=DreamSNES 0.9.6 Benchmark Tests= - Old Benchmarks Reference

Anything DreamSNES-related can be asked in this forum, such as general questions about the emulator, how to burn it, etc.
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

=DreamSNES 0.9.6 Benchmark Tests= - Old Benchmarks Reference

Post by 404NotFound » Sun Mar 03, 2002 9:43 am

In this thread I hope to establish a TRUE speed percentage for games to determine thier playability. All games tested are domestic to my knowledge and are tested using the following settings:

----------------------------------------
Autosave - Yes
TV-Mode - No
Fullscreen - Off
Refreash - 60Hz
Sound - Mono (very important)
CPU Optimizations - Yes
Mode 7 Interpolations No (why is this even included in DSNES?)
Language - English
----------------------------------------

It is given that all games tested work with DSNES.. (duh) All speeds were displayed by pressing S3 on the keyboard. This command does not show framskip, so I could not determine the frame skipping.

Boldface roms run very well or almost full speed.
--------------------------------------------------------------------------------------

#
007 - 80%
100% Cotton - 85%
2020SuperBaseball - 90%
3 Ninjas Kick Back - 90%-100%
460 Kenen Monogator - 70%
777 Fighter - 95%
777 Fighter 2 - 70%
777 Fighter 3 -70%
7th Saga - 78%

A


AA Tenis - 100% (Maybe even past full speed)
ABC Monday Night Football - 70%
Ace Striker - 68%
Ace Tennis - 40%-50%
Acrobat Mission - 100%
Act Raiser - Froze going into battle. Otherwise 100%
Act Raiser 2 - Doesn't seem to run
Addams Family: 100% (drops to 80% when there's a lot on screen)
Addams Family Values - 100%
Addams Family-Puggsleys Scavenger Hunt - 100%
Adventure Island - 70% (High score bug)
Adventure Island 2 - 90%
Aero the Acrobat -100%
Aerobiz - 90%
Aero Fighter - 80%
Airrait - 100%
Aladdin - 75%-80%
Albert Odessey - 80%
Albert Odessey 2 - 75%-80%
Alien-3 - 80%
Aliens V. Preadator - Over 100% (Screen flickers slightly)
Alfred Chicken: 100%
Aunser Jr. Road to the Top - 70%
Amazing Tennis - 90%
Amazing Spiderman - 100%
American Gladiators - 95%
Ancient Magic - 100%
Anglique - 70%
Animaniacs - 70%
Another World: 80-85%
Anthrox Christmas Demo - 100%
Applesd - 100%
Arcades Greatest Hits - 100%
Arcana - 100%
Archer Maclean's Super Defender - 100%
Arans Spirits - 100%
Ardy Lightfoot - 90%
Area 88 - 100%
Aretha - 100%
Aretha 2 - 80%
Arkanoid - 90%
Art of Fighting 2 - 80%
ARYOL - 77%
Asterix - Freezes
Asterix and Oldex - 90%
Astral Bout - 93%
Astral Bout 2 - 100%
Astro Go Go - 100%
Axelay - 75%

B

Bahamut Lagoon - 65%
Batman Forever - 80% (I love the character names "mad jon" "mad phil")
Batman Returns - 70%
Battletoads in Battlemaniacs = 85-100%
Battletoads and Double Dragon = 100%
Beach Volleyball: 75-80%
Beavis and Butthead = 84-100%
Big Sky Trooper - Froze
Bio Metal: 45-55%
Black Thorne - 90% (Can we say Prince of Persia?)
Bobbes World - 70%
Boogerman: 100%
Bounty Sword - 80%
Brainies - 90%
Brain Lord - 70%
Breath of Fire - 75%
Breath of Fire 2 - 65%
Bubsy - 80%
Bugs Bunny - 100%
Bust a Move - 80%


C

Cameltry - 100%
Cannon Fodder - 100%
Castlevania IV - 70%-80%
Castlevania V - 80%-80%
Chaos Engine: 85-100%
Choplifter III - 80%
Chos in the Windy City - 100% (what the?!?)
Chrono Trigger - 70%
Civilization - Froze
Clay Fighter 2: Judgement Clay - 75%
Clue - 100%
Contra 3 : Stays around 75%, but it's playable.
Cool Spot - 90%
Cybernator - 100%

D

Dark Water - 80%
Demolition Man - 95%
Demon's Crest - 60-65%
Desert Strike - 90%
Devil's Curse - 80%
Dino Dini Soccer: 85%
Dirt Trax FX - 22%
Doom - Didn't work (bad dump?)
Donkey Kong Country - Very rarely below 100% (seems to be true for all DKC games)
Dragonball Z - 60%
Dragonball Z Hyper Dimension - Hard locked DC
Dragon's Lair 55%-75% playable
Dragon Quest I+II - 100%
Dragon Quest III - Black Screen
Dragon Quest IV - 70%
Dragon Quest V - 80%
Dragon The Bruce Lee Story - 80%
Dr. Franken - 95%
Dropzone: 100%

E

EarthBound - 100%
Earthworm Jim - 60%
Earthworm Jim 2 - 75-80%
Energy Breaker - 90%
Equinox - 80%
EVO - 85%

F

Fatal Fury Special - 65%
Fifa 96 70%-95% very playable
FIFA 98 - 80%
Final Fantasy III - 60%
Final Fantasy IV - 100%
Final Fantasy V - 80%
Final Fight - 90%
First Samurai: 65-75%
Flashback - 100%
Frogger 100%
Front Mission - 90%
F-Zero = 60%
F-Zero 2 - 60%

G

Ghoul Patrol 100% (can we say Zombies Ate My Neighbors?)
Gods: 85-100%
Go Go Ackman - 80%
Go Go Ackman 2 - 100%
Gradius 3 - 80%
Gundam Wing Endless Duel - 70%

H

hardball 70%-80% playable
Harvest Moon - 65%
Head on Soccer 75%-90%
Home Alone - Black Screen
Home Alone 2 - 100%
Hook - 65%

I

Illusion of Gaia - 90%
Inspector Gadget - 90%
International Superstar Soccer - 93%

J

James Bond Jr. - 95%
Jim Powers: 90% (the scrolling in it is bad)
Judge Dredd - 55%
Jungle Book - 85%
Jungle Strike - 100%
Jurassic Park - 90%
Jurassic Park 2 - 95%

K

Kick Off: 100%
Killer Instinct = 90-100%
Kirby's Avalanche = 65-93%
Kirby's Dreamland - Hard Lock
Kirby's Dream Course - 95%
Kirby Superstar = Did Not Run, Locked Up DreamSNES
Knights of the Round - 70%

L

Lady Stalker - 100%
Last Action Hero - 100%
Last Fight (?): 90-100% (Not sure of the name, it's Japanese)
Legend of the Mystical Ninja = 75-86%
Lemmings - 100%
Lemmings 2 - 100%
Lethal Weapon - 95%
Libble Rabble - 100%
Live a Live - 90%
Lost Vikings - 90%
Lost Vikings 2 - 100%
Lufia Fortess of Doom - 100% (90% in battle)
Lufia 2 - 100%

M

Magical Drop - 95%
Magical Drop 2 - 100%
Magical Quest Starring Mickey Mouse - 80%
Magical Knight RayEarth - 100%
Major Title: 80%
Mario RPG - Does not run
Mechwarrior - 70%
Mechwarrior 3050 - 80%
Megalomania - 100%
Megaman 7: Im suprised...this one runs 100% in the levels (slowdown to about 85-95% sometimes when lots of stuff is going on) Around 90-95% during Boss Battles.
Megaman 9 (Rockman and Forte?) - 100%
Megaman Soccer: Constant 75%
MegaMan X - 90% normal, 80% with plasma shots on screen
MegaMan X2 - Not working
MegaMan X3 - Not working
Mickey Mania - 90%
Micro Machines - 85%
Micro Machines 2 - 95%
Mortal Kombat II - 70%
Mr Nutz: 100%
Mystic Ark - 65%
Mystic Quest - 100%

N

NBA Jam = 100%
NBA Live '95 75%-90% very playable
NBA Live '98 - 100%
NFL Quarterback Club - 70%
NHL 98 -90-100% (Sound Errors)
NHLPA Hockey '93 75%-90% very playable
Ninja Gaiden Trilogy - 70%

O

Ogre Battle - 70%
Out To Lunch - 100%

P

Pac Attack: 90-100%
Pac in Time - 90%
Parodius - 75%
PGA Tour '96 - Will Not Run
Phalanx: 90-100%
Phantom 2040 - 100%
Pilot Wings - 50%
Pinball Dreams: 90-100% (very jerky)
Pinball Fantasies: 75-80% (very jerky)
Pit Fighter: 70-100% (it's terrible!)
Pitfall: 85-100%
Pocky % Rockey - 95%
Popful Mail: 85-100%
Prince Of Persia - 90% (120 minutes? The PC version has 60 minutes.)
Putty Squad: 100%


Q

R

Rock 'n Roll Racing: 75-85%
Rockman and Forte - above 95%
RPM Racing: 70%

S

Saturday Night Slammasters: 75-85% (there's a score bug & the rom freezed after the first fight)
Seiken Densetsu 3: 60%
Sensible Soccer: 80-90% (jerky)
Side Pocket: 100%
Soccer Kid: 95-100%
Spiderman In Maximum Carnage = 71-76%
Spindizzy Worlds: 80%
Starfox = 30-35% (severe graphical corruption)
Street Fighter II = 90-100%
Super Bowling: 70%
Super Double Dragon: I hate this damn game...i liked the origional....but anyways....80-85% but it seems a bit slower
Super Dunkshot: 70%
Super Ghouls N Ghosts = 78-82%
Super International Cricket: 78-83%
Super Pang: 75%
Super Punch-Out = 100%
Super Mario Brothers All Stars
.........Super Mario Brothers - 90%-100%
.........Super Mario Brothers - Lost Levels - 90%-100%
.........Super Mario Brothers 2 - 93%-100%
.........Super Mario Brothers 3 - 90%-100%
Super Mario Kart - 70%
Super Mario World - 100%
Super Metriod - 90%

T

Teenage Mutant Ninja Turtles 4: 75-80%

U

V

W

Wolfenstsin 3D: 70%
WCW Superbrawl: 80%

X

Y

Yoshi's Island = 45-55%

Z

Zelda 3 - 95%
Zombies Ate My Neighbors = 100%

--------------------------------------------------------------------------------------
Well there you have it. Updated 3-25-2002. If this topic does not stay stickied when i edit it, which it may not, be sure to check the full thread for further updated lists. I am halfway through... about.
Anyway there you go. Feel free to contribute.

Thanks to: mirrorsaw_2000, Tails Prower, catfishman, and Dr. Zoidberg for thier contribution to the list.

The old =Dreamsnes Compatibility List= can be found here.
Last edited by 404NotFound on Thu Apr 11, 2002 5:53 pm, edited 14 times in total.
dsj1
DCEmu Crazy Poster
DCEmu Crazy Poster
Posts: 31
Joined: Wed Oct 17, 2001 7:44 pm
Has liked: 0
Been liked: 0

Dreamsnes .96 not-full screen?

Post by dsj1 » Sun Mar 03, 2002 10:42 am

I didn't even know dreamsnes has an option for non-full screen? Does it really make a big difference? How does killer instinct play-speed wise? How about super punch-out? This emulator is really good though- it will someday be very close to perfect I would guess! But those guys are "brilliant" programers. Just think if they did this as a full time job! It would be as good as Zsnes on the pc. WOW!
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

Dreamsnes .96 not-full screen?

Post by 404NotFound » Sun Mar 03, 2002 11:26 am

dsj1 wrote: I didn't even know dreamsnes has an option for non-full screen? Does it really make a big difference? How does killer instinct play-speed wise? How about super punch-out? This emulator is really good though- it will someday be very close to perfect I would guess! But those guys are "brilliant" programers. Just think if they did this as a full time job! It would be as good as Zsnes on the pc. WOW!
*head pops like a balloon*

Well the point of this thread was for people to add onto the list and make a nice accurate speed benchmark for all games, but to answer your question and to state the meaning of life: <Statement taken out to prevent mass human race chaos>


I'll post up some more after my DC has a rest from all the title scanning and i get some more time.
[This post was edited by: I@N]

if you believe the above plz get ur head checked :)
Last edited by 404NotFound on Mon Mar 25, 2002 7:12 pm, edited 1 time in total.
mirrorsaw_2000
Insane DCEmu
Insane DCEmu
Posts: 119
Joined: Mon Feb 18, 2002 5:18 pm
Has liked: 0
Been liked: 0

Post by mirrorsaw_2000 » Sun Mar 03, 2002 12:40 pm

Testing Conditions: Same as 404s

Actraiser = 86-100%
Contra 3 = 65-75% (drops to 50-60% at the first boss)
F-Zero = 60%
Super Ghouls N Ghosts = 78-82%
Super Punch-Out = 100%
Street Fighter II = 90-100%
Killer Instinct = 90-100%
Mortal Kombat = 65-75%
Mortal Kombat II = 65-75% (sound issues)
Mortal Kombat III = 58-65% (sound issues)
Zombies Ate My Neighbors = 100%
Yoshi's Island = 45-55%
NBA Jam = 100%
Starfox = 30-35% (severe graphical corruption)
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

Post by 404NotFound » Sun Mar 03, 2002 1:58 pm

Sweet, i'll burn a copy with my whole rom collection and start at the top later.

But i got stuff to do right now.
Last edited by 404NotFound on Mon Mar 25, 2002 7:14 pm, edited 1 time in total.
dsj1
DCEmu Crazy Poster
DCEmu Crazy Poster
Posts: 31
Joined: Wed Oct 17, 2001 7:44 pm
Has liked: 0
Been liked: 0

roms_lst

Post by dsj1 » Sun Mar 03, 2002 5:22 pm

Hey-

Use dcfactory to burn dreamsnes .96 it'll put an automatic roms-list in there for ya! It'll still access the drive (plus the .mp3-although you can just take out the .mp3 file and make up a name with no mp3 file in the sound .dir if you want!)but not nearly as bad as before! Try it it works great!
Thanks for the info all-
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

roms_lst

Post by 404NotFound » Sun Mar 03, 2002 5:37 pm

dsj1 wrote: Hey-

Use dcfactory to burn dreamsnes .96 it'll put an automatic roms-list in there for ya! It'll still access the drive (plus the .mp3-although you can just take out the .mp3 file and make up a name with no mp3 file in the sound .dir if you want!)but not nearly as bad as before! Try it it works great!
Thanks for the info all-
Yeah i use DCFactory, but with 400 roms you know :)
Last edited by 404NotFound on Mon Mar 25, 2002 7:14 pm, edited 1 time in total.
Mr.TA
DCEmu Cool Newbie
DCEmu Cool Newbie
Posts: 15
Joined: Tue Feb 12, 2002 8:01 pm
Has liked: 0
Been liked: 0

Post by Mr.TA » Sun Mar 03, 2002 7:31 pm

The way you folks are trying to benchmark DS is all wrong IMO. In the real world there would be no frame skipping at all. Since DS doesn't provide us with that option all we have is the emu telling us a percentage of frames displayed out of the amount of frames skipped. Not so good for "real" benchmarking. So what we have here is a delima...on one side we have a frame skipping percentage given by DS and on the other we have a whatcha see percentage. I'll deal on the latter.

Let me try to clarify things a bit...a real SNES displays a 60hz or 50hz video signal depending on the region your from (NTSC = 60hz,PAL = 50hz) which means 60fps or 50fps. So 60/50fps would indicate 100% speed,got it? I'm sorry to be the bearer of truth...but DS is NO WHERE NEAR A REAL 100% on any game! If you must tack on a certain REAL percentage on DS,it would be closer to 12%-25% or about 7 to +15'ish fps depending on the game and your settings.

(Note:)These are my own benchmark numbers for the games I have tested in my personal collection crossed referenced with a real SNES and my PC with SNES9X v1.39.

So it breaks down like this: (I'll use NTSC for simplicity sake)

Real:
100% = 60fps
75% = 45fps
50% = 30fps
25% = 15fps
10% = 6fps

Frame skip:
100% = 0 of 0 frames
80% = 2 out of 10 frames
50% = 5 out of 10 frames
30% = 7 out of 10 frames
10% = 9 out of 10 frames

Frame skipping gives the illusion of faster gameplay by doing just that...skipping frames. In DS's case tho it's IMO to much,but its a neccessary evil to keep the emu speed up. Speaking of speeding the emu up...I don't see how DS is gonna get much faster without a bunch of/complete re-writes. Sure writing a SH4 (or even better) a ARM7 ASM CPU core for the SPC700 sound CPU will perhaps speed things up a bit more,but I'm sorry to say that DS will prolly never reach a full 60fps w/sound,without frame skipping. :( Altho,I'd love to be proven wrong! :D

With all that said,a full/near full speed SNES emulator is possible on the DC. I just don't believe that it'll come from Gary Henderson's source code. The mythical emu I speak of will have to be coded from the ground up to get any kind of real SNES emulation...and that my friends will be a lot of work.

Listen up ya'll! Don't for one second think that this post is some derogatory statement againist the DreamSNES team,quite the opposite. The DC scene and everyone in general (including me) owes the DS team a great debt of gratitude for all their hard work and accomplishments...and I mean that $h!t too!


Welp,I've taken up enuff of yer time (and prolly confused most of you! LOL!) so I'll end this post/rant with a few suggestions for the DS team (providing their reading this) on making the future DS even better!

1. Map button B on the main screen to play/pause mp3 playback...its a lil annoying hearing the song/s skip and sputter while loading the rom dir. Or set it up to pause when going into the roms menu and then resume after the list has loaded.

2. Give the users the options they want! Like adjustable sound freqs,adjustable frame skip levels,turning on/off transparencys,turning on/off auto frame skip,a FPS counter for the folks that don't have/want a KB. Basically,make it user configurable just like the PC ver.

and... :gulp:

3. Go open source with DS! There are several programmers lurking around that I think would be more than happy to lend a hand on the project.
mc
DreamSNES Creator
DreamSNES Creator
Posts: 114
Joined: Sun Feb 24, 2002 8:23 pm
Location: Link?ping
Has liked: 0
Been liked: 0

Post by mc » Sun Mar 03, 2002 8:21 pm

Mr.TA wrote: The way you folks are trying to benchmark DS is all wrong IMO. In the real world there would be no frame skipping at all. Since DS doesn't provide us with that option all we have is the emu telling us a percentage of frames displayed out of the amount of frames skipped.
Uh, no, that's not at all what the percentage says. The percentage
shows the relative speed of the CPU emulation (relative to a real
65C816 that is). I.e. it shows the ratio

number of clock cycles processed by the emulated CPU
----------------------------------------------------
clock frequency of real SNES * amount of time

for some amount of time (currently about 1/4 of a second).

It relates only to CPU emulation, not graphics processor emulation, so
there is no information about frameskip in there. Rather, it answers
questions like "Does Dixie Kong climb up that rope at the correct
speed right now, or should she be moving faster?"

So a 100% figure means that the actual game progresses in the same
speed as it would have on a SNES. Nothing more, nothing less.

This is the concept usually understood as "speed"; the distance
traveled in a certain amount of time. Not exactly how many
intermediate points were passed on the way.

(Low frameskip is of course also nice, but the emulator prioritizes
correct speed over low frameskip, as this tends to result in the best
playability. Playing games in slo-mo is no fun.)

I hope this clears up the confusion.
Flying at a high speed
Having the courage
Getting over crisis
I rescue the people
Mr.TA
DCEmu Cool Newbie
DCEmu Cool Newbie
Posts: 15
Joined: Tue Feb 12, 2002 8:01 pm
Has liked: 0
Been liked: 0

Post by Mr.TA » Sun Mar 03, 2002 11:30 pm

Indeed it does mc. But other than testing/debugging the CPU core it doesn't serve much of a purpose to the casual end users,and has little to no relevance to "in game" speed benchmarks.I may be wrong (its been known to happen) in the end when all is written and compiled every emulator is judged by its "in game" speed. Not neccessary how well or close the program emulated the CPU/s.

Example & Quiz - Dave's Dgen,you can select M68000 cores: StarScream (fast) or the other one. HA! I forgot the name (accurate) which one do you think most people used?

If you guessed StarScream your a winner,and yer prize is......ummmmm,my first born! You can come and pick him up anytime after 9am. :D

BTW,he likes the PPG!
mc
DreamSNES Creator
DreamSNES Creator
Posts: 114
Joined: Sun Feb 24, 2002 8:23 pm
Location: Link?ping
Has liked: 0
Been liked: 0

Post by mc » Mon Mar 04, 2002 7:41 am

Mr.TA wrote: Indeed it does mc. But other than testing/debugging the CPU core it doesn't serve much of a purpose to the casual end users,and has little to no relevance to "in game" speed benchmarks.I may be wrong (its been known to happen) in the end when all is written and compiled every emulator is judged by its "in game" speed. Not neccessary how well or close the program emulated the CPU/s.

Example & Quiz - Dave's Dgen,you can select M68000 cores: StarScream (fast) or the other one. HA! I forgot the name (accurate) which one do you think most people used?

If you guessed StarScream your a winner,and yer prize is......ummmmm,my first born! You can come and pick him up anytime after 9am. :D

BTW,he likes the PPG!
Of course the meter is intended primarily for testing, that's why it's not documented.

However, It still seems that your definition of "in game speed" differs from most other peoples.
If a sequence that's supposed to finish in 17 seconds finished in 17 seconds, then most people would
consider the emulation to have the correct "in game speed".
Flying at a high speed
Having the courage
Getting over crisis
I rescue the people
Mr.TA
DCEmu Cool Newbie
DCEmu Cool Newbie
Posts: 15
Joined: Tue Feb 12, 2002 8:01 pm
Has liked: 0
Been liked: 0

Post by Mr.TA » Mon Mar 04, 2002 9:58 am

Let me see if I can try to clear this mess up...

The benchmarking that I'm refering to is the what you see/play being emulated on DS verses what a real SNES does. Hmmm, I'm guessing that your talking about a CPU timing benchmark?
mc
DreamSNES Creator
DreamSNES Creator
Posts: 114
Joined: Sun Feb 24, 2002 8:23 pm
Location: Link?ping
Has liked: 0
Been liked: 0

Post by mc » Mon Mar 04, 2002 10:07 am

Mr.TA wrote: Let me see if I can try to clear this mess up...

The benchmarking that I'm refering to is the what you see/play being emulated on DS verses what a real SNES does. Hmmm, I'm guessing that your talking about a CPU timing benchmark?
I'm talking about the percentage figure in the lower right corner. Which displays the speed of the
emulation, relative to the speed of a real SNES.
Flying at a high speed
Having the courage
Getting over crisis
I rescue the people
spag
DCEmu's #1 Cher fan
DCEmu's #1 Cher fan
Posts: 214
Joined: Mon Feb 04, 2002 12:17 am
Location: New York
Has liked: 0
Been liked: 0

Post by spag » Mon Mar 04, 2002 1:17 pm

you mean mario world plays at 100%? it seemed to be like 10-14 fps to me (10 being the lowest it ever got, 14 being the best).

maybe we should change all the benchmarking to frames-per-second?
User avatar
|darc|
DCEmu Webmaster
DCEmu Webmaster
Posts: 16189
Joined: Wed Mar 14, 2001 6:00 pm
Location: New Orleans, LA
Has liked: 25 times
Been liked: 6 times
Contact:

Post by |darc| » Mon Mar 04, 2002 2:35 pm

this topic is now stickied

for more compatibility info go here
http://www.dcemulation.org/phpBB/viewtopic.php?t=4155
im gonna unsticky that one, this is a better way of testing
It's thinking...
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

Post by 404NotFound » Mon Mar 04, 2002 3:14 pm

Once i download these last 23 roms i'll have a whole CD worth and will get started on testing...
Last edited by 404NotFound on Mon Mar 25, 2002 7:15 pm, edited 1 time in total.
mc
DreamSNES Creator
DreamSNES Creator
Posts: 114
Joined: Sun Feb 24, 2002 8:23 pm
Location: Link?ping
Has liked: 0
Been liked: 0

Post by mc » Mon Mar 04, 2002 4:16 pm

spag wrote: you mean mario world plays at 100%? it seemed to be like 10-14 fps to me (10 being the lowest it ever got, 14 being the best).

maybe we should change all the benchmarking to frames-per-second?
Do I really have to explain this one more time?

Ok, I'll give you an example. Let's consider a hypothetical game Super Mario Rollerskating. In the intro, Mario rolls past on the screen on a pair of rollerskates. This is an NTSC rom, so on a SNES it runs at 60fps. The Mario sprite is advanced by 10 pixels each frame. So it will take about 25 frames for him to reach the other side of the screen (that's below half a second, he's skating quite fast).

So, if we count time in 60'ths of a second, we can plot the displayed position of Mario on the screen thusly:

t x
- -
0 0
1 10
2 20
3 30
4 40
5 50
6 60
7 70
8 80
9 90
10 100

and so on. Now imagine that playing the intro of this game in DreamSNES requires a frameskip of 2 to run at full speed. This means that every third frame is actually rendered and output to the screen by the graphics engine. However, the CPU emulation still advances at the correct speed, so in this case the position of Mario as viewed on the screen will be

t x
- -
0 0
1 0
2 0
3 30
4 30
5 30
6 60
7 60
8 60
9 90
10 90

and so on. So in this case the speed is 100% (meaning that Mario will reach the end of the screen at the correct point in time), but the framerate is just 20 fps (or 33% if you like).

(Note that I have no idea what speed "Mario World" runs. I'm just trying to clear up the confusion about speed and framerate, hopefully for the last time...)
Flying at a high speed
Having the courage
Getting over crisis
I rescue the people
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

Post by 404NotFound » Mon Mar 04, 2002 8:06 pm

Updated. Pleaes refer to top for new info.

Edit: Any way i can keep this stickied and modify it?
Last edited by 404NotFound on Mon Mar 25, 2002 7:15 pm, edited 1 time in total.
kRYPT_
DCFactory Creator
DCFactory Creator
Posts: 597
Joined: Mon Nov 12, 2001 7:58 pm
Location: Ontario, Canada
Has liked: 0
Been liked: 0
Contact:

Post by kRYPT_ » Mon Mar 04, 2002 9:46 pm

I don't think so, not unless you're a mod.
However, I've re-stickied it for you.
404NotFound wrote: Updated. Pleaes refer to top for new info.

Edit: Any way i can keep this stickied and modify it?
404NotFound
DCEmu Ex-Mod
DCEmu Ex-Mod
Posts: 4970
Joined: Thu Nov 29, 2001 3:40 pm
Location: The Canadian-Mexican border.
Has liked: 0
Been liked: 0

Post by 404NotFound » Tue Mar 05, 2002 2:09 pm

k cool, i'll just keep reposting it instead of editing it.
Last edited by 404NotFound on Mon Mar 25, 2002 7:17 pm, edited 1 time in total.
Post Reply