Can my PC run SMG1 and 2?

[rpglord]

(06-08-2012, 12:35 AM)AnyOldName3 Wrote: Hang on, I've now seen your LLE anti-stutter solution, so will believe the lack of stuttering. The main thing to consider is whether or not it will reintroduce the bugs of HLE.

In fact, a full set of settings screenshots may be useful for attempted replication of these results. Also, anything else we should know (ie are you using some strange build, have you recently been fired from Nintendo for stealing SMG2's source code and creating a static recompilation of it in x86-64 machine code) would be useful.

Unless you can tell me how to take all settings windows in one screenshoot,taking one screenshoot for each setting window sounds boring
I already told you which settings make a difference in this game-any other setting has no impact on perfomance,so what would you gain by looking at each at every setting ?
First and foremost,you must set efb copy to texture instead of ram.
I think everyone knows this gives big speedup in pretty much every game. In 15+ hours of playing SMG2 I have not encounter any instance where copy to ram would be required-no gameplay or graphic bugs with copy to texture.
Secondly,you need to enable accurate vbeam emulation.
And thirdly,if you dont want sound to skip,you need to disable idle skipping option.
Thats all the option which make a difference speedwise in this game.
If you are still not getting same pefomance,it means something must be wrong on your end.
I am using currently 3.0-600 but I tested this with couple of revisions.

[Garteal]

I'm with rpglord on this one.
After trying it out myself, instead of relying on information from others:
A 2500K, at stock clocks (3.3Ghz, no Turbo-Boost), should be able to play Super Mario Galaxy 2, with LLE, at 60FPS most of the time.
Similar CPU's should get similar performance.

[werewolfyman]

(06-08-2012, 02:20 AM)Garteal Wrote: I'm with rpglord on this one.
After trying it out myself, instead of relying on information from others:
A 2500K, at stock clocks (3.3Ghz, no Turbo-Boost), should be able to play Super Mario Galaxy 2, with LLE, at 60FPS most of the time.
Similar CPU's should get similar performance.

This seems to cause sound desynchronization on games that run on a single core. (Pokemon XD is the only game that has done this for me with the method RPGlord discovered.)

This works great with any other games I have. (Even Pokemon Colosseum, XD's predecessor.)

Maybe it's because I have LLE on thread enabled?
Back in a minute.

EDIT: Nope, sound still desynchronizes.
Weird, LLE on thread helps in every other game that isn't single core only.
Can someone try other games that rely on one core?

This is on an i5 2300 with Turbo boost enabled. (3.1 ghz.)

[Shonumi]

(06-08-2012, 02:20 AM)Garteal Wrote: I'm with rpglord on this one.
After trying it out myself, instead of relying on information from others:
A 2500K, at stock clocks (3.3Ghz, no Turbo-Boost), should be able to play Super Mario Galaxy 2, with LLE, at 60FPS most of the time.
Similar CPU's should get similar performance.

... Le gasp ...

This makes me want to test it, even though I haven't started dumping my Wii collection. If this is true (no Turbo-Boost?, woah...) then it seems like this rewrites some of the common knowledge passed around here on the forums, e.g. SMG2 + LLE needs a 4.4GHz i5-2500K. Can you make a video too, Garteal, or maybe just screenshots?

Quote:This seems to cause sound desynchronization on games that run on a single core. (Pokemon XD is the only game that has done this with the method RPGlord discovered.)

The sound desync might be a result of how Accurate VBeam affects sound timings and such. I have not tried to run games in single-core mode though, I'll do that later tonight.

[werewolfyman]

Might just be a problem with XD.
Not really all that bad as XD has less problems with HLE than Colosseum does.

[rpglord]

(06-08-2012, 02:45 AM)Shonumi Wrote: ... Le gasp ...

This makes me want to test it, even though I haven't started dumping my Wii collection. If this is true (no Turbo-Boost?, woah...) then it seems like this rewrites some of the common knowledge passed around here on the forums, e.g. SMG2 + LLE needs a 4.4GHz i5-2500K. Can you make a video too, Garteal, or maybe just screenshots?

I think maybe a lot of people just misunderstood "common knowledge".
If you want the game to run 100% speed 100% of the time then you do need about 4.4 ghz sandy bridge.
Main problem was sound skipping second the game wasn't running 100% thats why I can see some people saying if you want perfect experience you need overclocked sandy.
But the thing is,even before I discovered this trick,game was running 100% speed 90% of the time,so sound skipping in some rare places wasn't really a big issue.

[Shonumi]

(06-08-2012, 03:04 AM)rpglord Wrote: I think maybe a lot of people just misunderstood "common knowledge".
If you want the game to run 100% speed 100% of the time then you do need about 4.4 ghz sandy bridge.
Main problem was sound skipping second the game wasn't running 100% thats why I can see some people saying if you want perfect experience you need overclocked sandy.
But the thing is,even before I discovered this trick,game was running 100% speed 90% of the time,so sound skipping in some rare places wasn't really a big issue.

My main concern isn't the sound issue, but the FPS. Garteal said "most of the time" he's getting 60 FPS. A lot of people will say that you need an OCed Sandy/Ivy Bridge to play with LLE on SMG2 and maintain a high frame rate as well as a high game speed. That's why I want to see a video or something, to see what "most of the time" looks like compared to OCed setups.

[rpglord]

(06-08-2012, 03:20 AM)Shonumi Wrote: My main concern isn't the sound issue, but the FPS. Garteal said "most of the time" he's getting 60 FPS. A lot of people will say that you need an OCed Sandy/Ivy Bridge to play with LLE on SMG2 and maintain a high frame rate as well as a high game speed. That's why I want to see a video or something, to see what "most of the time" looks like compared to OCed setups.

Well,I have posted the video of game running with LLE.
It never drops below 50 fps on i5 2400,so game speed is not an issue. There is almost no difference between game running on 50 fps or 60 fps - most people will not notice the difference.
Not to mention if you have pal version of the game,you should be able to get 100% gamespeed all the time if you disable "euro60" option.
Also,videos are not really good way to show this,since they cause perfomance loss, unless of course recorded with camera

[Shonumi]

(06-08-2012, 03:50 AM)rpglord Wrote: Well,I have posted the video of game running with LLE.
It never drops below 50 fps on i5 2400,so game speed is not an issue. There is almost no difference between game running on 50 fps or 60 fps - most people will not notice the difference.
Not to mention if you have pal version of the game,you should be able to get 100% gamespeed all the time if you disable "euro60" option.
Also,videos are not really good way to show this,since they cause perfomance loss, unless of course recorded with camera

I'm not talking about whether there is a noticeable difference between 50 FPS or 60 FPS or whether if game speed becomes an issue, I'm talking about what it takes to effectively reach and maintain 60 FPS with SMG2 on LLE (NTSC of course). People have said that only a Sandy/Ivy Bridge could do this once OCed.

It's something like a gold-standard for Dolphin. If a rig can do 60 FPS more or less constantly (albeit, minor drops are expected) with LLE on that game, then it's technically a beast, as far as Dolphin gaming goes. Even if there is no visible difference in quality between 50 FPS and 60 FPS, it's still more difficult for machines to get 60 FPS rather than the 50 FPS.

If Garteal gets 60 FPS on SMG2 with LLE and that 60 FPS is consistent with only minor drops, then that changes what's been said here on the forums.

[rpglord]

(06-08-2012, 04:12 AM)Shonumi Wrote:

(06-08-2012, 03:50 AM)rpglord Wrote: Well,I have posted the video of game running with LLE.
It never drops below 50 fps on i5 2400,so game speed is not an issue. There is almost no difference between game running on 50 fps or 60 fps - most people will not notice the difference.
Not to mention if you have pal version of the game,you should be able to get 100% gamespeed all the time if you disable "euro60" option.
Also,videos are not really good way to show this,since they cause perfomance loss, unless of course recorded with camera

I'm not talking about whether there is a noticeable difference between 50 FPS or 60 FPS or whether if game speed becomes an issue, I'm talking about what it takes to effectively reach and maintain 60 FPS with SMG2 on LLE (NTSC of course). People have said that only a Sandy/Ivy Bridge could do this once OCed.

It's something like a gold-standard for Dolphin. If a rig can do 60 FPS more or less constantly (albeit, minor drops are expected) with LLE on that game, then it's technically a beast, as far as Dolphin gaming goes. Even if there is no visible difference in quality between 50 FPS and 60 FPS, it's still more difficult for machines to get 60 FPS rather than the 50 FPS.

If Garteal gets 60 FPS on SMG2 with LLE and that 60 FPS is consistent with only minor drops, then that changes what's been said here on the forums.

It all depends on what you consider "minor drops"
Like I said,it drops below 60 fps roughly 10% of the time,usually just when you land on planet
If you consider that minor drop,then yes you could say oc'd sandy is not required