A common thing that I read on the forums is how much faster one rom is over another. Someone will claim that their rom runs much faster than others or just users will say that this one is better than another. One of my very first contributions to XDA was benchmarking every rom that was available for the HTC Mogul at that time. I decided to do the same thing for all HTC Hero roms based on the Sprint 2.20.651.1 leak. This consists of: DamagedCake 1.2b, DamageControl v2.05, DarchVanilla v2.1, Fresh Rom 2.0d, RegawMOD v2.0.2p, and untouched Stock.

Now first things first I have a statement to make. I chose to do this because I saw some people bullying others on the forums for making their own roms (I’m looking at you Regaw). I set out to run these tests to prove that every single rom based on this build would benchmark very close to the same. This means that when it boils down to it, there is basically no merit in saying that one rom is faster than another. It is simply a matter of needing to pick which one has the options already in it that you prefer. The modifications that we do to these roms will NOT dramatically speed up the core functions of the phone. It would require kernel development to actually make a real difference in the speed. Yes we can possibly make things run a little smoother and add features, but when it comes down to it, they will all be the same speed. Let me try and make my point here because I’m rambling…. I have been making roms for the Hero since November and have come to expect other people to work off of my work. It’s exactly how this stuff works and is exactly what we are doing to HTC. But people typically give credit where credit is due. Then you get some people that like to go around bashing people for making a rom, yet steal work from other people without as much as a thanks. Funny again that I am talking about Regaw. You may notice that he released a patch for his rom 2.0.2p with no release notes at all as to what it fixes other than a vague mention of wifi tether. So what’s in this rom?

  • A copy of estrongs file explorer, straight from fresh rom. Not my app at all, but still something that I have been including in fresh forever.
  • A copy of AutoKiller, straight from fresh rom. Again, not my app but something that has only been in fresh rom
  • A copy of the config file for the new XDA keyboard included in fresh rom with English and Spanish being the only languages enabled… taken straight from fresh rom
  • An “sdcard” folder containing above mentioned config file, taken straight from fresh rom so that it matched the line taken from the update-script.
  • The fix I used to make all apps show up in the market in build.prop. This fix doesn’t belong to me but the way I entered it shows it was likely copy/pasted from fresh rom.

That is the entirety of the patch. He also uses my ported Espresso artwork, but that has been taken by so many roms without credit that I gave up on that. Look… I don’t have a problem with regaw. And I’m really not trying to bash. My whole point of this was that every single person is entitled to make every rom that they want. So when he goes and gives someone crap for making their own rom… it really bothered me. Coincidentally, when I expected every single rom to perform about the same, his rom performed the worst on both benchmarks.

Sorry for the rant. None of this is related to the benchmarking. So let me get down to the details. Just like a dyno on a car, benchmarking is not a real world test. While the tests were consistent and what they show holds merit, I’m not saying that it doesn’t mean another rom can’t run smoother. But my point here was to prove that no rom is actually dramatically faster than another.

My method:

  • Factory wipe
  • Flash rom
  • Boot phone up and install Benchmark Pi and Neocore. Leave phone plugged in to usb and enable “stay awake while charging”
  • Reboot and allow phone to boot up entirely (waited for activity to stop in debugger)
  • Run Benchmark Pi 3 times and take the average of all 3
  • Reboot and allow phone to boot up entirely (waited for activity to stop in debugger)
  • Run Neocore 3 times (with the sound disabled) and take the average of all 3 (the only exception here was on DamagedCake. While running Neocore’s 3 tests there was 1 test that was significantly lower than the other 2. I considered it a fluke and discarded it’s result. I then ran it a 4th time and used that instead.)

And now I present to you the results. The “winning” rom is highlighted in red. But again, my goal was to prove that basically every rom is just about the same. Note that Benchmark Pi specifically only tests the processor where as neocore tests a lot more.

Well there you have it. Hopefully this ends the “this rom is faster than this rom” debate, and also makes some people realize that they were acting a bit rude. Pick the rom that is customized the way you like it. Or at least close. Don’t like any? Then make your own!! The more the merrier! And if you use someone’s work, make sure you give them credit!

Excel file containing all results