Shared publicly  - 
 
Quick guide to a faster, stabler Android:  triple wipe, clear dalvik, reset battery stats, format /system regularly, limit the clock when screen is off, use random schedulers (the less characters in its name, the better), amusingly-named CPU governors (the more your 8-year old giggles at its name, the better), and top it off with an off-spec LCD density for improved app compatibility. And now the new hotness, use the less-than-random non-blocking PRNG to mess up the pool used by cryptography and apparently nothing else.

Equivalent guide for a faster, stabler car: Paint flames on the side doors, and hammer a 4x2 plank to the trunk's lid.


Screw this, I'm doing cars.
399
64
hermes 2xgreat's profile photoCristian Garjy's profile photowaldek andrukiewicz's profile photoSébastien Clermont's profile photo
73 comments
 
Finally some serious guide! Full data wipe also? ^^
 
LOL!!! --> "The less characters in its name, the better"  .... =D
 
You're forgetting speed holes, removing the muffler, compressing the shocks, installing Bose speakers, adding political bumper stickers, using lots of the colour red.
 
If you were to package this in an app and charge $.99...
 
I heard about that. How come most people experience improvements? The louder the muffler the faster you go theory?
 
+Jeromy Layne The louder exhaust creates a sonic boom in the back of the car, which helps propel it forward. It's a cheaper variant of the Orion Project
 
And I thought that only by attaching a Cavallino Rampante sticker it would go faster :(
 
Over here the wide 4" exhaust means you can hear the little boy racers from 5 miles away......they love that!

To me it makes the car sound like a tractor lol
 
Haha, (stomach hurt). So aside from the jokes why exactly is it useless?
 
When I first read the thread on xda, I was thinking - 90% of all lag removed by such a simple thing? How the fuck would the android devs, working on butter, NOT address this?

At least his app helps to empty the battery faster.
 
So, jokes aside, the non-blocking pool, whilst less secure, does actually speed up the device and reveal an issue that needs to be addressed, does it not?
 
You stupid newb; real pros know that scheduler performance is inversely proportional to the number of VOWELS in their name! Judging them by the overall number of characters is idiotic.
 
+Vito Cassisi , +Jeromy Layne  The entropy pool stuff is bullshit. The only users of /dev/random are libcrypto (used for cryptographic operations like SSL connections, ssh key generation, and so on), wpa_supplicant/hostapd (to generate WEP/WPA keys while in AP mode), and the libraries that generate random partition IDs when you do an ext2/3/4 format. None of those 3 users are in the path of app execution, so feeding random from urandom does nothing except make random... well... less random

The only reason devices feel faster is because by constantly polling the PRNG, it keeps the device's I/O in constant use (which in turn, depending on device, will make the CPU stick to higher clock frequencies to keep up)
 
It's XDA again isn't it? It's always XDA...
 
+Ricardo Cerqueira I haven't had time to look into it, but wasn't part of the issue that apps which didn't require good (for lack of a better word) random numbers were requesting them, thus depleting the pool?
 
+Vito Cassisi I can come up with tons of plausible techno mumbo-jumbo to justify almost anything, it doesn't make it true :) Even if that was the case, it'd only affect those apps, since there are no other users of /dev/random
 
+Kakashi Hatake wanna talk about PEGASUSQ?
Yo dawg, be a smartass all ya want, I got a fucking horse with wings managing my CPU.
 
with lg phones you shouldn't leave the magic power of RILs unmentioned....
 
When's CM10.1 going to be ready for my Renault Scenic? 
 
+Gaurav Pandit to be fair, some people on reddit cleared up that myth pretty fast in that topic.
 
+Ricardo Cerqueira, when moving to cars you'll get to experience the fantastic argument of torque wrench vs. the German tradition of "gutentite". 

Yeah. Pretty sure it's the same people.
Rob M.
 
Are we developers now?!
 
Oh that reminds me.. I need to stock up on blinker fluid and muffler bearings. 
 
+Ricardo Cerqueira you forgot to list "install the most amazing and expensive task killer to give you 90% of free memory all the time"
 
+Pedro Carneiro I haven't seen people recommending task killers in quite a while, so I suspect that particular trend is/has died down. There'll always be people complaining about free RAM, but it isn't nearly half as bad as it used to be.
 
You mean the Tornado I put into my throttle body and the magnets on the fuel lines don't help? What about the octane booster in my naturally aspirated 4 cylinder? I mean, it has The Tornado installed so its almost a turbo!
tobby o
 
I only have one word for all this: NOS
 
Not sure about v6, but v8 seemed to help my p930 quite a lot - or it did on ICS. 
 
+tobby olupot Real NOS actually works. Not cheap or efficient though. NOS stickers on mom's car, OTOH, don't.
 
I should follow your suggestion... I will port suckysuck governor to djangomanouche and "no" scheduler to avoid phone boot... I think that a turned off device will works great for battery drain... Carburano (a never dev but a utopistic user) 
Rob M.
 
+Alvin Brinson We so need a NOS governor and a Vortex Generator scheduler!
tobby o
 
+Alvin Brinson according to many testimonials, people are getting an unprecedented boost in performance. Maybe it's more like VTEC kicking-in.
 
Amen. I'll never understand why people "dirty flash".

As for the car... I'll stick with the Blouch 20G and supporting mods. 
 
Don't forget STP! You gotta add STP! 
Rob M.
 
Airboy tuned my car, CM tunes my phone, I remain nearly clueless about the rest.
 
Thanks a lot. I was seriously considering rooting my n7 .
 
I almost blew a gasket when I initially started reading this. Haha. 
 
You forgot to add random letters to device naming schemes such as the HTC EVO 4G LTE 2 or Motorola Droid RAZR MAXX HD 2.
 
+Micha Ols Yes, they did. But majority of them have seen the amazing placebo effect already.
 
At first I started thinking "does he who is a good developer actually buy in to this crap?", but soon realized the sarcasm. Had a good laugh. :)
 
To incease quickness to your vehicle, add lightness.  It's a simple Colin Chapman message, that starts with the driver and ends with the seats and soudproofing removed from the car.  Live fast, but not faster than the Honda race car in the fast lane that you passed on your way to work today.  Speed fast.
 
In all honesty, I'm quickly and sadly realizing that on many phones and tablets stock is the fastest and cleanest way to go.  Both my tablet and gnex are slower than their stock counter parts when rooted, rom'd, and overclocked.   I would have said they were faster until I did side by side tests and lost pretty much every test I could think of...  it really doesn't seem to matter what rom, kernel, etc... the stock wins pretty much every time.  
 
+Ryan O'Neill Yep! Noticed pretty much the same thing. People look at me crazy when I tell them I just run stock rooted with a few changes.
 
+Ryan O'Neill Right on. And that's why I have decided, after a long winding journey, to stick with Nexus series, and use CM if I feel itchy.
 
This is for the most part bad info or one big joke (I'm praying for the later of the two). Sounds like a bunch of crap.
 
Although keeping a clean internal SD card actually helps
 
Actually, it's not really that you should use less characters in a name. It depends on what the characters are. Since data is stored in binary format as a series of 1's and 0's, you should choose names with characters that have more zeros than ones because everyone knows that zeros are nothing and so they don't take up any space. Hehe.
 
GSam estimates my battery life at well over 90 hours with the screen off for some time on LeanKernel vs no more than 35 with CM's kernel.  Idle time battery life can be significantly improved without impacting screen-on performance. My bus ride to work is the same length of time every day and I get to work with another 7-8 percent of battery left using a non-stock kernel vs a stock one on my GNex.
 
Why not creating a page in CM's wiki with these myth busting facts properly explained? It's a good service to the community, especially because many users might ruin a CM experience with crappy mods or kernels and then blame CM.
 
 "limit the clock when screen is off"
I'm always doing that. am I a bad person now?

I know on x86 you Race to idle is what you want but I don't think it behaves just like that on a phone.
When my n4 runs at 1,5 ghz in my pocket, it get's hot. Getting hot means draining the battery, doesn't it?
It takes a lot more power on full speed than on a few hundred mhz and I think this power loss on top speed far outnumbers the time it's not sleeping.
Just like 4 cores on 25% are far more efficient than one core on highest clock.

Found this neato thread about this.
http://lwn.net/Articles/281629/
First commet seems about right:
"The dynamic power consumed by a CMOS circuit is proportional to the _square_ of the voltage, while the clock speed is proportional to the voltage itself. Leakage power is constant. So, at least for embedded processors, like ARMs, where leakage power is low, voltage scaling can be a win over sleeping: you halve the speed and power drops by nearly a factor of four."


Then again I wonder why the phone even get's hot.
Ideally there would be no Apps running and then It would be bets to set highest possible speed for short tasks...

...

A cm wiki mystery team would be awesome!!
 
Yeah turns out the prng thingy is placebo lol
 
Just kick back and have a drink.
 
Here's what I do to win @ androidz

Factory reset x2
Format system x3
Format data x3
Format boot x5
Wipe battery stats x5 (10x to be on the safe side)
Wipe cache 3x
Wipe dalvic 3x each boot
Add a comment...