SmokedMeat
Gamer™
3,567 Pikmin more powerful.
-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
How much more powerful is the Switch vs Wii U?
- Thread starter DOT DASH DOT
- Start date
Jaded Alyx
Member
This is pretty much what I remember from those tech threads way back.
DonF
Member
I believe that undocked it downclocks for better battery and since the screen is 720p it needs less horse power. It's not that the dock has extra processing power, but it uses more power when docked.
DOT DASH DOT
Member
I'll say one thing though, it does actually make a tangible difference going from 720p Wii U to 900/1080p Switch gaming, even if we are talking about a mainly resolution increase, it does make Nintendo games look so much better, and coupled with the fact that the hardware is league's above the Wii U in quality and design.
zenspider
Member
No idea if the item limit was hardware related, but we have no idea what data is actually being loaded for each item (AI, physics, animations, etc.).
Either way, this is a horrible metric for general "power".
nogoodnamesleft
Banned
I think that if you can't tell the games apart at "first glance" and DF needs to point out differences I don't think it's an unfair call tbh.
Slick Butter
Member
wii u's like 22 gamecubes duct taped together i guess
handheld mode switch is maybe 26 gamecubes soldered together
docked switch could be like 60 gamecubes
handheld mode switch is maybe 26 gamecubes soldered together
docked switch could be like 60 gamecubes
I wish there was some developers benchmark floating around.
Please remember this is a handheld that will up clock for TV output. Your closest analog in terms of power is Xbox 360+ with RAM Pak in handheld mode and Xbox 360 PRO when docked.
In terms of CPU Vs
It is equivilent to the espresso clocked at 2ghz with bonus full SIMD or roughly on par with xenos (the xenos was pretty inefficient) not bad for a 1ghz CPU. Or about 80% of the jaguar CPUs in the PS4 (on a single core basis)
GPU
A bit simpler.despite been 192 gflops in handheld mode due to been more modern is at least as good as the 240glops Xbox 360 if not better, and double or better in docked mode.
So in theory you should be able to run say modern warfare ghosts 360 at a full 720p with better textures in handheld and run it at 1080p when docked
Please remember this is a handheld that will up clock for TV output. Your closest analog in terms of power is Xbox 360+ with RAM Pak in handheld mode and Xbox 360 PRO when docked.
In terms of CPU Vs
It is equivilent to the espresso clocked at 2ghz with bonus full SIMD or roughly on par with xenos (the xenos was pretty inefficient) not bad for a 1ghz CPU. Or about 80% of the jaguar CPUs in the PS4 (on a single core basis)
GPU
A bit simpler.despite been 192 gflops in handheld mode due to been more modern is at least as good as the 240glops Xbox 360 if not better, and double or better in docked mode.
So in theory you should be able to run say modern warfare ghosts 360 at a full 720p with better textures in handheld and run it at 1080p when docked
lwilliams3
Member
Only the GPU clock speed changes docked/undocked. It still has a much more modern architecture with GPU features that matches the PS4-pro. The usable RAM quantity and CPU is also around 3.25x the Wii U's.
Skittzo0413
Member
... it sorta does though. The dock itself does nothing, but with the Switch connected to the AC adapter it upclocks the GPU cores by 2-2.25x, which is how you get higher resolutions in docked mode.
frankie_baby
Member
360 is ancient tech, undocked switch would probably be more like the equivalent of 500 xbox360 gflops
DeuceGamer
Member
Doesn't the RAM also change clockspeeds slightly?
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
WiiU runs MK8 fine at 60fps online. Running UE4 has nothing to do with strength.
Skittzo0413
Member
Someone correct me if I'm wrong, but the general FLOPS numbers we see (like 192 undocked) represent only the maximum theoretical floating point operations per second of that hardware, right? And more modern chips tend to be able to get far closer to that maximum than older chips, due to various architectural changes that are beyond my understanding, yes?
So hypothetically, the 360 may have averaged something (completely pulled out of my ass) like 75 GFLOPS while the Switch undocked can average something like 150 GFLOPS. Is that a good way to understand it?
According to DF Nintendo added the ability to run Switch and docked at the same highest RAM speed (1600 MHz I believe).
Gotdatmoney
Member
The Switch's GPU is stronger then what is in the 360 by a big chunk. Like the Switch wouldn't break a sweat running any 360 game in handheld mode what so ever. Compare Skyrim 360 vs Switch.
Engine support is much more dependent on feature set than power correct?
you have the ps360 and wii u positions mixed up
not sure if serious...
Skittzo0413
Member
Errr, the Wii U was more powerful than the PS3 and 360. Overall, anyway.
I think devs just often didn't put in the time to actually optimize their games all that well (and with that install base size who could blame them).
frankie_baby
Member
To be fair i dont actually know, just know that more modern stuff is more powerful flop for flop, even the wii u at 176 glops had a chunk more GPU power than the 360 at 240 and switch is obviously far more modern than that
Skittzo0413
Member
I dunno, that's how I've seen it explained before and it's always made more logical sense to me to think about it like that (more modern = can get closer to the theoretical peak) rather than X flops = 2* Y flops. It's just weird to quantify it in that way- a floating point operation is a floating point operation, why would one method of doing those operations result in a different level of performance from another if they wind up with the same speed.
But I readily admit I'm definitely no expert here.
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
Correct.
lwilliams3
Member
That chart has a lot of errors, including the comparison of the XBX, PS4Pro, PS4,etc.
frankie_baby
Member
Notice how close he puts the wii u to the wii, theres no error its just his agenda
butman
Member
GabeTheGrouch76
Member
1 GameCube = 9.4 GFLOPS
Wii U = 176 GFLOPS / 9.4 = 18.72 GC's
Switch
157 GFLOPS / 9.4 = 16.7 GC's
157 GFLOPS / 176 = 0.89 Wii U's
Switch
393 GFLOPS / 9.4 = 41.81 GC's
393 GFLOPS / 176 = 2.23 Wii U's
So there you have it.
Switch undocked / portable is 40 GameCubes / 21 Wii Us duct-taped together.
Switch undocked / portable is as powerful as 16 GameCubes / 0.8 Wii Us duct-taped together.
Switch docked is 40 GameCubes / 2 Wii Us duct-taped together.
Skittzo0413
Member
^ Besides switching docked and undocked those numbers make no sense. You're dividing the Switch GFLOPS by the number of GCs that the Wii U is.
But I appreciate the effort.
But I appreciate the effort.
frankie_baby
Member
Switch docked is 393 gflops undocked is 196
Undocked has two modes, one 196 and one 157 (153?) GFLOPS.
GabeTheGrouch76
Member
Yeah I didn't think the Wii U numbers through lol, here's a revision:
157 GFLOPS / 176 = 0.89 Wii U's
393 GFLOPS / 176 = 2.23 Wii U's
As for the docked and undocked numbers, I was only going by impressions posted earlier in the thread. I've seen people say Switch is more powerful undocked than docked, while others say otherwise. 'Tis a bit confusing.
frankie_baby
Member
True but no one is gonna make a game that needs power using the slower mode
Skittzo0413
Member
No worries, I definitely get that mix-up. As for undocked vs docked, undocked is verifiably less powerful than docked, since people have done tests to see that the Switch's power draw is a lot lower in undocked/portable mode. On top of that we have DF's leaks about the exact clock rates which is where we get the above numbers from (undocked was bumped up to a max of 196GF btw).
But as people have said this entire thread, just comparing FLOP numbers doesn't really tell you much since more modern chip sets will still be more powerful than older ones even if the raw FLOP numbers are lower.
I think the undocked/docked confusion may have come from comparing BotW's framerate. The reason it runs better in undocked mode is because it renders half the pixels and likely tones down some effects like edge AA.
frankie_baby
Member
A couple of games run a tiny bit worse in docked but obviously they run in higher resolution
The switch gpu got a speed boost from 304mhz to 384mhz (ish) so it is 196glops exactly half of docked
Nikana
Go Go Neo Rangers!
GOAT COMMENT
blu
Wants the largest console games publisher to avoid Nintendo's platforms.
Flipper didn't use flops for anything else but vertex computations. This comparison is pointless.
GabeTheGrouch76
Member
Oh no I wasn't being serious with my analysis, I was doing it to join in on the fun. :v
I also get what you're saying though. When looking up the numbers I've read posts explaining how the GFLOP numbers for Switch (undocked) and Wii U are lower compared to the PS3 and 360 (which are both around the low 200s), but PS3/360-era games on the former two platforms would have superior performance since they had newer architecture under the hood.
Geno Breaker
Banned
Same console, greater leap.
Yeah but the time span between SMW and SMW2 is greater than the one between the splatoon games. And Yoshi's Island uses the Super FX 2 chip.
ShadowFox08
Member
I thought the bandwidth was the biggest problem? At least gpu, CPU, and RAM scale proportionally.
Oh ha no I meant the ports of things like FF7, KH3, and other big multiplat games. Yeah it won't struggle with WiiU ports.
Skittzo0413
Member
Oh yeah I understood haha, I just like to put a little disclaimer for that comparison since a lot of people seem to still think flop to flop comparisons tell you the whole story.
But your first post which had the docked Switch being 21 Wii U's had me giggling.
frankie_baby
Member
Shouldnt really be needed as tegra uses tile based rendering (or something like that)