Radeon pro 400 series

inhalexhale1

macrumors 65816
Jul 17, 2011
1,039
685
PA
Just saw this, if anyone was curious for comparisons. From notebookcheck.com:

"Compared to the similar Radeon RX 460, the Pro 460 may be either the same or even feature more shaders but at a lower clock speed. The performance of the card could be similar to Radeon RX 460/ GTX 965M."

The comparison is interesting, since the Surface Book (MS's laptop that's priced similarly), uses the 965M.
 
  • Like
Reactions: Appleaker

macduke

macrumors G4
Jun 27, 2007
11,043
14,857
Central U.S.
Have the GPUs only come this far? Doesn't seem that crazy to me for some reason. I always look at the TFLOPS because that seems easy to compare and this GPU seems pretty similar to my late 2012 iMac at work with 1.9 TFLOPS of performance. So if by workstation you mean desktop from four years ago, then I guess? Please correct me if I'm wrong as I don't know a lot of the intricacies of GPUs these days, but I thought floating point is a good measure of raw output? I also know that my iMac, which is non-retina at 2560x1440, can get slightly laggy sometimes when doing visual things like switching spaces and expose if I've been working in a bunch of apps for a while. So that makes me worried about trying to run an external 4K display, much less multiple 5K displays plus the built-in retina display. I think my mid-2012 rMBP is around 0.75 TFLOPS and definitely gets laggy sometimes trying to run it's retina display.
 

mateo124

macrumors 6502
Oct 15, 2011
252
21
Have the GPUs only come this far? Doesn't seem that crazy to me for some reason. I always look at the TFLOPS because that seems easy to compare and this GPU seems pretty similar to my late 2012 iMac at work with 1.9 TFLOPS of performance. So if by workstation you mean desktop from four years ago, then I guess? Please correct me if I'm wrong as I don't know a lot of the intricacies of GPUs these days, but I thought floating point is a good measure of raw output? I also know that my iMac, which is non-retina at 2560x1440, can get slightly laggy sometimes when doing visual things like switching spaces and expose if I've been working in a bunch of apps for a while. So that makes me worried about trying to run an external 4K display, much less multiple 5K displays plus the built-in retina display. I think my mid-2012 rMBP is around 0.75 TFLOPS and definitely gets laggy sometimes trying to run it's retina display.
Take a look at TFLOPS numbers on the 10 series Nvidia GPUs. It's pretty clear performance wasn't their goal here, rather being able to charge over USB type c was.
 

Trahearne

macrumors 6502
Oct 6, 2014
418
73
Take a look at TFLOPS numbers on the 10 series Nvidia GPUs. It's pretty clear performance wasn't their goal here, rather being able to charge over USB type c was.
Energy seems to be too, given that AMD claims these GPUs all operating under 35W. Most people instantly compare it to GTX 1060 mobile, which is supposedly operating at >2x the power budget.
 

mateo124

macrumors 6502
Oct 15, 2011
252
21
Energy seems to be too, given that AMD claims these GPUs all operating under 35W. Most people instantly compare it to GTX 1060 mobile, which is supposedly operating at >2x the power budget.
Exactly. Using any other card would have required more power, which USB type-C cannot provide. Heat would also be an issue seeing as to how much smaller they've made it.
 

Appleaker

macrumors 68020
Jun 13, 2016
2,197
4,190
Have the GPUs only come this far? Doesn't seem that crazy to me for some reason. I always look at the TFLOPS because that seems easy to compare and this GPU seems pretty similar to my late 2012 iMac at work with 1.9 TFLOPS of performance. So if by workstation you mean desktop from four years ago, then I guess? Please correct me if I'm wrong as I don't know a lot of the intricacies of GPUs these days, but I thought floating point is a good measure of raw output? I also know that my iMac, which is non-retina at 2560x1440, can get slightly laggy sometimes when doing visual things like switching spaces and expose if I've been working in a bunch of apps for a while. So that makes me worried about trying to run an external 4K display, much less multiple 5K displays plus the built-in retina display. I think my mid-2012 rMBP is around 0.75 TFLOPS and definitely gets laggy sometimes trying to run it's retina display.
No, GPUs have gone a lot further. Apple has only come this far.
[doublepost=1477609634][/doublepost]
Exactly. Using any other card would have required more power, which USB type-C cannot provide. Heat would also be an issue seeing as to how much smaller they've made it.
It charges through Thunderbolt 3 and that can provide up to 100W. But are you talking in terms of the total power requirement? Because then you're right, because going to a 75W chip (which they would never do due to the design) wouldn't be possible.
 

akis-k

macrumors member
Original poster
Feb 4, 2010
49
25
Greece
Have the GPUs only come this far? Doesn't seem that crazy to me for some reason. I always look at the TFLOPS because that seems easy to compare and this GPU seems pretty similar to my late 2012 iMac at work with 1.9 TFLOPS of performance. So if by workstation you mean desktop from four years ago, then I guess? Please correct me if I'm wrong as I don't know a lot of the intricacies of GPUs these days, but I thought floating point is a good measure of raw output? I also know that my iMac, which is non-retina at 2560x1440, can get slightly laggy sometimes when doing visual things like switching spaces and expose if I've been working in a bunch of apps for a while. So that makes me worried about trying to run an external 4K display, much less multiple 5K displays plus the built-in retina display. I think my mid-2012 rMBP is around 0.75 TFLOPS and definitely gets laggy sometimes trying to run it's retina display.
By saying workstation, I am referring to AMD's promo that these cards are optimized for professional applications and not for being powerful enough. You could have an Nvidia quadro card for example that is much less powerful than a gaming card, but is much more efficient when working with professional applications.
 

h9826790

macrumors G5
Apr 3, 2014
14,408
6,957
Hong Kong
Apart from all the fuss with the touch bar, the price and the new design, it appears that apple has put inside their notebooks a workstation class Gpu, for the first time! This is very good news for content creators for sure! Some links that I found http://m.marketwired.com/press-rele...in-graphics-processors-nasdaq-amd-2170518.htm, http://creators.radeon.com/radeon-pro/
Don't be too excited, it's just named "Pro", I can see nothing make this 460 a workstation class GPU yet. No ECC RAM, No specific driver, No special support. Most likely no special optimisation in the pro software, and possible high failure rate due to poor cooling system.
 

macduke

macrumors G4
Jun 27, 2007
11,043
14,857
Central U.S.
By saying workstation, I am referring to AMD's promo that these cards are optimized for professional applications and not for being powerful enough. You could have an Nvidia quadro card for example that is much less powerful than a gaming card, but is much more efficient when working with professional applications.
Oh ok thanks. I was not aware that there were cards out there like this for mobile.
 

theitsage

Suspended
Aug 28, 2005
795
860
I don't see how any professional could get work done if a new MBP was to connect to two 5K displays. Top of the line Radeon Pro 460 has similar performance to the budget RX 460 Desktop GPU.

You're better off going with the non-dGPU MacBook Pro and spend the money on an eGPU build.

AKiTiO Thunder3 - $289
Dell DA-2 220W - $20
Radeon RX 480 - $250
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.