rMBP 15" vs base Mac Pro

[UBS28]

According to geekbech, the rMBP 15" is as fast as the latest Quad Core Mac Pro in CPU performance?

That's pretty impressive considering the Mac Pro is a workstation and more expensive. The GPU performance is significantly lower though.

 

[dusk007]

That geekbench is a bit misleading.
Due to Turbo boost the mobile quads are quite fast and offer a responsive system but

The mobile Quads reach 3.3-3.6 Ghz on turbo mode but that only last for a while. Long enough for the benchmark but the base clock is much much lower. When the GPU is busy the CPU has much less room to turbo up.

The server quad core cpu has a base clock of 3.7Ghz and virtually no turbo with just 3.9Ghz. It essentially runs ALWAYS at max performance.

Also while a rMBP can provide performance when needed it gets quite hot and loud. The Mac Pro like most well built desktops can run at full load for hours while being barely any louder than normal and normal is fairly quiet.
Also the GPUs are on another level. GPUs scale much better than quad core CPUs if they have the TDP to play with. A D300 is supposed to be a 270X which is like a 760 desktop grade or a 780M.
There is also two gpus but they don't work together so that should not count for much. They always can run at their highest speed setting and with AMD's zero core probably need next to no power when not needed.

 

[kelon111]

The Mac Pro is designed to take higher workloads so of course the CPU would be able to maintain turbo speeds for much longer. A simple benchmark is generally not a good judge of long and heavy workloads.

 

[alphaod]

According to Geekbench, the top end rMBP is almost as fast as the base Mac Pro; the top end rMBP costs more than the base Mac Pro

That said, the rMBP is a portable computer with no expansion, whereas you can upgrade the processor on that Mac Pro in a year when the prices come down a bit (or buy a second hand part) to make it a lot faster.

 

[maplingstorie]

Get the Mac Pro if:
1) You're doing graphical intensive work or video editing
2) You don't need portability
3) You're rich

Get the Macbook Pro if:
1) You need a lot of processing power
2) You love portability
3) You can't live without the gorgeous retina display

I kind of believe that geek bench is right in regards to the score between the Mac Pro and Macbook Pro. CPU processing has nearly reach the threshold in terms of processing power, that's why NVIDIA introduced CUDA to aid in computational power. That's way I see it (I maybe wrong. forgive me) and CPU chips are getting more less power hungry with a small boost of performance each generation. Not to mention Mac Pro has higher L3 cache (faster RAM) so applications open faster.

 

[SarcasticJoe]

I will have to take a moment to point out that it's well known that Geekbench is woefully inaccurate and is limited to testing memory (RAM and HDD/SSD) and CPU, but only in a way that reflects performance in regular consumer applications with fairly little memory being used.

GPU's do not factor in at all and because of the small amounts of data used in the CPU tests, the main feature of server and workstation CPU's, a really big Level 3 (L3) cache, doesn't get to show it's worth. Combine this with both using the same SSD and the automatic error correcting in EEC memory the Mac Pro uses leveling the playing field, it's only the natural conclusion.

RAM in general is woefully slow compared to modern CPU's and the purpose of cache is to act as a much faster place where the most commonly used parts of what's in RAM is stored. This means that on average the CPU doesn't have to wait as long for data from RAM. The applications for which workstation and server CPU's are used (like editing audio and video) generally involve working with large sets of numerical data and in those tasks CPU's like those in the Mac Pro are miles ahead of anything else.

The only reason why Geekbench is so commonly used is because it's the only well known cross platform benchmark. On the Windows side everyone uses PassMark or Futuremark benchmarks, which aren't available on OSX. Personally I'd recommend that rather than Geekbench, people should use Cinebench and Uniengine Heaven or Valley.

 

[mrsavage1]

That geekbench is a bit misleading.
Due to Turbo boost the mobile quads are quite fast and offer a responsive system but

The mobile Quads reach 3.3-3.6 Ghz on turbo mode but that only last for a while. Long enough for the benchmark but the base clock is much much lower. When the GPU is busy the CPU has much less room to turbo up.

The server quad core cpu has a base clock of 3.7Ghz and virtually no turbo with just 3.9Ghz. It essentially runs ALWAYS at max performance.

Also while a rMBP can provide performance when needed it gets quite hot and loud. The Mac Pro like most well built desktops can run at full load for hours while being barely any louder than normal and normal is fairly quiet.
Also the GPUs are on another level. GPUs scale much better than quad core CPUs if they have the TDP to play with. A D300 is supposed to be a 270X which is like a 760 desktop grade or a 780M.
There is also two gpus but they don't work together so that should not count for much. They always can run at their highest speed setting and with AMD's zero core probably need next to no power when not needed.

Actually the 780m is a lot stronger than a single D300 in apps like games. the 780m is comparable to the D700 in games on OSX. The D300 however will be a lot stronger than a 780m on workstation apps. The mac pro has 2 D300s and if both are used, they will squash the 780m in FCPX or apps that utilise dual gpus and opencl.

 

[dusk007]

I don't know about osx but who if you care for gaming performance, OSX is a bad point to start. You just play in OSX when a reboot to Windows is too tedious and OSX offers good enough performance.

In pure performance the D300 is a lower clocked 270x a D700 is a 280x with lower clocks. A 780M reaches a fire strike score of 5400. A 280x is around 8400. Even with lower clocks that is not the same.
On Tomb Raider the 280X is about 50% faster. In BF4 same difference if you ignore anything lower than 1080p at full details. A 270X is about 30% shy of a 280X.
I don't know the clocks of a D300 but I doubt it doesn't at least tie with a 780M in Windows.

In any non gaming related stuff the FirePro should destroy a 780M and even with the workstation drivers the K3100 doesn't fare too well. AMD just has the better architecture for GPGPU this time around. Maybe Maxwell will change that.
Eitherway you have a GPU can run full tilt while the CPU is busy. Something that a thin notebook just cannot handle too well. A Desktop like an iMac or Mac Pro is just a lot better in that regard.

 

[SarcasticJoe]

I don't know about osx but who if you care for gaming performance, OSX is a bad point to start. You just play in OSX when a reboot to Windows is too tedious and OSX offers good enough performance.

The reason why people bring up games is because they're fairly good tools to test GPU performance and these days they graphics is far from the only thing they're used for.

Also, the Macbook Pro the Macbook Pro only has the option for a 750M, not a 780M, which is about 50% faster than a 750M. Let's not forget the Mac Pro only comes in dual GPU configurations, which increases the gap even further.

In other words: No matter how you try to misrepresent data, the Macbook Pro is not performance wise up to snuff with the Mac Pro.

 

[dusk007]

In other words: No matter how you try to misrepresent data, the Macbook Pro is not performance wise up to snuff with the Mac Pro.

LOL you did read my post? Did you?
Reading comprehension is not your strong suit is it.

A little hint, it does help to read an entire post and maybe also the post that it responds to, in order to understand text in context. Though reading my post again, there are a few errors which I will leave standing but I cannot understand how you got from it what you think you read.

BTW a 780M is NOT just 50% faster than a 750M. It is about 2-4 times faster. Usually about 3 times. And I had the 780M in my post because a few post before I guessed it to be about equal to a D300.