Jump to content

Compcache - Yea or Nay


Guest PhonePersona

Recommended Posts

Guest PhonePersona

This comes up a lot of times and I don't know if there's already a thread for it or not, but I figured I'd put it out there.

I know what Compcache does theoritically.

What I want to know is if it is actually of benefit to ALL Android users of it there's a subset of 'power users' if you will, that benefit more from it than your 'basic' user.

So please share your thoughts and experiences a la Compcache if you will :)

Thanks.

Link to comment
Share on other sites

As usualy I am a ignorant unknowing fool, can someone please explain what compcash is, and what pros, cons it is to use it.. I see that its a choise to have or not having it on the custom bakery...

Link to comment
Share on other sites

Guest teknologist

I think the answer breaks down to what usage (apps installed also) you make of your device.

Try with, and without it and decide for yourself. Pretty sure there is no final and universal answer.

With the online kitchens you can choose whether you activate it or not.

Personally I use it and don't have any of the lags *a few users* users are reporting.

And I think the same goes for Paul....

Edited by teknologist
Link to comment
Share on other sites

Guest RemoteSojourner

It was great for me when I had few apps. But as the number of apps I use grew, it started lagging heavily and at one point with MCR 3.0 my phone was behaving worse than it was when I bought the phone with the notorious laggy HTC stock rom.

With 3.1, I made sure that compcache was gone and the difference was obvious. Everything is great now.

Link to comment
Share on other sites

Guest paaland

On MCR 3.1 I removed /system/init.d/ramzswap.sh which I assume is that turns on compcache on 3.1.

With compcache on I experience lagging, especially when sliding between screens in HTC SenseUI and scrolling large lists like all apps, market etc.

Edited by paaland
Link to comment
Share on other sites

On MCR 3.1 I removed /system/init.d/ramzswap.sh which I assume is that turns on compcache on 3.1.

With compcache on I experience lagging, especially when sliding between screens in HTC SenseUI and scrolling large lists like all apps, market etc.

It seems that the correct way of disabling compcache is to "comment out the proper lines in ramzswap.sh". Does anyone know what these proper lines are?

Source: http://android.modaco.com/index.php?s=&amp...t&p=1137371

Link to comment
Share on other sites

Tired both, currently running 3.2 without compcache, and it still flies, loads of free memory. I got to admit, i don't notice much if any difference between the two since i have now tried both, it actually seems to feel a little quicker coming out of sleep mode with no compcache, but its hardley anything. I think I'm sticking with no compcahe now as that seems to work best with my usage/main apps i use.

Edited by Higgsy
Link to comment
Share on other sites

So it's enough to remove /system/init.d/ramzswap.sh?

By removing the file you also remove some other optimizations which are not related to compache. I commented out (prepend lines with #) the following lines from /system/init.d/ramzswap.sh, and that seemed to disable the compcache:

#/system/xbin/insmod /system/lib/modules/ramzswap.ko disksize_kb=131072

#/system/xbin/swapon /dev/block/ramzswap0

#echo "10" > /proc/sys/vm/swappiness

Propably the lines related to lzo are useless also if compcache is not used. tun.ko module is needed for other purposes (vpn??). I don't what the xvmalloc module does. scaling_governor is probably related to "teknologist's cpu boost" when the phone wakes up from sleep.

EDIT: Remember to reboot the phone after the changes. After the phone has restarted you can execute "free" command in the command line to see if compcache is disable (cache sizes should be zero).

Edited by Salski
Link to comment
Share on other sites

Guest PhonePersona

I've got about 40 applications and an over-clocked processor on my Hero.

For the most part my Hero is extremely fast and responsive.

I tried without Compcache and it seemed a tad bit slower to me.

Link to comment
Share on other sites

Guest marco31x
I've got about 40 applications and an over-clocked processor on my Hero.

For the most part my Hero is extremely fast and responsive.

I tried without Compcache and it seemed a tad bit slower to me.

How did you over-clock your processor? I also got an HTC Hero :)

Link to comment
Share on other sites

Guest PhonePersona
How did you over-clock your processor? I also got an HTC Hero :)

I used the 'SetCPU' application.

It allows you to optimize the CPU in a bunch of ways including establishing profiles to manage your phone's CPU when its in different states e.g. Charging, Sleep...etc.

Link to comment
Share on other sites

Guest teknologist
I used the 'SetCPU' application.

It allows you to optimize the CPU in a bunch of ways including establishing profiles to manage your phone's CPU when its in different states e.g. Charging, Sleep...etc.

Well according to setCPU developer it shouldn't work on the Hero as I don't disable perflocks in the kernel...Weird...

Link to comment
Share on other sites

Guest teknologist

well after all feedback on this topic I have disabled compcache totally in my 3.2b5 based Hero...

I think Paul's last ROM is so fast that the benefits we had with CC in the first ROMs have become more a of downside in the latest ROMs.

IMHO I think there is something wrong in the way compcache works (is developed) and I don't think the occasional sluggishness are related to compression/decompression but rather an excessive fragmentation over time or something else...

My Hero feels great with 3.2b5 without compcache and this is even after 48Hrs+ uptime...

So from now on I will advise to disable compcache...

a big thx to everyone on the topic for giving the feedback !

Edited by teknologist
Link to comment
Share on other sites

Guest Konrad Walsh
Well according to setCPU developer it shouldn't work on the Hero as I don't disable perflocks in the kernel...Weird...

it has always worked and shows a noticeable difference for me

Link to comment
Share on other sites

Guest teknologist

That's good to know then !

BTW, how do you test that it works ?

HTC's perlocks is supposed to override any CPU freqs depending on the state of the screen...

Edited by teknologist
Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.