Translating...

Cancel

Ricardo Cerqueira
2 years agoEditedPublic
I should've learned by now that sarcasm doesn't translate well on the internet :)

So to be clear... IMNSHO, the recent entropy pool fad 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 conceivable reason some devices may 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 and/or ramp up the IO scheduler).
42
Florio Alagna2 years ago
Is it actually possible to miss the sarcasm in your previous post? If you replaced the whole thing with a size 78 SARCASM print, it'd possibly become less obvious lol.+2
Fabio Lo Brutto2 years ago
+Florio Alagna they probably skipped the second part of the post.
most of the time people don't read the all thing...
+2
Florio Alagna2 years ago
It was pretty evident at triple wipe already, but some people probably really do that haha.

Who doesn't get it, deserves the public shame following their eventual rant.
+2
Solomon Seal2 years ago
+Florio Alagna So many "guides" directly from the cooks (not developers) of the copy and paste roms say that you must do multiple wipes or it won't work. 

I don't even bother reading the rest of that thread, skip to someone with reasonable (accurate) install instructions. Then maybe they know what they are doing with the rom.
+1
Harsha Ramesh2 years ago
+Fabio Lo Brutto .. the second part was the fun part =D ... Now I want to go paint fire stripes on my car and take it to the races =D+2
Gaurav Pandit2 years ago
Was just about to search cyanogen community to see if anyone mentioned that thread. Thanks for clarifications.

Also, a dedicated /sarcasm tag would have made html5 more useful.
Nicholas Nezis2 years ago
Once Comcast starts putting out their own roms, you'll have to unplug the device and battery for 30 seconds. You know... to be sure.+6
Steven Harper2 years ago
Actually once I have a RIL crash on my i9100 I have to pull the battery to get the modem back.+1
Clive Neo2 years ago
Yes, I've found out to my cost that sarcasm doesn't translate through a screen, btw, i haven't got a bloody clue what your talking about in ur post lol :-)
Fabio Lo Brutto2 years ago
the worst thing(as always) is that this new placebo effect has spread through every tech\android site like it's the cure of every disease known to humankind.
It'll probably persist for a few years in the mind of people.
Rejoice! a new "thing" has born
+7
Jeremy Thornhill2 years ago
Thanks +Ricardo Cerqueira , I'm glad to hear somebody smarter than myself confirm my suspicions about this fad :)+3
Andrea Pergola2 years ago
+Ricardo Cerqueira after reading your post I was curious so I downloaded the app on my xoom and found that it makes the cpu stay from 400 to 1000 mhz constantly. More or less it does the same job of performance governor (actually it does it better as the cpu is always at highest clock). Didn't tested for more than half an hour, but I guess the battery would drain faster..

Thank you for these posts, I read about this app 2 or 3 days ago, but never installed suspecting it was not so miraculous (or cm would have included it since November ;)). Btw your sarcasm in the previous post was pretty clear also for me without having so much android/linux/coding knowledge... maybe common sense should be enough :)
+1
Douglas Jenkins2 years ago
But, but, but, faster is ALWAYS better, eh? Thanks for the heads up, +Ricardo Cerqueira 
Pedro Carneiro2 years ago
+Ricardo Cerqueira  you probably already seen this: 

https://code.google.com/p/android/issues/detail?id=42265

Quote from dalvik maintainer about people buying the magical app:

"@21: people who've just paid $1.49 are unlikely to say "i'm an idiot"."

Hahaha! Gotta love Google! :)
+4
Steve Kondik2 years ago
I don't care what you say Mr. Arcee because its faster and you must be wrong!+13
+Steve Kondik Screw you Mr cyanogen, that's just the result of ro.makeitgofaster=true. You clearly don't understand this Android stuff, did the post hit "What's Hot" or something?+16
Derek Ross2 years ago
+Ricardo Cerqueira It damn well needs to. Yesterday we had 4 different make Android more awesome posts saying how good Seeder was on there.
Tim Skorick2 years ago
Hahaha I can't tell if you're talking about a mobile app or magical speaker cables.  "Hey if the difference is PALPABLE I'm buying it now man!"
samuele tonon2 years ago
+Ricardo Cerqueira actually at the monent it still not clear what's happening on the random/urandom thing.  it appears it was a know bug til gingerbread but people are still investigating the issue you can see here
https://code.google.com/p/android/issues/detail?id=42265

i'm quite curious on how it will end
Adrian Ulrich2 years ago
+samuele tonon It's perfectly clear what is happening: The 'uber fix' simply causes your CPU to stay at a higher clock frequency. That's it.  Running a mild cpu-hog would to exactly the same.
And apart from that: Feeding /dev/urandom info /dev/random is just STUPID.
Randy Magruder2 years ago
So what's causing the pool to be depleted? Or is the monitoring itself not giving accurate results?
+Randy Magruder Linux ASLR, which is enabled from 4.1 onwards. The pool isn't expected to grow indefinitely, but nothing in regular Android userspace uses it at a rate that will cause it to fully deplete. And even if it was depleted, it would still not impact things like navigation and drawing.
Randy Magruder2 years ago
Well, I'm a Windows developer so I won't claim to know what I'm talking about.  I'm kinda learning as I go here.  I guess what got me was the discussion about userspace random pulling (indirectly) from dev/random by way of dev/urandom.  That and the comments about thread locks in the implementation of dev/random.  I know you can seize up with too much locking, but I guess it comes down to how often ASLR is requesting new random numbers.  Of course, it doesn't explain supposed speed increases on older versions of Android, but who knows.  Seems like all this is easily tested, though, by changing the threshold values for the entropy pool, and doing nothing else (e.g. no waking up the CPU every second).  How often is Linux ASLR pulling random values for depositing items in memory in a random way?
david kantor2 years ago
Fact is the internet says it works, xda says it works, I won't even run my oy test, I just will download and know it works.
If my phones don't seem faster all the time I will blame Google for some flaw in maps.
I will not look back to this thread for insight as I read that it doesn't work after I read that it does.
First thing i read says its good, so anything sales I read later must be written by apple fan boys and therefore not trustworthy
grant borden2 years ago
Im a p999 rooted user. We have on this device what is refered to as memory leak wich causes lag. Ive tried a few "fixes" like v6 and one other script that was simply called "lag fix". RC i know damn well that you know a world and a half more abt this stuff than i do, and i should know better than to post this cuz ill probably be laughed out of here but for some reason(probably the reasons posted above) this app wich i didnt buy from the market but got off xda for free works for my phone. After reading your post i felt like an idiot for hittimg thatnk on the Seeder thread. Plz dont burn me here lol i just wanted to give some user input.
Ryan Brown2 years ago
+grant borden I've had a stock G2X (p999) since it launched, and it does indeed have some bad memory leak problems. It will eventually get itself into a situation where there's about 30 MB of RAM free and it thrashes constantly until a reboot. My wife's phone has the exact same issues. I assume it's a known issue and LG just doesn't care to fix it. That's par for the course with them.

That said, I doubt that that running an app that fiddles with a random number generator will have any appreciable effect on RAM allocation.
Craig Andrews2 years ago
There are many users who are using the "replace /dev/random with /dev/urandom" method (rm /dev/random && ln /dev/urandom /dev/random). This method doesn't cause the CPU governor/usage to change at all, and these users report performance improvements.

I'm not saying that this isn't a placebo situation - I just think there's more research to be done, and more data to be gathered, before anyone should go about broadcasting blanket "not so humble" opinions.
+2
Randy Magruder2 years ago
The dev who started all this on xda has posted an update:
-------
Hey guys,

Just wanted to say a few things -

1. My initial analysis (dalvik reading /dev/random) does indeed not apply on ICS, Jellybean, etc.
2. Nevertheless, it seems that rngd does have a significant effect on these builds for some people, including me.
3. I'm going through the kernel's RNG code right now trying to figure out if perhaps there's some lock contention, or something else funky going on. I have entropy debugging turned on, and there are a lot of nonblocking reads going on, constantly.
4. Perhaps it is just kicking the governor, but I don't see any change to time_in_state with it running. I suspect it really has something to do with the input entropy pool.

I never expected Seeder to take off like this; it's a little scary. I initially published a build, and had very positive feedback on XDA, so decided to put something on the marketplace. Though I've been a *nix admin and coder for a while, this is my first Android app.

It doesn't seem to work the way I thought it did on later Android builds, and I will get to the bottom of this.
------

To me personally, the million dollar question is: Who is doing all these non-blocking reads?
+3
Sara Oliveira2 years ago
I have one word for you: placebo+1
Randy Magruder2 years ago
Sara, I bet you haven't even tried it.+2
Sara Oliveira2 years ago
Randy, I was talking to Ricardo ;-)
I didn't use the word "placebo" on purpose. Even though I can't reproduce the effect, even on shitty phones, I believe people are getting it. (although there's probably a lot of placebo involved, since "smoothness" isn't exactly something that can be quantified) It's the logic/explanation behind it that doesn't make sense.
Gaurav Pandit2 years ago
Has anybody tried to see if it affects CPU in any way? I am just an Android user, and not a developer, but if its overclocking the CPU, it might make some things faster. Just a wild guess.

PS: Never mind. Just saw few comments above mine, and someone already mentioned it.
Micha Ols2 years ago
+Fabio Lo Brutto I fear you are right.
Need to wipe my battery stats 3 times now, because my WiFi signal is weak again.
+4
Craig Andrews2 years ago
It seems that Steve Kondik disagrees with all of the blanket "placebo" statements and is doing some homework instead of non-productively blowing off the issue like most comments here. See https://code.google.com/p/android/issues/detail?id=42265#c114+4
+Craig Andrews Yes, and like I told +Steve Kondik 15 minutes ago, his tests don't match the claims being made. AFAICT, feeding the PRNG through ioctls won't unblock existing input/blockdev contention. He may have stumbled across an entirely different issue.+3
Martin Brabham2 years ago
Damn, and I've already added a spoiler to my device and switched my ringtone to a turbo spooling.  I'm also seeding /dev/random from my accelerometer so when my phone is slow I just shake it to go faster.+3
People how doesn't now shit nothing about this talking "It a CPU overclock thing" BULLSHT 
+Francisco Enrique Pérez Abreu It's more complicated than that. But it's still bogus :)+1
Edvin Slama2 years ago
Well, I ned a 2 by 4 for my phone.It's kinda slow.
Sándor Mihók2 years ago
Hmm, CM developers say, its a hoax, because they implemented this guy's code in their kernel? Yep, its not a hoax, its a bullshit. Dont steal this guy's work...+1
Alex Grig2 years ago
So how could you Ricardo explain this: with enabled seeder I can export projects from Movie Studio while on the phone call or while screen is off. Without seeder Movie Studio stop export immediately!!! Will appreciate your explanation.