Support NeoGAF

[Dorago]

Yet it runs good on a Steam Deck …

It does run, but it's at 45 FPS with an 80 minute battery life.

Ask me how I know

 

[tr1p1ex]

at least we know from Outlaws that a game can't improve performance from an early build posted on the 'net.

 

[tkscz]

You can literally feel these 20fps in this video, what are they talking about? It was even worse than that and become better?

Did you have the video on mute? They said it's better than what was at GamesCon, but that it still has dropped frames down into the 20s. This is also not the final build.

 

[SABRE220]

I calculated that its roughly +20% over PS4 Pro CPU, without going into more modern features that the CPU support which again I can't really identify benefits for a game developer over a synthetic benchmark.

Its weaker than Series S CPU, and of course PS5, that's a given. I think it's pretty rare the cases where games are CPU bound on consoles in modern days unless the dev fucked up ala Capcom Dragon's dogma 2 NPC AI behavior.

Are there any actual benchmarks, that point to that conclusion? Also question is how does it compare to the Zen in the steam deck. I don't see it matching up CPU wise and might be why in CPU limited scenarios the deck has superior performance.

 

[Buggy Loop]

Are there any actual benchmarks, that point to that conclusion? Also question is how does it compare to the Zen in the steam deck. I don't see it matching up CPU wise and might be why in CPU limited scenarios the deck has superior performance.

For steam deck I actually posted one today. As of today like I detail in my post, overall performances are weaker than steam deck for CPU but was the better choice if you have both in front of you to implement in an APU. Switch 2's CPU is better IPC than Zen 2 and at higher clocks for a thicker portable it would wipe the floor. As of now, with the rumors that Nintendo will give back a core and burst clocks to 1.7GHz, if that happens then Switch 2 will take back the lead. Steam deck would have benefited a lot for more cores.

For PS4, PC got the luck (/s) of having a jaguar processor, AMD Kabini Athlon 5370 & a binned Xbox one APU A9-9820



I had an old anandtech link for the A9-9820 that no longer seems to direct to the right place, but this one is close enough for reference.
I'll keep the old number since anyway its a tad higher.

AMD A9-9820 - Xbox One APU - Jaguar @2.35GHz - 294
IPC interpolation means
PS4 - Jaguar @1.6GHz - 200
PS4 Pro - Jaguar @ 2.13GHz = 266

As I detailed in my original post, Switch 2 equivalent 1.1GHz clock would be = 336
Another ballpark with A78 from another product is :
Mediatek Dimensity 1200 - A78 @3GHz - 975
Switch 2 A78 @ 1.1GHz = 357

Somewhere between +26 to +34%

Which is not surprising. The IPC of old AMD Bulldozer / Jaguar / excavator generation were terrible. AMD had to do massive work to change things around with Zen

They had a massive uplift for the first Zen core compared to excavator so Jaguar is 2012 IPC roughly. Then Zen 2 had a good uplift on top of that. A78 IPC is ~7% higher than Zen 2. So all this combined is why a 1.1GHz A78C will beat a 2.13GHz Jaguar and with a good margin.

 

[Gamer79]

 

[blacktout]

Weird I haven't seen you in any of the Star Wars Outlaws threads.

 

[Gamer79]

Weird I haven't seen you in any of the Star Wars Outlaws threads.

Nothing to say. Star Wars games are mostly crap and only 30fps, ewww!

 

[FoxMcChief]

I don't know, why would you be apologizing?

I apologize for laughing at watching all the the Switch 2 haters eating crow. You'd think I wouldn't still laugh at it, considering how many L's they've collected, but here I am still laughing at the morons.