PS4 CPU and Nintendo Switch CPU, Comparison and info (Revised)

[Renton577]

I had a similar post to this a while back and have done some more research, so if you refer to the pictures here I have come to the conclusion that the Athlon 5150 is the Jaguar module that is in the PS4, as you know there are 2 of them both being 4 cores like is listed in the specs and each having 2MB of cache like listed as well and even having the same core clock, if you refer to the listed picture as well that gives the total score of the PS4 4,860 on multicore and 896 single core, since core count does not effect the single core performance, This could also be a good example of why the PS4 Pro struggles to reach 4K on titles seeing as the CPU is still the same with a higher clock. Now when it comes to the Nintendo Switch it uses a Tegra X1 so for reference I used an Nvidia Shield TV and halved the score seeing as the Switches CPU runs at half of the Shield TV's clocks, This give the Nintendo Switch a score of 758 Single Core and 2,173 multicore, this is in the base mode while in performance mode where the CPU clocks are bumped to 1.75Ghz that changes to 1,320 single core and 3,780 multicore, In this mode the performance disparity between the two consoles when it comes to CPU isn't to bad. I'm still hopeful we will get some decent 3rd party ports on the switch with all of this in mind seeing as the settings are usually there if a PC version has been made, lowering quality settings should let the switch run games like a PC would run games at low setting with current games or even medium if they have been optimized but still I feel its up in the air since its not just about if the console can run it, its if the developer is okay with the quality of it on the system. This also is a good look at why like I had said before the PS4 Pro cant really handle true 4K. With the Xbox One X's customized Jaguar chip it may help with overhead but even then I don't think we will see true 4K until the next console generation. This was just a little tidbit of info I wanted to share and hope its interesting to others as well.

[The Sloth]

 the switch has better IPC and single core score, the PS4 has 8 cores, clocked higher, so better multi core score, X86 is much better than ARM, also geek bench is not a good program to see scores. 

7 minutes ago, Renton577 said:

I don't think we will see true 4K until the next console generation. This was just a little tidbit of info I wanted to share and hope its interesting to others as well.

next generation, if Sony or Microsoft uses ryzen will might see 60 fps ( or 144fps with optimization, we can dream). 

 

[vanished]

Are these two not architecturally so vastly different that comparing them is questionable at best?  How could you even find a reliable benchmark to run on both?

[Renton577]

1 minute ago, Ryan_Vickers said:

Are these two not architecturally so vastly different that comparing them is questionable at best?  How could you even find a reliable benchmark to run on both?

Geekbench benchmarks were used for both, This is just a generalized comparison on both, seeing as the architectures are vastly different I don't expect this to be 100% accurate, None the less I found it interesting.

[manikyath]

i fear you made one critical mistake; especially in the case of nintendo, all that matters is that the game fucking runs, and somehow nintendo made mario kart work on a total of 88MB RAM in the original wii.

[PlayStation 2]

1 hour ago, manikyath said:

i fear you made one critical mistake; especially in the case of nintendo, all that matters is that the game fucking runs, and somehow nintendo made mario kart work on a total of 88MB RAM in the original wii.

Meh, MKWii is probably built off of Double Dash!!'s engine. So (basically) double the RAM and you can do more than what the GameCube could've done.

 

In any case, comparing an Nvidia Tegra chip to and AMD Jaguar chip is generally stupid as you can't make proper comparisons to them without tilting to either side.

[Mira Yurizaki]

I don't see how it's stupid to compare architectures as long as you're feeding them the same exact workload. Otherwise it would be stupid to compare NVIDIA against AMD because their GPUs use vastly different architectures and require mutually exclusive optimization methods.

 

The real problem for me with GeekBench is that it's tested not only on different hardware, but on a different software configuration. I'm certain that Android behaves differently than iOS than Windows. So the only fair comparison I will accept from any cross platform test is if only the hardware changes as little as possible. Otherwise, you have to take it with a grain of salt.

[Renton577]

On 8/13/2017 at 9:07 AM, M.Yurizaki said:

I don't see how it's stupid to compare architectures as long as you're feeding them the same exact workload. Otherwise it would be stupid to compare NVIDIA against AMD because their GPUs use vastly different architectures and require mutually exclusive optimization methods.

 

The real problem for me with GeekBench is that it's tested not only on different hardware, but on a different software configuration. I'm certain that Android behaves differently than iOS than Windows. So the only fair comparison I will accept from any cross platform test is if only the hardware changes as little as possible. Otherwise, you have to take it with a grain of salt.

I think the closest comparison then would be to Linux since Android is built on Linux, these benchmarks would be closer then when it comes to the actual OS being as similar as possible. But yes I agree Geekbench throws the same workloads at the processors regardless of architecture so it should have some use when comparing the two.

[AshleyAshes]

On 8/14/2017 at 8:47 PM, Renton577 said:

I think the closest comparison then would be to Linux since Android is built on Linux

By that rational, MacOSX runs like like Android...

 

This thread started with an apples to oranges comparison and now it's someone desperately trying to argue how similar apples are and that apples are just redder, less acidic oranges with a thinner peel...