Why Wii U is weaker than the 360 and PS3, explained intuitively.

#41EazyERockPosted 1/10/2013 9:50:18 PM
all that work you put into your essay TC and still, i have no care to give.
---
not changing this sig until RG3 wins Rookie of the Year
Bayonetta 2 is a Wii U exclusive.. Ya Mad Bro?
#423D_Shado(Topic Creator)Posted 1/10/2013 10:14:35 PM(edited)
neoVIVI222 posted...

The Wii U has a shorter pipeline than the 360 and PS3, I think the Wii U is around 5 stages and I know the 360 is around 24 stages. The best way to think about pipelines is how you do laundry, do you put one load(one instruction) in and wait for it to be fully done washed, dried and folded before starting the next or do you fill up the washer once it the first load in the washer is no longer using it. A simple pipeline would be 1) Get the instruction, 2)Decode the instruction 3)Load in data, 4)Execute, 5)Store. So the Wii U is probably has a similar pipeline to this while the 360 has broken this up into many parts.

So why pipeline? Because you get more done in a given time period but there are trade offs which will be in next post.


Lol, I know what a pipeline is buddy.

What I'm wondering is why you seem to think a shorter pipeline = more instructions per second. From your own analogy, it seems intuitively obvious that you could have at most X operations occurring at once, where each one is at a particular stage in the pipeline. Now, of course that isn't possible with many classes of instructions, but in general more pipelines is a GOOD thing.

EDIT: Just saw your clarification post. Good reasoning, certainly worth merit. Yes pipelining is a bit of a double-edged sword I suppose, but I'd need to see some stats before agreeing that the consoles with more pipeline stages actually do suffer from it in practice.
#43SavageCabbage88Posted 1/10/2013 10:18:20 PM
I don't get it. Why don't they throw the bricks into a wheelbarrow or into the back of someone's pick-up?
#44aether17Posted 1/10/2013 10:39:18 PM
3D_Shado posted...

EDIT: Just saw your clarification post. Good reasoning, certainly worth merit. Yes pipelining is a bit of a double-edged sword I suppose, but I'd need to see some stats before agreeing that the consoles with more pipeline stages actually do suffer from it in practice.


http://www.youtube.com/watch?v=PKF9GOE2q38

Look at 5:16 to 7:38

I think this explains it quite well. Yes, I know, some people aren't fans of "The Megahertz Myth", but forget about that, the pipeline explanations are far more important.

And I believe Espresso has 4 stages (according to people on NeoGaf and B3D), while Xenon indeed has 24.
#45DarkAdonis123Posted 1/10/2013 10:57:36 PM
jmichaelbp posted...
And yet ME3 runs better on the U than it does on the PS3.


How do you figure that?
---
If Platinum was HAND drawn, she'd be on paper, fool. This is a video game. They just made her on screen, no "drawing" involved -Delano7 on BlazBlue
#463D_Shado(Topic Creator)Posted 1/10/2013 10:59:26 PM
aether17 posted...
3D_Shado posted...

EDIT: Just saw your clarification post. Good reasoning, certainly worth merit. Yes pipelining is a bit of a double-edged sword I suppose, but I'd need to see some stats before agreeing that the consoles with more pipeline stages actually do suffer from it in practice.


http://www.youtube.com/watch?v=PKF9GOE2q38

Look at 5:16 to 7:38

I think this explains it quite well. Yes, I know, some people aren't fans of "The Megahertz Myth", but forget about that, the pipeline explanations are far more important.

And I believe Espresso has 4 stages (according to people on NeoGaf and B3D), while Xenon indeed has 24.


1. In this day and age, the Megahertz Myth feels like a strawman to me. We all realize that Mhz isn't the sole contributor to performance, and clearly a P4 at 3.4 Ghz wouldn't match an i5 at 2.4. I don't recall claiming other consoles were faster merely due to speed.

2. For the 45 vs. 82 seconds part, he never shows the code generating the stats. I would suspect it was intentionally made to support the G4. Anyone can make a benchmark that will be good to one CPU and bad to the next.

3. In his example, the items in both pipelines are moving at the same speed. I think in practice, the P4 would probably compensate for the branch misses soles.
#47FFnutPosted 1/10/2013 11:07:18 PM
I go to sleep and wake up to this now?

*yawns* Shado, stop reposting the same topics just because you get beaten in the previous one.

For anyone who doesn't know what I mean, Shado posted this exact topic earlier, only without breaking down the metaphor of brick-carriers = CPUs. Essentially, he wanted to fool people into saying "Well the first two are obviously faster" so he could turn it against them when the metaphor was revealed.

The correct answer to his question is "the first two groups are faster, but keep in mind they deliver a lot more bricks." And since Shado decides to deconstruct his metaphor, I'll deconstruct my answer: the bricks that are being delivered are "bricked" consoles.

TL;DR: Shado, if you want the metaphor to work STOP INCLUDING BRICKS IN THE METAPHOR.
---
Official Naoya of the Shin Megami Tensei IV Board
#48aether17Posted 1/11/2013 12:02:04 AM
3D_Shado posted...
aether17 posted...
3D_Shado posted...

EDIT: Just saw your clarification post. Good reasoning, certainly worth merit. Yes pipelining is a bit of a double-edged sword I suppose, but I'd need to see some stats before agreeing that the consoles with more pipeline stages actually do suffer from it in practice.


http://www.youtube.com/watch?v=PKF9GOE2q38

Look at 5:16 to 7:38

I think this explains it quite well. Yes, I know, some people aren't fans of "The Megahertz Myth", but forget about that, the pipeline explanations are far more important.

And I believe Espresso has 4 stages (according to people on NeoGaf and B3D), while Xenon indeed has 24.


1. In this day and age, the Megahertz Myth feels like a strawman to me. We all realize that Mhz isn't the sole contributor to performance, and clearly a P4 at 3.4 Ghz wouldn't match an i5 at 2.4. I don't recall claiming other consoles were faster merely due to speed.

2. For the 45 vs. 82 seconds part, he never shows the code generating the stats. I would suspect it was intentionally made to support the G4. Anyone can make a benchmark that will be good to one CPU and bad to the next.

3. In his example, the items in both pipelines are moving at the same speed. I think in practice, the P4 would probably compensate for the branch misses soles.


The Megahertz Myth wasn't my point though, my point was the explanation on how code will have issues like presented in the comparison. The P4 doesnt have much of a "saving grace" aside from the high clock speed, and actually the 360 CPU is only how it is because of its clock speed. If it was clocked lower, it would choke quite easily. Recall that its based on the G4 processor line, which introduced high clock speeds to many. It was impressive with how high things were clocked, but the IPC, and efficiency was incredibly low. Xenon is no different (it's based on PPC G4). It's not a "bad" CPU, but it's just very inefficient, with the high clock speed in order to compensate for that. Multithreading doesn't help much either considering how the 3.2 ghz becomes 1.6 ghz during that process.

I also think you should factor in the DSP and I/O processors into the equation, something both Cell and Xenon have to deal with themselves, at times taking quite a bit of CPU power. Espresso essentially ignores those and focuses on more important code.

Also note, Espresso does 5 instructions per clock, like Broadway (if it truly is Broadway based).
#49calhoun1389Posted 1/11/2013 2:20:59 AM
omniryu posted...
It affects on how many third party titles we get.

Go ahead and try to name 5 multiplatform games that were made worse for all systems just so the Wii U can get a copy. I'll wait. I'm 98% sure that you can't because it's all the same case as with the Wii and multiplatform games where the Nintendo system gets a port with reduced features and special control support.

These arguments are as petty and BS as when people were claiming that the Xbox 360 was holding back the PS3's multiplatforms, unwilling to accept that both systems can play 98% of the same games at the same quality (albeit, more discs on 360). It's all pretty much people nitpicking and just TRYING to find something wrong with a system they don't like just to justify to themselves and others why they don't like it, as well as convince everyone else that their opinion is the correct one.

If someone wants to buy something, leave them the hell alone, we all have the right to buy whatever product we want.
---
My first born, Jackson was born on Friday, May 11, 2012.
My Anime List: http://tinyurl.com/c6ylh9y (Updates regularly)
#50Killeryoshi8Posted 1/11/2013 2:27:30 AM
jmichaelbp posted...
And yet ME3 runs better on the U than it does on the PS3.


No it doesn't.
---
Blockscape - http://www.youtube.com/watch?NR=1&v=SFKlNskaRPU&feature=fvwp