Mac Pro is in my future, but there does seem to be a bit of a hole in the product line.
----------
True enough. One of my friends with a hackintosh has spent a number of sleepless nights getting things to work after updates over the years. That's worth $. That said, he makes the point that Apple has nothing as fast, and he's pretty sure the new Mac Pro isn't going to be as fast either.
The other problem I think he'd point out in your 2nd paragraph, is that with fixed GPUs, it isn't going to last 5+ years unless you're content to fall far behind on that front (or, unless Apple makes upgrades somehow available).
----------
I doubt most potential Mac Pro buyers are simply trying to shave .1 second off a web-page load. They are looking at things like how many seconds they have to wait each time they rotate a 3D scene a bit for the preview to redraw, or what happens when they scrub or apply an effect to video, or how many hours (or days) it will take when they hit the 'render' button.
If I'm rendering a 3D scene that is going to take 4 days, that 2% will save me a couple hours. That's worth money on it's own, but if means meeting or missing some deadline, it might be worth many times the price of the machine. But, we're generally not talking about only 2% difference either.
Yes, the high end 2010 iMac had a GeForce 320M card. That mini is the same design as the current version. Easy for Apple to offer a discrete GPU as an option if they wanted to. Whether they want to is another debate... maybe not even a debate.
http://apple-history.com/mac_mini_mid_10
I know, I know. But pros would also want to know how well this machine screams in 64 bit.
I understand about OpenCL, but generally an individual GPU core is a lot slower than an individual CPU. The performance comes from using a lot of GPU cores at once (there could be dozens or hundreds or thousands). So there are two parts to taking advantage of GPUs for general processing:
(1) state the problem so that it can be solved in a highly parallel way
(2) write the code in OpenCL according to that design (well, there are other GPGPU languages/frameworks, but the idea's the same).
Some tasks fit (1) more naturally, like video transcoding, but for others it can be hard to get much parallelism (if any).
I was looking for links because I keep seeing post (like the one I responded to) where it sounds like people have seen Apple claim that Mavericks will take advantage of GPUs to speed up general computing. That would imply that Apple has somehow come up with a solution for (1) for at least some general computing tasks. That would be very interesting because it seems to me (1) is the major impediment to taking advantage of GPU computing power. (Not to dismiss (2) -- also an important factor since developing code is time-consuming and expensive).
Of course, maybe those posters misunderstood Apple. Or maybe I'm misunderstanding them. Etc., so I was hoping for some links I could look at. GPGPU is very powerful, but is currently only being used for a very narrow set of things. I'm interested in anything that will GPGPU more generally used and useful.
Well duh. If you want to browse the web it would be a waste of money too. This is a specialty professional workstation.
True Milo... they are useful to you if you know the score of your current machine, so that you have something to compare it to. However the code that is written for these benchmark tests are usually ALOT more efficient code than those developed by large market applications.
These 'benchmark' programs are programmed specifically to do one task and then record the results.
While this at least gives you a 'scale' to determine how these results compare to other results, it still doesn't give you a clear indication of how well a program would run on it. (and the score number is only useful if you have another score to compare it to on the same exact hardware you have) As an example, I've had gaming PCs that benchmarked very high on the 'score', but would run certain games like crap.
I've also seen the SAME EXACT PC have a higher or lower score depending on where it was plugged in, in the house (maybe cooler or warmer climate? or difference in electrical outlet?)
Yep... so you either:
Get an iMac with a questionable lifespan
Build a Hackintosh that is one kernel update away from doom,
or buy the MacPro which might lock you into outdated GPUs quite soon.
---
So here's a question.. which of these will be faster for Lightroom/Photoshop?
a) hackintosh with 4-core i7 and a GTX 680 or 780
b) base MacPro
I suspect it does come down to the specs of that hackintosh motherboard.
----------
Agreed. I'd be more tempted to go iMac if it had a better track record.
which leads me to ask, (and i'm not trying to be trollish). Is the Mac Pro or even a hackintosh the best option here?
if you're not limited by software to a specific OS platform. Like Photoshop and LR are cross platform, if these are the only options you've listed, It sounds more and more like Apple is walking away from a lot of business.
They're trying to make the "professional computer" into a throw away computer. I dont know many professionals who treat their PC as an appliance, that is to be tossed when a new model is released. They are very particular about components and constantly, almost ridiculously upgrade their hardware internals to keep as up to date as possible.
as you said, if they want to maintain the same sort of hardware upgradability, they need to go hackintosh. Which is 1 major patch away from breaking.
Or Go "appliance" based computing and fork out $3000+ grand every 6 months to a year (or whenever Apple decideds to do their refreshes), rather than component upgrading.
In light of this, Plus the actual concerns i've read constantly from other Mac Pro threads about forcing everything external, lack up upgradability, Forced to use proprietary hardware and parts, its almost more worth it to go Windows based PC for workstations and just skip the new Mac Pro completely.
I honestly think Apple might have misfired with this tube. They need something, Some ace up their sleeve that makes this device truly worth it for professionals looking for a workstation. I don't think OSx, nor "design" is enough here. Professionals who are more concerned with performance and time spent on tasks, aren't going to pay the premium for the "tube" if it means long term performance limitations
Well, or you get a windows machine. That is if you don't have to have the OSX environment. And (sadly?) that's what I'm going to do.
The breakpoint might be the issue of GPU upgradeability. Thinking about that more, that (and not supporting the current OS) are the biggest problems with my 6.5 year old MacPro.
Or rather... even if the GPUs are upgradeable, no guarantee that there will be good ones available. We only got modern cards for the current MacPros with 10.8, and guess what, Apple throws all that out the window again, pissing of the card makers. Any wonder why they don't develop many cards for the Mac?
oh.. ok. i see what you were asking more clearly now..
the way i see it is that apple opted to ditch multi-socket CPUs in favor of multiple GPUs because basically, anything that can make use of multiple CPU cores can be programmed to take advantage of the more numerous GPU cores instead..
it's as if they're saying 'if you need fast linear processing, use CPU.. if you need fast multithreading, use GPU'.. looking into the future a bit, it seems fairly likely to have a (relatively) cheap means of plugging in GPU modules via thunderbolt for people who can make use of things such as renderfarms.. the current train of thought is to link a bunch of computers together in order to get more CPU cores going but not only does that take up (relatively) mega amounts of space, it also costs a lot more and uses more power than a (theoretical) GPU module.
but you (@iSee) already see what i'm getting at (even if it's theoretical).. re: the mavericks thing? yeah, i don't know what apple has said or how people may be misinterpreting what they said but i guess one possibility would be that apple is offloading some of the smaller (maybe even non parallel?) aspects of the operating system onto the GPU simply as a means to free up some CPU juice which will be used for application specific tasks..?
i really don't know but honestly, for what i would consider 'general computing', an ipad's (or definitely an air or mini) setup can handle that stuff just fine.. the elephant in the room with regards to speeding up general computing has (imo) nothing to do with CPU/GPU/RAM etc and everything to do with UI and input/output devices.. the computers have reached a point where they're plenty fast enough to handle most(by far) computing tasks as they can easily keep up with the speed in which humans and computers can currently interact.. i mean, i can have a thought in a split second but it may take upwards of a week to get it into a computer and back out into a form which i can communicate the idea with other people.. that's where the real slowdown with modern computing is..
dunno, this computer and pretty much everything else in apple's current line up is perfectly fine for 'general computing'.. they're also perfectly fine for most of the workload beyond general use.. there's only a slim percentage of people (slim still may mean millions of individuals) who actually need the type of processing power available from the new mac (and many people can and will easily max it out at times).. i guess my point is that if people are sitting around hyping up how mavericks is going to use openCL/GPU in it's day-to-day operations as if it's some sort of meaningful sales point than it's likely the new mac is way more powerful than their actual needs and they may just be finding some ways to justify purchasing one instead of just saying 'damn, that thing is sweet.. i'm going to buy it just because'
Not necessarily. It's great for things like rendering but for real time things like audio plugins it's much easier to do it with more CPU cores.
And I haven't really seen this discussed, but isn't the TB interface interconnected with the graphics hardware? Before this machine was announced some people were saying that a mac pro couldn't have TB at all because it didn't have integrated graphics.
Apple is saying they have dual GPU because of the processing power it ads, but isn't it possible the real reason is because they had to do that to have dual channels of TB? Many people have said they'd rather have single GPU and dual CPU, but would that configuration still be able to have dual channels of thunderbolt?
Last edited:
that's baloney.. i think your post should of ended after you said "i don't know many professionals".
pros upgrade on an as needed basis (need more storage? buy another hard drive..etc).. otherwise, they're generally running their computers into the ground in the same configuration as they bought them then replacing the whole system after 5-7 years..
you're talking about hobbyists and/or things you read on forums such as macrumors
yeah, that could be true. i really don't know as i don't have experience in audio.. i'm a cad/3d modeler/still image renderer so pretty much everything i say has to do based on my experience in those types of apps.. i realize i could be completely off the mark if i try to apply that knowledge to other areas of computing.
hmm.. i'm not really sure but my understanding is that the tb controllers are deeper in the system than the GPUs themselves and must be on the motherboards which is where the integrated graphics thing may have come into play.. a (current) mac pro can't have thunderbolt because the motherboards don't allow for it.. not because of which type of graphic cards they have.. (again, could be wrong but i think this is the case?)
i don't think this is what's really going on.. someone more tech savvy than myself would have to answer this though..