Can i overclock it in "LiquidE_1.100.05"?

[Guest ktplus]

thank you :)

[Guest Vishal Arya]

thank you :)

not yet , overclocking would only be possible when the official firmware with the source code is released by acer

1.100.05 is a leaked rom , till then you would have to be on 1.6 to over clock ... unless some one also manages to leak the source too ;)

[Guest pajn]

This one (for milestone) uses a kernel module to overclock. Maybe it's possible to do the same for the acer liquid?

http://code.google.com/p/milestone-overclock/

Does anyone knows if we have permissions to add a kernel module?

[Guest Borkata]

This one (for milestone) uses a kernel module to overclock. Maybe it's possible to do the same for the acer liquid?

http://code.google.com/p/milestone-overclock/

Does anyone knows if we have permissions to add a kernel module?

Nice catch man :) I will look at it and try to setup for our device if possible ;)

[Guest Nesli]

Every android phone should have kernel module, but we need to modify this one, if we want to use it...Maybe better way is look into the nexus/desire OC/UV kernels for changes/information, same CPU is used in our machines...

..PS Our CPU is also undervolted (-0,1V), it is a reason, why it cant reach (in mostly cases) 1ghz on A1 running 1.6..When we´ll be able to revert voltage back to normal, we should OC to 1,13-1,2 Ghz... :)

PPS But everybody is waiting for a kernel source, cause we´re missing "hardcore" devs..Maybe somebody, who has at least a bit of kernel knowledge should ask OC devs from desire/nexus for help..(I think Desire also doesnt have source code open)

Edited May 19, 2010 by Nesli

[Guest Borkata]

The application and hack Milestone use doesn't need to change kernel as long as I see, so no need for kernel source :) Have to research further when I have time...

[Guest Nesli]

Yeah, kernel rest in peace;) But there are simmilar code as in kernel module used for OC..Milestone needs this, cause of lock bootloader and no possibility to flash custom/modified kernel..

[Guest wytek]

damn....i wanna that .apk on liquid!!! anyone have tryed if qworkon liquid too?? :)

p.s.

Other phones

In theory this technique is perfectly feasible in any android phone, and will surely work in phones based on Texas Instruments OMAP3. You may say it's unneeded if you can use custom firmwares with overclocking enabled; but you must flash a specific kernel to get a particular speed. It would be much better to be able to set any maximum frequency/voltage on the fly without flashing or rebooting. For instance, you could overclock to 1.0 GHz before a browsing or gaming session and then return to a lower frequency like 600 MHz, though still overcloked, for battery savings and safety. It's up to you. And that's what this project enables.

You'd have to port the kernel module over to the new kernel, which is to say, you'd have to find out where in the kernel memory are a couple of tables. I've written a detailed howto in the Disassembly page (in progress).

mmmm......

Edited May 19, 2010 by wytek

[Guest pajn]

Nice catch man ;) I will look at it and try to setup for our device if possible :D

buhu, that was what I wanted to do :( But go on man, it is probably a higher chance to make it if someone with knowledge does it :(

EDIT: But I have to try anyway! But I got stuck pretty fast

dd bs=1 count=320 skip=2147552924 if=/dev/kmem of=/sdcard/msm_clock_init

/dev/kmem: read error: Invalid argument

0+0 records in

0+0 records out

0 bytes transferred in 0.004 secs (0 bytes/sec)

Was /dev/kmem not activated when the kernel was compiled or are the something I have to do to make it work?

EDIT2: when doing cat /dev/kmem the phone reboots :)

Edited May 19, 2010 by pajn

[Guest wytek]

i have tried the apk on my rooted liquid, after program start i give su permission but the programm don't see the root....

Edited May 19, 2010 by wytek

[Guest pajn]

i have tried the apk on my rooted liquid, after program start i give su permission but the programm don't see the root....

For the first, the ram addresses are wrong. And then probably something more...

[Guest ahsid]

ahhrr this seems really interesting, someone really has to find a way to make it work !!!

[Guest vache]

Strange, somes lines of our kallsyms :

80457960 d acpu_freq_tbl

80457964 D acpu_freq_tbl_768

80457c34 D acpu_freq_tbl_998

[Guest ahsid]

Strange, somes lines of our kallsyms :

You think you can make the module work ?

[Guest barmyman]

You think you can make the module work ?

ohh, it would be at least great...

[Guest krumbalu]

yep vache c'mon try at least :)

[Guest Borkata]

Strange, somes lines of our kallsyms :

I went through source code of 1.6 and we should look for:

acpu_freq_tbl for CPUfreq frequency table

and

acpu_freq_tbl_list for MPU opps table.

Once I have time I will dump myself and disassemble.

I just wanted to share my findings in the code of kernel. Lets hope 2.1 is not so much different.

Also if anyone can post their greps and attach dumps it will be awesome.

[Guest ahsid]

I went through source code of 1.6 and we should look for:

acpu_freq_tbl for CPUfreq frequency table

and

acpu_freq_tbl_list for MPU opps table.

Once I have time I will dump myself and disassemble.

I just wanted to share my findings in the code of kernel. Lets hope 2.1 is not so much different.

Also if anyone can post their greps and attach dumps it will be awesome.

I'll help but I'm running 2.1

[Guest vache]

This is what i managed to extract :

http://shoobee.eu/upp/download.php?file=530extract.zip

Found nothing interesting for now, but still investigating...

[Guest pajn]

@Vache how did you read from /dev/kmem? or, can you please post the full result after disassemble?

[Guest HustlinDaily]

Once you get that information, would it ossible to get JIT compiler working on Liquid?

[Guest jayziac]

I think effort should be put into porting JIT compiler instead of overclocking. overclocking uses up battery & overheats, and doesn't get as much of a speedup as having a JIT compiler working.

[Guest ahsid]

guys to add JIT we kneed kernel sources and that's not going to happen for a while...

[Guest vache]

@Vache how did you read from /dev/kmem? or, can you please post the full result after disassemble?

Use "busybox dd" instead of just "dd" like said here : http://code.google.com/p/milestone-overclo...iki/Disassembly