Emulator

I'm sure this will be closed, but you can use this anyway.

https://skydrive.live.com/?cid=F76F2C2B2135891B&id=F76F2C2B2135891B!105

I just upped a few revisions that will fill some gaps.

something is strange with those build of yours
they were either very wrong revisions or build was done differently
i barely reach 101% with those 2 builds you posted to narrow the gap

so far with all above i used build bots builds.

That was when the fifo busy branch was merged, maybe it was slow because it was slow. :p

sure, fifo branch slow thing is the price of love, :) other solution is create options but i think there no good idea, btw the performance lost is depend the game but is not very noticiable. Enjoy the estability and "no options" + "peace" && "Love"

Bye

Please report me what games are more noticeable, I can disable some options and test these games (not for improve only for report what feature can affect the FPS)

SMG1 is something my niece is playing all the time.
That's why i noticed this big gap in speed
its impossible to play it on stock 3.3.ghz of 2500k any more.
I lowered it during the summer time.

Its the best game to check for speed issues.
Just standing in game after entering is good place for speed test as it posses all qualities for benchmark

Note
i do not know if fifobusy is responsible for this
i just tested one of fifo builds(pre merge)
with id 3.0-446 and info

• Sped up to the Dolphin debugger by not checking for breakpoints when stepping. etc..

And in it smg is fast as 363 first tested above with speed of 123-124%

i do not recall the exact revision when fifo was merged ,but it doesn't sound like it anyway.

The issue was somewhere between 456-491 (first post above)

I guess it'd be stupid to suggest that you didn't make sure all the settings were exactly the same with each build you've tested?

i am very precise about those things ;-)

if i remember correctly fifo merge was before 3.0-450? and problem started after 3.0-458
and before
3.0-491 as i stated fifo 3.0-446 build didn't had problems either
so problem is after fifo and between those two.

its this crazy git hashing i can't make heads or tail which revision is which

Yeah, those guys and their crazy GIT hashes, someone should complain. :D

speed tests should be done with a benchmark tool like fraps and work out the average over a period of time, some games will work fine on the menu, or even on 1st level but can drop once effects start showing or you come into contact with water etc. not reliable to just say "120%" because i can get 500% in zelda ww in some areas and 120% in others.

yes,good reference spot is always important
and for smg 1 that starting place is just that
it has depth,color,mob's and all rest required for such

In sonic adveture 2 i can get full speed with the 3.0 486.

And now with 3.0 735 i can only get 45 at 49 fps.

Both tested with the same config and in the stage 1 City Escape

With the 3.0 763 I can get fullspeed in sonic adventure 2 again and the sound is ok now too.