Load times
6 years ago
Texas, USA

What's this thing about taking 13 seconds off your time because of "load times"? What are the requirements for being allowed to do this? And shouldn't it be added to the rules?

Uusimaa, Finland

Homebrewed 3ds loses 13 seconds to loads in any% so thats why. It should probably be explained somewhere

Edited by the author 6 years ago
United States

Yeah i never got a chance to explain it anywhere, so lemme do it now.

So, I homebrewed my 3ds recently, and when I went to do runs, I realized I was losing time to nothing. So I decided to investigate. Comparing my PB video to footage from the homebrewed ds, I was definitely losing time to loads. So I tested every load. I lost exactly 17 frames every time I entered a level or finished a level, and I lost exactly 4 frames every time I used a pipe or door or something else like that (the game runs at 60fps). Adding up every load in Any%, that equated to a consistent 13.05 seconds lost every run to nothing. So in the Discord we pretty much agreed that for homebrewed systems, we would allow load removal to make the times fair. In the whole experimentation, I got curious to see whether other things affected load times. So I tested them. Comparing one of my old runs on the new3ds to my 27:24 on old3ds, it lost no time to loads. Comparing eddaket's run on cartridge to my 27:24 on digital, no time was lost to loads. That also proved no time was lost to loads just based on different systems (like different wiis for NSMBW). So I concluded that time is only lost to loads when the 3ds used is homebrewed, the term "homebrewed" meaning the system has arm9loaderhax and luma3ds installed, and is running on a separate emuNAND (it's complicated I know).

Hopefully that wasn't too confusing and it cleared things up.

Uvideo likes this
United States

Alright, I made a pastebin and added a link to it in the rules for every category: https://pastebin.com/fXM0sBEn

Texas, USA

Just read through it, thanks for clearing it up!

asheevee_ likes this
Navarre, Spain

One question, if you have it homebrewed but you also have a physical copy of the game, does it affect the loading zones too???

United States

physical vs. digital makes no difference, if it's homebrewed it loses time no matter what version of the game you're on or the system you're on.

United States

I may have made a mistake there -- watching your PB video is showing me that the loads are on par with the non-homebrewed systems. I don't know if that's because you don't have homebrew or if it's because cartridge.

United States

What could also be happening is you're losing time to loads, but the NTR CFW is actually making up for the lost time by overclocking the system a bit. I guess more investigation is needed.

Navarre, Spain

I have arm9 and luma and I'm using a download. I will try tonight playing without the homebrew to see if there is any kind of change so we can test all correctly :) Thank you Eevee

asheevee_ likes this
United States

UPDATE: A new revolutionary homebrew method has just been discovered, and I HAVE NO IDEA WHETHER IT AFFECTS LOADS. welp time to do all that testing again will update the pastebin with any discoveries as i find them. eventually i will remove anything pertaining to a9lh from it because it is now obsolete and if you still have it go get boot9strap at https://3ds.guide/updating-to-boot9strap