Yes, I expect software performances to catch up. Problem is, the nmp GPU aren't getting any younger, so unless apple or third party come up with GPU upgrades, you may not get to enjoy those gain for long. The tech world doesn't stand still for anybody.
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.
Has anyone decided to buy an iMac instead?
- Thread starter td2243
- Start date
- Sort by reaction score
I don't understand how these chips can't reach there top turbo rating yet the i7 equivalents can be OC'D well beyond their highest rating and stick those numbers consistently.
I'm assuming this is somewhat of a rhetorical question? Everything is a trade-off... performance, stability, and power consumption (or heat generation and cooling capacity) are all trade-offs. Turbo boost is just a form of built-in over clocking that doesn't compromise stability or power consumption, unlike most enthusiast over clockers who compensate with expensive cooling solutions or even sacrifice stability in the interest of even more performance.
I did acknowledge your testing a few posts later in this thread (#40); yes, they don't hit their top speeds, but they DO top out at about the same speed in testing. So if you have a single-threaded workload, the 4, 6, and 8-core units will be about the same speed, even if that speed is not, in fact, 3.9GHz.
Although VirtualRain answered this well, a Nikon D800 raw still is about 50MB; a Canon 5D Mark III is 30MB. On documentary jobs we may have five such cameras shooting 500 photos per camera per day. That is about 100 gigabytes per day.
Certain activities are time consuming, especially with lots of photos. E.g, Lightroom import and generating 1:1 previews. Or Photoshop HDR processing. Other activities (e.g, shake reduction sharpening) aren't that slow per se, but every control movement imparts a several second delay. So the artistic loop of repetitively inspecting, changing, re-inspecting is greatly impeded by even a few sec delay.
If you are doing production-quantity work under a deadline, a substantial performance improvement in your workflow "common path" is worth almost any price.
It's also not the Xeon as much as it is the other intangibles that come along with the "Pro" rig. Bus speeds, better GPU, etc. If they had an i7 version, that would be just as good.
Exactly.. All of those are available in a simple PC, so a workstation class machine isn't really needed.
Sorry... and good point... you're absolutely right that all three of these CPUs offer similar clock speeds on lightly threaded workflows.
----------
Sure... a "simple PC" can be equipped to perform the same as a "workstation class machine" but then I'd argue it's no longer a "simple PC". In my mind, a "simple PC" is something you do email, web, and MS office work on and perhaps even organize your snapshots or do some home video editing.
The fact is, some tasks, like RAW photo processing (in volume), HD Video with effects, 3D CAD and rendering, scientific modelling, and even graphic arts, require more horsepower than the average "simple PC" offers so that the technology gets out of the way and lets you be fluidly productive.
----------
Even if the software starts to take full advantage of the hardware, I don't anticipate needing a GPU upgrade until I need a whole new machine. Nothing I do is wholly dependent on the GPU (or any other part of the system)... performance of every part of the system is important and needs to be in some kind of balance. By the time I need a new GPU, I'll also want a faster SSD, CPU, more memory, etc.
There is significant truth to that, and my group does professional video and photo work using iMacs and PCs every day.
To a degree some "workstation"-class products like Xeon CPUs and nVidia Quadro GPUs are an effort to differentiate a higher-margin product which may not have major performance advantages, or those advantages may exist in a narrow range most users won't see.
OTOH, if you operate in a zone where the performance advantages exist, they are very important. E.g, if your common workflow involves efficient multi-threaded code, this could run several times faster on a 12-core machine than a 4-core machine. Likewise if you frequently do operations which are GPU accelerated and your software harnesses that.
Then there is production reliability and where your support comes from. It is easy to bolt together a fast PC which is pretty reliable, but chasing down the last infrequent glitches -- pushing systems integration testing to the limit -- isn't something many people have the time or inclination for.
If you play games or use a computer casually, those periodic glitches may not matter. You can take time to post on forums, look things up, try new BIOS versions, run stress tests, etc. If you are a professional trying to get work done, you are busy with your main job, and don't have time for chasing computer problems. You want to avoid problems and if they occur, want a single number to call for support.
I have spent a lot of time debugging my overclocked home-built Windows PC, and my 2013 iMac is just as fast and never crashes. I am happy with my iMac but I can definitely see a place for the Mac Pro which justifies the cost.
Most users will never use the full power of the Mac Pro so I would reccomend the imac the Mac Pro is a great machine but most people will never use all of that power so there is no point . I liked the old Mac Pros but now I just think they are a waste of money unless you really need all that power , They are not easily upgradable either .
I'm thinking about buying a nMP with Intel Xeon E5 3.7GHz base model since I run AutoDesk Revit in Win 7. Revit doesn't take advantage of muti-cores. I have a MP 1,1 and I'm looking to upgrade. I'm not sure if I should buy a nMP or iMac i7. I love how you can upgrade the MP's.
Then new mac pros are not easy to upgrade like the old mac pros . I just dont like the new look of the mac pro . They should have just kept both models .
My Mac Pro clients have had pretty much the same problem with the iMacs they have pushed hard. Most of them after 3 years and AppleCare expired I've managed to extend their lives by tearing down, re-pasting with Artic Silver 5 (original paste like dry powder by then) and some fitting better noctua fans inside which not only run faster but quieter also. Compare that to the towers which apart from an annual blow out with compressed air have kept going on and on. Now many of them have gone past 3 years most have had CPU upgrades, flashes to 5,1 and are still very much going strong indeed. Most are buying nMP models but that's to actually supplant and not replace the 4/5,1's instead it's the iMac's that are being retired from front line hard work and relegated to less stressful tasks!
The real pragmatic issue is that practically nobody using a modern Mac has purely single threaded workload. OS X is going to try to distribute some of the dozens of threads any normal instance is running at the time over multiple cores.
To get to a purely single thread, single core execution the OS would have to heard the dozens of processes and their associated threads onto one single core. You have to do that with no increase in cache misses or timesharing overhead.
Go to the utilities folder and open activity monitor and click to the CPU tab. All of that has to be herded precisely to absolutely maximize the Turbo setting. It doesn't particularly try. Because in not trying all that hard the overwhelming vast majority of the max Turbo benefit can be achieved. Turbo to 3.8GHz instead of 3.9GHz? the sky isn't going to fall. All the more so if the base rate is 3.0GHz. 800MHz is a substantial increase.