• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [ROM][TH3ORYROM] PUR3AO5PTH3ORY V1.0 [AOSP/CM7] DKS Kernel (8/3/2011)

Ooh, new AOSP rom! I can't wait...but I have to...and I just got my current rom set up to my liking :mad: nandroid is a good thing :) Thanks!
 
May want to hold off on flashing for time being. Per the xda thread he'll be releasing a new version tonight with fixes (don't know what's wrong, haven't gotten that far)

Unreleated side note, I wish he wouldn't let his 3 year old name and spell his ROMs :rolleyes:
 
May want to hold off on flashing for time being. Per the xda thread he'll be releasing a new version tonight with fixes (don't know what's wrong, haven't gotten that far)

Unreleated side note, I wish he wouldn't let his 3 year old name and spell his ROMs :rolleyes:

I will be starting my download around 12:30 tonight at the earliest so it should be posted by then.

Related side note to your unrelated side note, he is just talkin' in 1334 (leet:rolleyes:) it shows he is a haxxor and cool...I personally like to...uh...over punctuate (when I am typing)
 
I will be starting my download around 12:30 tonight at the earliest so it should be posted by then.

Related side note to your unrelated side note, he is just talkin' in 1334 (leet:rolleyes:) it shows he is a haxxor and cool...I personally like to...uh...over punctuate (when I am typing)

I know it sounds stupid, but I've never wanted to try any of his ROMs because the titles are so foolish :o It's a good thing I have a profound love for AOSP or else I wouldn't touch this one either lol
 
I know it sounds stupid, but I've never wanted to try any of his ROMs because the titles are so foolish :o It's a good thing I have a profound love for AOSP or else I wouldn't touch this one either lol

I've yet to try CM7 because CM twice threw my OG droid in a bootloop. I haven't tried a theory rom because I don't like sense, it can have a stupid name and I will still try it (Froyo really?)
 
Tried it for like an hour. Every time I would try to crank the CPU to where I run thunderbread (mildly uv, 1.8 GHz) I would get freezes, even though I am on the same kernel. Never even got to installing the KAndeeyyshjkjh store or whatever. Need my phone tonight, and LT2.5 is running great for me.

Think I will stick with LT as my DD for now. It still has what I want--AOSP with lots of switches and levers (what't the point of f%$king with your phone if you can't f$#$ with your phone?).

I think some of these ROM's need a clearer mission statement. Here is my take:

CM7: be CM7 for the TB
OMGB: alternative pure AOSP for the TB (I guess?)
OMFGB: OMGB with a few goodies
LTB: CM7 with tons of easily configured under the hood control
PUR3AO5PTH3ORY: ? not really sure--his other roms seem to be about lot of flash-able flashiness, so maybe a flashier CM7? (Got to admit I want to steal the boot animation)
 
May want to hold off on flashing for time being. Per the xda thread he'll be releasing a new version tonight with fixes (don't know what's wrong, haven't gotten that far)

Unreleated side note, I wish he wouldn't let his 3 year old name and spell his ROMs :rolleyes:

dammit! i just downloaded it,lol...

the names are goofy(i wont pretend to understand them either) but i have become a droidth3ory fan.... his roms definately WORK. even running full blown sense they are fast,smooth and reliable. id go so far as to say if we didnt have AOSP yet,id prolly be falling back to a senseless th3ory rom eveytime i flash somethin and get tired of it :p th3ory is similar to me to incredibly re-engineered on the inc.

brizey,why dont you just try flashing whatever kernel you were running in LTB?

also,alot of his flashable flashiness is becasue he wants to get a rom to run as smooth and fast as possible before effing it up with eye candy. :)
 
Why overclock so much, do you really notice that much of an increase in performance?


Because I can, lol. And battery life is not much of a concern for me...working from home and all that. Plus I get 3000+ quad scores, which really helps to score with the chicks. Really, though, it is just because I can.


brizey,why dont you just try flashing whatever kernel you were running in LTB?


Well, I sort of am using it, unless theory mods the kernel. I have been using dream kernel...BUT...I had a reboot this morning with LTB, and I have not been using dream kernel very long. It is probably my settings...I am going to back off a bit on the OC. Something to consider is that the operation points are different. Theory uses the dream kernel, but not speedtweak. It uses vipermod instead, so I used vipermod to undervolt -25 mV. On speedtweak, I start with normal ("slightly" undervolted) then bump up the max cpu frequency. The frequency steps in the theory setup are relatively coarse. In other words, even though the kernel is the same, the actual operating points are not. And like I said, I got a reboot this morning. I am going to back off to 1.6 GHz with LTB and dream kernel for the day, and try theory tonight.

One thing is for sure...I can run 1.8 normal all day on Imo 3.5.8 with no issues. But my cpu is not a good undervolter. I get sleep death sometimes at -50 mV on Tiamat, and Imo freezes on the extreme undervolting settings. I get the best quad number per clockspeed with dreamkernel.
 
after a day with this im impressed. its incredibly smooth and fast,even at stock speeds(im pretty boring when it comes to OC,UV),battery life was decent,data/gps/bluetooth/etc. all worked fine.

i did have some minor issues with 1.1 and titanium backup,but in all fairness idk for sure it was a titanium issue. i restrored cm7 and redid some backups before flashing 1.1.1 and it worked fine there :)

i had no data during setup,wich was a little freaky... but it appeared and h as been fine after sync and a reboot.

only other thing is pretty minor... i had to check "automatic"in settings/time and date and reboot to get my time to be right.

one more observation that you may like or may not... the keys on the keyboard are huge. while i do have fat thumbs,id personally prefer them to be a lil smaller so that the test window is bigger when typing in landscape mode.

i like it :D ill give it a few more days,at least
 
Because I can, lol. And battery life is not much of a concern for me...working from home and all that. Plus I get 3000+ quad scores, which really helps to score with the chicks. Really, though, it is just because I can

1.4 is good enough for 3k+ scores. I've noticed anything above 1.4 and the performance isnt noticeable and actually slows down over time.
 
Because I can, lol. And battery life is not much of a concern for me...working from home and all that. Plus I get 3000+ quad scores, which really helps to score with the chicks. Really, though, it is just because I can.

1.4 is good enough for 3k+ scores. I've noticed anything above 1.4 and the performance isnt noticeable and actually slows down over time.

This is why I don't run quadrant at all. Too many factors that are not accounted for in quadrant plus too many things you can do to get artifically high scores. I usually do not mess with OC until I notice something that stutters, then I usually go with the 1.4. I've never seen a need to try to push the processor in this phone to work almost twice as fast as it does stock
 
1.4 is good enough for 3k+ scores. I've noticed anything above 1.4 and the performance isnt noticeable and actually slows down over time.

Not for me its not on any kernel I have tried. Even at 1.6 I usually get 2800, 2900. That goes for Imo, Tiamat, and dream. I never even got close on sense ROMs.

(Edit: and that inlcudes BAMF and its DB cheat.)
 
This is why I don't run quadrant at all. Too many factors that are not accounted for in quadrant plus too many things you can do to get artifically high scores. I usually do not mess with OC until I notice something that stutters, then I usually go with the 1.4. I've never seen a need to try to push the processor in this phone to work almost twice as fast as it does stock

Lol, philosophy crash. I over clock everything I can. I have not driven an unmodified car since I was a teenager. I like AOSP roms because they are faster and offer more control under the hood.

I agree with Quadrant not being a very good benchmarking tool. Antutu (or whatever it is called) is even worse, IMO. It assigns a huge chunk to SD R/W speed, which has little bearing on day to day performance.
 
Lol, philosophy crash. I over clock everything I can. I have not driven an unmodified car since I was a teenager. I like AOSP roms because they are faster and offer more control under the hood.

I agree with Quadrant not being a very good benchmarking tool. Antutu (or whatever it is called) is even worse, IMO. It assigns a huge chunk to SD R/W speed, which has little bearing on day to day performance.

To each his own, while I don't see the need for it personally I do understand that people like to do it. I have had my trial and tribulations with overclocking, well one and I was young and dumb at the time, when you are overclocking and don't know what the eff you are doing, the smell of melting plastic is not a good thing, just glad my parents believed it was a lightining strike that fried that CPU :D

Overclocking on the phone is much safer now as it has checks and measures built in where they should turn off on their own if you exceed a certian temp
 
Not for me its not on any kernel I have tried. Even at 1.6 I usually get 2800, 2900. That goes for Imo, Tiamat, and dream. I never even got close on sense ROMs.

(Edit: and that inlcudes BAMF and its DB cheat.)


That's kinda strange, but every phone is different and quadrant doesn't mean much in the grand scheme of things. CM7 with Imos kernel runs anywhere from 2800 to 3100 usually. Which shows how inaccurate it can be. But since you brought up cars, its like a dynojet. Is it as accurate as an eddy current dyno? No, but if you use the same car, same dyno, similar conditions its useful to show gains and losses.

For fun, on BAMF stripped at 1.4 I posted a 3600+ using the cheat script. Its in the 'let's see your screens' thread.

For even more quadrant fun, I ran this last night after reading this thread. CM7, Imo's 3.5.8, 1.4ghz, Ix gov.. no script manipulation..

snap20110806_012430-1.jpg
 
Just found a new benchmark that is CPU intensive: Linpack for Android. Just does a floating point test.

One thing I found was that smartass is not as good as interactiveX at ramping up quickly.

The test only lasts about 1.5 seconds, so it is not a good "burn in" test, but at least is doesn't include crap like R/W speeds that are not going to be very affected by tweaks.

I wish you could set it to run for an hour or two and that it had a cpu temp monitor. Actually, I just wish we had a cpu temp monitor at all--seems absurd that we do not.
 
antutu is actually my personal favorite benchmark app. unless you just want high scores,in wich case just use quadrant ;)

while sd read/write may not seem to have much prectical performance value,alot of folks keep music and video on the sd card,so IMO its definately a releveant thing to test.

i personally dont see much point in any of it. i like ibricks dyno analagy,i can see how benchmarks can have some relevance to determine changes in performance on the same device.

i however go by the feel of my phone. a big number in a benchmark app doesnt neccessarily mean that your phone has any percievable difference in speed or smoothness :)
 
For streaming, sd card speed is functionally binary. It is either fast enough to stream or not. A benchmarking app is not going to tell you that.
 
so watching a video thats stored on an sd card is streaming? doesnt matter how fast the phone reads the card? sorry if my impressions were incorrect,i thot this was what the app was testing? :eek:
 
so watching a video thats stored on an sd card is streaming? doesnt matter how fast the phone reads the card? sorry if my impressions were incorrect,i thot this was what the app was testing? :eek:

Yes, it is streaming. All compressed video has a video bit rate--look at the file details of an avi in windows. As long as the reading of the card can stay ahead of the video playback, any additional speed is irrelevant other than for "cushion". That's why people buy cheapo 5400 rpm drives as "data" (read: downloaded movies) drives: they are fast enough to stream 1080p video so the extra speed makes no difference.

The app does test the read/write speed. Something else to be aware of: there are lots of ways to measure speed. For video and music, you want large total throughput--the higher class number SD cards sort of guarantee this type of speed. If you read through the stuff on running CM7 from an SD card for the nook color you will see something else--people recommend class 4 Sandisk cards because they have a very small block size, which makes them faster for short burst read/writes--which is more how an operating system will use the card.

And that is the problem with benchmarking. Something optimized for one application may not be as suitable for another. And often you don't know which something the benchmarking suite is looking at.

By the way...I am experimenting with running this ROM at stock speeds. :eek: One thing I found is that I can undervolt another 25 mV. I can go to -50 with no off-screen death, and no freezes up to 1 GHz. If I overclock again I may do custom voltages.

And to get back on topic: This ROM is incredible for only being in existence for a week. I know it is sort of a CM7 kang, but me likey.

Now, I just need to figure out how to get a turbo and a liquid cooling system in this thing...maybe I just need a bigger case...
 
Back
Top Bottom