By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Gaming Discussion - New Ipad Pro more Powerful than Xbox One

TheBraveGallade said:
The switch is at least three quarters of an xbone...

Only 25% difference between Switch and Xbox One? Whatever helps you sleep at night, I guess.

Xbox One: dynamic resolution 960x1080 to 1920x1080 at 60 fps
Switch (docked): dynamic resolution 640x360 to 1280x720 at 30 fps


Around the Network
Pemalite said:
teamsilent13 said:

No because RAM.

More to it than that.

haxxiy said:

It's almost surely based on half-precision floats.

Indeed.

haxxiy said:

Apple claims the A12X to be twice as fast graphically as the A10X, which was slower than the 820  and 821 Snapdragons on graphic benchmarks. In turn, we know those two are rated at about 410 - 520 single precision GFLOPS. Besides, if that were wrong and it were indeed FP32, Apple wouldn't pass the chance of claiming it was faster than the PS4 using FP16 as well, they surely love their marketing claims, after all.

More to performance than flops.

haxxiy said:

I've read somewhere that FP16 is sort of enough for mobile, but you probably know better if that claim holds water or not.

FP16 is usually opted for it's performance/power saving characteristics, it does come with a heap of caveats however.

JRPGfan said:
Why do companys get away with lying like this?

Are apple really saying its FP32 is more than 1.3 Tflops ?

More to performance than flops. We have had this discussion many times prior.
Graphics Core Next is old and antiquated, it's inefficient compared to more modern designs.

I thought the graphic architectures of Apple and the like were still years of development behind even AMD's stuff, so comparing computing performance would actually be flattering to them and not the opposite.

I mean... Imagination Technologies, man.



 

 

 

 

 

That NBA 2K19 demo was running at full Retina resolution which is 5.6 million pixels .... that's more than 2.5x 1080p resolution .... so it was rendering the game at 60 fps at a resolution way beyond an XBox One S. Not quite 4K resolution, but not far off.

I think their claim might be correct, that GPU must have some serious grunt underneath it to be able to pull that off. The CPU totally destroys the XB1S or PS4 CPUs.

If it can run NBA 2K19 at that resoution, I think there's a good chance it can run just about any other XBox One S game at a much lower resolution of 900p-1080p (which is where most XB1S games are rendered at).



haxxiy said:

I thought the graphic architectures of Apple and the like were still years of development behind even AMD's stuff, so comparing computing performance would actually be flattering to them and not the opposite.

I mean... Imagination Technologies, man.

AMD's latest graphics architectures are years ahead.
But the Xbox One's GPU is derived from hardware that came out 2012. Keep that in mind as we are about to enter into 2019.

Soundwave said:
That NBA 2K19 demo was running at full Retina resolution which is 5.6 million pixels .... that's more than 2.5x 1080p resolution .... so it was rendering the game at 60 fps at a resolution way beyond an XBox One S. Not quite 4K resolution, but not far off.

I think their claim might be correct, that GPU must have some serious grunt underneath it to be able to pull that off. The CPU totally destroys the XB1S or PS4 CPUs.

If it can run NBA 2K19 at that resoution, I think there's a good chance it can run just about any other XBox One S game at a much lower resolution of 900p-1080p (which is where most XB1S games are rendered at).

More to hardware, graphics and performance than just resolution.



--::{PC Gaming Master Race}::--

haxxiy said:
Pemalite said:
Formatting of your post is all messed up.

Need more low-level details to see if Apples claims actually hold any water... On the CPU front, mobile chips have been beating Jaguar for years.

It's almost surely based on half-precision floats.

Apple claims the A12X to be twice as fast graphically as the A10X, which was slower than the 820  and 821 Snapdragons on graphic benchmarks. In turn, we know those two are rated at about 410 - 520 single precision GFLOPS. Besides, if that were wrong and it were indeed FP32, Apple wouldn't pass the chance of claiming it was faster than the PS4 using FP16 as well, they surely love their marketing claims, after all.

I've read somewhere that FP16 is sort of enough for mobile, but you probably know better if that claim holds water or not.

Dude, the old ipad pro have about 3x the gpu performance compared to snapdragon 820/821. In GFXbench aztec ruins benchmark

Snapdragon 820/821 gets 6.1 fps on aztec ruins high settings and 16.8 fps on normal

Ipad pro (2017 model): 18 fps on high and 47.8 fps on normal.

Both tests are on offscreen so resolution is 1080p on both devices.

 

Now you say that this new ipad pro model is twice as fast as the old one putting it about 6x of snapdragon 820/821 and you say this device has 410-520 gigaflops, so you are saying this new ipad pro has 2460-3120 gigaflops (6x410 and 6x520).

LOL

Where you got source that snapdragon calculates in fp32? From what I know it's fp16 but I can't find source on either.

 

Sources:

Pixel 1 XL phone using snapdragon 821

https://gfxbench.com/device.jsp?benchmark=gfx50&os=Android&api=vulkan&cpu-arch=ARM&hwtype=GPU&hwname=Qualcomm%20Adreno%20(TM)%20530&did=39701616&D=Google%20Pixel%20XL

Ipad pro (2017 model)

https://gfxbench.com/device.jsp?benchmark=gfx50&os=iOS&api=metal&cpu-arch=ARM&hwtype=iGPU&hwname=Apple%20Inc.%20Apple%20A10X%20GPU&did=50838650&D=Apple%20iPad%20Pro%20(10.5-inch)

Last edited by Trumpstyle - on 02 November 2018

6x master league achiever in starcraft2

Beaten Sigrun on God of war mode

Beaten DOOM ultra-nightmare with NO endless ammo-rune, 2x super shotgun and no decoys on ps4 pro.

1-0 against Grubby in Wc3 frozen throne ladder!!

Around the Network

What matters is sustained performance, but yes this new ipad pros gpu should have similiar performance to xbox one S looking at gfxbench gpu benchmarks. I seen some people posting about the teraflops numbers. Huawei mate 20 gpu have 489 gigaflops and this ipad pro should be about 4x more powerful so we get a pretty high number for the ipad pro.

But we know from digitalfoundries fortnite analysis that switch in handheld mode beats the iphone X. So we can guess the sustained performance of this new ipad pro, thanks to being a tablet this allows for higher thermals and using 7nm instead of 10nm I'm guessing in total this will increase the sustained performance about 3x compared to the iphone X. Putting the new ipad pro about half or less sustained performanced compared to xbox one s.

Keep in mind that the gpu in ps4 is a cut-down gpu that uses 2.8 million transistors this new ipad pro have 10 million transistors, so the teraflops numbers for the new ipad pro might not be so crazy.



6x master league achiever in starcraft2

Beaten Sigrun on God of war mode

Beaten DOOM ultra-nightmare with NO endless ammo-rune, 2x super shotgun and no decoys on ps4 pro.

1-0 against Grubby in Wc3 frozen throne ladder!!

pointless if devs cant get the juice out of the ipad lol
devs will not dedicate the amount of time of console versions to mobile versions



It is more powerful!

...

in a few applications under certain conditions...

...

for a few minutes until it throttles...

But on paper it is faster! Yai, I guess!



Official member of VGC's Nintendo family, approved by the one and only RolStoppable. I feel honored.

JRPGfan said:
Why do companys get away with lying like this?

Are apple really saying its FP32 is more than 1.3 Tflops ?
because I seriously doubt a tablet has that.

It's large (12-inch), 5 years newer and it costs $1000. It wouldn't surprise me if it was close to the Xbox one.



It cost 1800$. It better be faster than the Xbox one and most laptops too.