Setvsel
version: 1.51

Last update of the program in the header:30.06.2011



For the operation of the program requires root user rights.

Description:
The program is designed and works with the phone motorola Defy. Maybe it goes for other phones supporting the program milestone overckock. The program can be set to three existing moves any frequency to the latest stable 1200. In this case you can set your voltage for each stage.

Stable operating voltage for motorola defy

300 - 20
600 - 32
800 - 42
1200 - 60

That device is not left in a reboot after applying these settings, you must first apply the settings:

300 - 30
600 - 40
800 - 50

The values ​​are not recommended to put less.

standard values

300 - 33
600 - 48
800 - 58

In the program there is an item "up_threshold", it is given in percent and indicates when a load on the CPU, the CPU frequency will be increased to vsel3.

For example, if up_threshold 50%, the frequency is maximal, only when a load above 50%, to set up_threshold less than 50% personally do not recommend because is a clear capacity overruns. The standard value of 86% (and next to it), that's enough for most users.

There is a point, "Limit to VSel1 while screen is off", allows you to apply vsel1 setting, always when the phone is asleep, that significantly extends battery life on a single charge.

Directly below this point, the action can be set, with a decrease of the charge accumulator, less than a certain value, for example "do not increase the frequency, if the battery charge is less than 20%."

As you can see, CPU voltage at 300 MHz, and thus the power consumption can be reduced more than doubled, and in fact this is the frequency at which the phone is overwhelmingly lives most of the time.
Of course, it can be assumed that the stability depends on both the temperature of the CPU and the low temperature threshold is changed for the worse, so watch.

Consider that in experiments in any case not include the box Apply @ Boot, so you can just turn the power off and reboot in normal form, well preserved and, of course, too, does not interfere.
Miletone overclock should be disabled, or it is better to remove it altogether.
If you have when you start the program shows three fields to select the frequency, then you do not have root, or use, remove and put setvsel version from the market.

After a reset, all the parameters are set, it is normal. The current line shows the current parameters set in the system.

The voltage and frequency must be sought for each individual device, picking up the golden mean ratio parameters "performance-power consumption." To test the stability in addition to the program you want to install from the Market free, utility - "Stabilty Test". At the bottom of the screen there is a line of the program on "Stability test" + digging "Start", choose Vsel, the parameters of which we want to test the stability and run test for 2-3 cycles (think this is enough) + stability of your settings will show a successful test in Quadrant, as in unstable settings - program will "take off".

Android Market: market: // details id = com.SetVsel.Inteks.org?
Android Market Web: https://market.android.com/details?id=com.S...e=search_result
Xda-developers: http://forum.xda-developers.com/showthread.php?t=898036


Download: version: 1.51
SetVsel_1.51.apk(669.56 KB)




Post has been editedmainbord - 09.07.11, 13:43

To be more precise to say that with less stress to experiment with care, the minimum threshold is different, I have for example, a 300 MHz stably with 15 Vsel, worth 17, and the remaining lower voltage. On the other hand, someone Signed Off, that of high voltage than in the header bit flaky in general experiment, but gently.

I fully agree with SteedInCoat. The voltage and frequency must be sought for each individual devaysa, choosing a middle ground options ratio "performance-power consumption." To test the stability in addition to the program you want to install from the Market, also free, utility - "Stabilty Tes". At the bottom of the screen there is a line of the program on "Stability test" + digging "Start", choose Vsel, the parameters of which we want to test the stability and run test for 2-3 cycles (think this is enough) + stability of your settings will show a successful test in Quadrant, as in unstable settings - program will "take off" and show a very low result due to "crash" during the test. I suggest to all users, similar to the branch Benchmark, unsubscribe on their parameters SetVsel and how their changes affect your devaysa 2-mind criteria: increase in performance and battery life (and therefore more or less). Unsubscribe offer in form (details below its parameters):

Motorola Defy FW 3.4.2-155-2 (Froyo)
vsel1 - 18/300
vsel2 - 38/600
vsel3 - 54/1000
up_threshold - 86%
quadrant score - 1350

+ To this, in any form, of the battery life before and after customization by SetVsel.

I offer information to add to his cap.

Post has been editedelvispreasley - 06.03.11, 10:49

I have all the stable with the following parameters:
vsel1 - 19/300 (18 rarely reboot)
vsel2 - 34/600
vsel3 - 63/1200
up_threshold - 87%
Quadrant score - 2321

Post has been editedQuarx - 06.03.11, 13:44

to Quarx - like you with battery life under these parameters? Quadrant is quite impressive! I think in the future to overclock itself to 1200, but now I want to test the current settings for the duration of battery life device.

Ps: there is a suspicion that your quadrant is also due to the good "discharge" (the removal of) programs, eat operatvnuyu memory - share what you are discharged (removed), how and how much you have free RAM? He is now struggling with the workload of RAM - at the "killing" of all applications, the maximum that I could dobitsya 270 MB, which later turned into 180 ...

elvispreasley
Put his position as curator of the fact thatQuarx worth firmware with sence from HTC, and they have firmware optimized for Quadrant, 1001 I repeat, the numbers in the quadrant is absolutely nothing said about anything.

Just sign the said firmware 3.4.2-155-2 ... In general Quarx waiting for an answer, to establish the truth)
BTW: still happy to hear advice on RAM release

Firmware should Barebones - Lawnchair v1.1 (no difference what firmware 2.2)
http://forum.xda-developers.com/showthread.php?t=977989
+ Tweak build.prop

Stagefright more optimized than OpenCore, hence the result.

Although even at 800MHz, everything works fine and quickly, but if you can make it even quicker and smarter without consequences, why not do it! :)

Tsiferki mean nothing ... That's almost the best resulthttp://rghost.ru/4649405.view- happy :)
P, s Who knows why quadrant planet so buggy?

Post has been editedQuarx - 06.03.11, 15:41

vsel1 - 300 @ 27
vsel2 - 600 @ 45
vsel3 - 1000 @ 67

I came across a percent is not amenable overlokingu: hysterics:
at lower voltages, the system fills up the tests, but it works =)

And in the description of the above just goes to reboot, if the percent load operation.

PS with settings for bild.rop quadrant issued in 1777 on the 800 MHz =)

Strange. I use the drain 3.4.2-155-2. In SetVsel he noticed Enable Jit option (by pressing the button with hardware evil squares). Included. Quadrant added points. It turns out that the default Jit off?

I use the following settings:
vsel1 - 300 @ 18
vsel2 - 600 @ 36
vsel3 - 1000 @ 53
600 @ 32, in my opinion, will not work.

At the default settings of the Enable Jit still nothing has changed, and so and so points are obtained from 1230 to 1270. I must have something mixed up. Probably, this option is for 2.1?
With overklokom 1000 @ 50 1457 points.

Report on SetVsel

1. 17/300
2. 30/600
3. 60/1200

Post has been editedSteedIncoat - 07.03.11, 18:23

Overclock the processor up to 1400/75, Quadrant does not take off, the record in 2811, of course with tweaks build.prop

When the CBO will be put?)

Seriously - the result is a very worthy of praise, and, in principle, draws on the "local" record, but I think 1400 - is one too many, because increase in the frequency of 75% from the default can be quite detrimental to the health of CPU ... I've got such suspicions)

BTW: with Tviko build.prop Quadranta result on the aforementioned parameters - 1815)

Post has been editedelvispreasley - 07.03.11, 23:13

I did not resist and broke up to 1200 percent)

Motorola Defy FW 3.4.2-155-2 (Froyo)
vsel1 - 19/300
vsel2 - 38/600
vsel3 - 63/1200
up_threshold - 86%
quadrant score - 2180
build.prop = true everywhere)

Post has been editedelvispreasley - 07.03.11, 23:34

This is purely for the sake of the experiment, I suspect that this is not the record value that can be achieved at this frequency, but for a long time did not test on the basis of exactly the same about the CPU health reasons.


Of course there were more after all optimizations, especially SetVsel, consumption decreased and a half times, with family consumption parameters I was 30% В± 5%, Now 20% + -, depending on the load of 5% per day. These include:
1.10-20 minutes of calls
2.Ves day from 8 to 23 working Bluetooth
3.1-3 hours of music
4.1-1,5 hour work with applications, in particular GPRS, synchronization small things and there any

At the same time, per night, in flight screen takes less than 2%
Similar results written and other users.


Spiteful httsshnik, turnip cards glumly
In his village devayse battery
Previously, it was necessary chuhat turnip finger
Because Defi, eats three times less

Posted 07-03-2011, 23:39:


Something is not enough, I 2500-2600

Post has been editedSteedIncoat - 07.03.11, 23:44

My report SetVsel:

1. 16/300 (Less tried not to)
2. 34/700 (When fitted directly into 32 reboot)
3. 63/1200 (At 62 in a few seconds in StabilityTest - reboot)

up_threshold - 90%
quadrant score - 1718(Specifically cleaning memory, all the main run)

Each embodiment of the voltage / frequency testil slightly more than 15 minutes in StabilityTest v1.4
Sorry if not the topic, but what build.prop and what it eats?

The main topic is already painted on 155 pages

Build.prop corrected, then the following results in the quadrant:

on the drain voltage settings / frequency 1710
63/1200 and at up_threshold - 50%: 2196

Looked critical values ​​at which my machine is not stable. Here are the results that can be used:
1) 17 @ 300 (15 mistake stub 16 passes the test.)
2) 33 @ 600 (31 error 32 extends quadrant)
3) 39 @ 800 (37-reboot 38-extends quadrant)
4) 49 @ 1000 (47-reboot, 48 runs quadrant)
5) 63 @ 1200 (61-error-b2 passes quadrant)
My record in Quadrant: at 62 @ 1200 + 50% + JIT received 1666 (1537: 1895: 4147: 302: 450). More and I could not squeeze. But with editing build.prop, @ 62 1200 + 50% + JIT received in 2228 (5012: 1500: 3610: 289: 443).
SteedIncoat How you turned @ 1200 2500-2600 ??? Discover the secret :)
I would like to learn, edit build.prop not cause deterioration in the unit?
With regards to long work charge the battery, then it's difficult to draw conclusions for the nonlinear load the machine. But in my opinion, it was clearly more economical. Happy night "flight mode" - somewhere in those same 2-3 %%.