NexusHD2 Android Forum

tytung wrote:
Please read what cute_prince said in his [GUIDE]Overclocking NAND Devices with PLL2.




FYR, HD2 uses Qualcomm QSD8250 chipset.

Thanks but sorry, I didnt understand anything:) I want to be but Im not an expert:( I am only a last user of these fantastic works/ideas. Maybe a bit lazy, I dont know:)

My last question, is it applicable for leo?

tytung wrote:We have a OCed kernel too. But I'm not sure if PLL2 overclock or PLL4 overclock was used.

The question i my mind is one of feature vs. benefit.

I don't know the answer but while overclocking as described in the article may yield some performance benefits it may also shorten the life of the chipset and device (possible overheating is one thing that comes to mind). If my HD2 was not a working device I may be a lot more willing to experiment than I am currently. As I said, feature vs. benefit.

symbuzzer wrote:
Thanks but sorry, I didnt understand anything:) I want to be but Im not an expert:( I am only a last user of these fantastic works/ideas. Maybe a bit lazy, I dont know:)

My last question, is it applicable for leo?

Well, in my opinion, our solutions (NativeSD, DataOnEXT, and a OC'ed kernel) should be already better than their methods.

BTW, any discussion is welcome.

tytung wrote:
Well, in my opinion, our solutions (NativeSD, DataOnEXT, and a OC'ed kernel) should be already better than their methods.

I am very happy now, thanks:)

tytung wrote:
Well, in my opinion, our solutions (NativeSD, DataOnEXT, and a OC'ed kernel) should be already better than their methods.

BTW, any discussion is welcome.

I agree with tytung.
the described solution has the primary aim in having overclock in a different chipset.
they had to change the location of flash in order to avoid the wearing of nand.
and they got the idea from our DataOnEXT rom.... if you look at their initrd.gz they use the same code that we use:

Code: Select all

   
# Mount SD-DATA
    wait /dev/block/mmcblk0p2
    wait /dev/block/mmcblk0p2
    wait /dev/block/mmcblk0p2
    wait /dev/block/mmcblk0p2
    mount ext4 /dev/block/mmcblk0p2 /data nosuid nodev noatime nodiratime barrier=0 data=ordered commit=15 noauto_da_alloc nouser_xattr errors=continue

the use of multiple wait statement was my idea proposed the first time in securecrt's NativeSD thread....
so.... nothing new for us


--

tytung wrote:
Well, in my opinion, our solutions (NativeSD, DataOnEXT, and a OC'ed kernel) should be already better than their methods.

It is the best of all worlds. Many newbies (not all) simply do not spend the time to read what has already been done, and come up with "new" solutions that are almost always inferior to those which have come before it.

DataOnExt is not limited to EU versions (Tytung -- I do not think it should be indicated as such, re Aroma Installer text, as it will discourage some non-EU users from making good use of it).

I have a TMO1024 and choose the DataOnExt installation. This accomplishes two things for me by installing data/dalvik to Ext4 -- number of apps that can be installed is limited by the space I alot on my SDcard, and the dalvik-cache will have sufficient room to grow (which it does not have, if it resides on NAND). If the Dalvik space is filled and cannot grow along with the number of installed apps, you will be prevented from installing any additional apps (out of space message), regardless of how much free space still exists on your SD card.

MarkAtHome wrote:I have a TMO1024 and choose the DataOnExt installation.

Umm.. As you know, same here.




----
You can chose to ignore reality, but you cannot ignore the consequences of ignoring reality.

MarkAtHome wrote:. If the Dalvik space is filled and cannot grow along with the number of installed apps, you will be prevented from installing any additional apps (out of space message), regardless of how much free space still exists on your SD card.

Do you mean NAND? good point though.

On a slightly related point. I noticed with nativeSD that if the number of apps installed gets over 90% of available space, no further apps can be installed. With an ext4 partition of 10Gb and a number of NSD roms installed, this means that a user can't use a whole Gb of space.
Is there a way to increase the default 90% restriction in android?

Robbie P wrote:Do you mean NAND? good point though.

Hi Robbie -- yes, that is what I meant, and thought I said that (prior to your quote, in the same paragraph):

I have a TMO1024 and choose the DataOnExt installation. This accomplishes two things for me by installing data/dalvik to Ext4 -- number of apps that can be installed is limited by the space I alot on my SDcard, and the dalvik-cache will have sufficient room to grow (which it does not have, if it resides on NAND). If the Dalvik space is filled and cannot grow along with the number of installed apps, you will be prevented from installing any additional apps (out of space message), regardless of how much free space still exists on your SD card.

Your reply above was to my last sentence of the above paragraph. I think I answered your concern in that same paragraph, no?

Robbie P wrote:On a slightly related point. I noticed with nativeSD that if the number of apps installed gets over 90% of available space, no further apps can be installed. With an ext4 partition of 10Gb and a number of NSD roms installed, this means that a user can't use a whole Gb of space.
Is there a way to increase the default 90% restriction in android?

I have not used NativeSD since the very early days of Android on HD2, but I believe that Tytung modified DataOnExt so that apps can be shared, not requiring duplicate installs for each ROM. So, install the apps you want via an initial DataOnExt install, and those installed via NativeSD will be able to use them, too. Tytung?

leo chipset not same
this idea is not logical

symbuzzer wrote:Our EU HD2's nand memory is very small for new roms you know. For this reason, we are using different installation methods for best performance and flexibilty combination already.
XDA Senior Member Rootdefyxt320 found a new installation method for his HTC Explorer, and shared details with us on XDA.

News Link: http://www.xda-developers.com/android/h ... with-pll2/
Thread Link: http://forum.xda-developers.com/showthr ... ?t=2337886

Maybe we will see this new method on our Leos in next days, who know?