[Kernel] 3.2-oc,uv,explosions. razzle dazzle.. pretty much everything (12-07-2011)

---

Ok. 12-06-2011 kernel should work now.. both renditions.. seems like the addition of bfs knocked the kernel's proverbial dick in the dirt.. so no bfs yet.

I'm having a hard time keeping track of the names people are calling this kernel.. so in light of various others naming their kernels.. i shall call this one.. hmmm. i suck at naming things.. the first thing that popped into my head was...

Razzle Dazzle <--kidding. just an example of why I stick to appending dates on the filenames.. (and i will do the same for my children)
http://home.comcast.net/~ibladesi/TF...omVchanges.zip <---This boots at 1232mhz, and insults your manhood.

and for those menly men that like their tf101 fast all the time. 1592mhz boot clock..

http://home.comcast.net/~ibladesi/TF...omVchanges.zip <---

These two have very high vcore... it is meant for people who would like to adjust their voltage themselves. If you have trouble with the normal kernels, and suspect the undervolting is causing the issues... try this.. but remember to undervolt.. really...
http://home.comcast.net/~ibladesi/TF...1-flamable.zip
http://home.comcast.net/~ibladesi/TF...1-flamable.zip


previous:
http://home.comcast.net/~ibladesi/TF101-11-26-2011.zip

Newer stuff is in bold... just in case you can't read dates.
Testing out memory timings that may bite you.. If anyone can help cut through the bull**** that is MC/EMC clocks... just chime in here.. Originally I increased (slowed down) the timings and bumped up the clock by 33mhz.. and got fail. So... advancing the timings it is.. They are pretty pitiful stock.. What really pisses me off is the missing 33mhz (and thus 66mhz from emc).. we could use that... Well.. lets see if waking up the wimpy timings helps.

edit:1680mhz. The updater-script still says the max is 1656mhz for this zip.. but the max is 1680.. i promise
http://home.comcast.net/~ibladesi/TF...-2011-1680.zip


Same as above, but a slightly advance of the memory timings. Am I even using the correct word, I seem to be in car tuning mode.. Tighter memory timings, I guess? My plan is to push it until its crashy. This is not crashy: http://home.comcast.net/~ibladesi/TF101-08-24-2011.zip

edit: throttle_temp is at 56... voltage bumped up for freqs above 1224.. So this could get a little toasty
http://home.comcast.net/~ibladesi/TF...011-highmv.zip

Edit: Now on github. http://www.github.com/ibladesi

Also because some people asked..
OhYouLikeTheKernel?

I just installed.
Boot up ok! (no boot loop)
I set it up to 1544mhz. So far so good. I will keep testing and let you know of any problem.

Well.. it defaults to 1544.. go_maxspeed_load is set to 96.. I should've set that to 99 so it would have a better chance of hitting all the clocks. Asus had it at 80.. and that's good for stock clock, but for oc - it ends up skipping past certain freqs.

Thanks for trying it out.. ntfs compat - try that too if u get a chance.

Originally Posted by Blades

Thanks for trying it out.. ntfs compat - try that too if u get a chance.

no ntfs compatibility, at least with any of my drives.

Tested it with revolver 2.1.2

Results with kernels so far:

v7 : reboot loop, and CWM crashing after aprox. 20 seconds
v8 : running smooth @ 1.4 ghz, connecting hdmi crashes on both my tv as my monitor (dvi)
TF101-08-20-2011.zip : force closes on startup. Ending up with a popup in the middle of the screen saying 'ENGLISH' . HDMI working on tv, not on monitor.

I don't know what hdmi changes are made in the last kernels, even on stock i am not able to connect to my dvi-d monitor through hdmi (samsung 2343 , no speakers in it)
Monitor supports the resolution and supports hdcp. connector is fine (popcorn mediaplayer works fine with it, also hdmi 1.3a)
I still think\hope this is a software\driver issue. Not a hardware issue.

Governors acting weird

---

I tried the new Kernel and is working somewhat stable atm, a few FC at 1.544+ . Installed with no issues on my B60 + Prime 1.7.

As for the performance, is kind of a tough call. Even thus I'm reaching new high "stable" speeds it seems like they don't really follow the governors set. For instance, If I set 1.544 @ Performance, speed drops constantly below 1.4. Benchmarks also show a lower scores than what I used to achieve with Clensym-Blades v7a for intance on Antutu I used to reach scores of 6000+ at 1.4 Ghz now, even thus I'm running at a "higher" clock I get sub -5800.

I tried to enforce the 1.544 setting it as the min speed as well with the same result.

This all makes me think that there is some kind enforcement to throttle down using trigger, which I guess will be CPU temp, however, my CPU did not go higher than 32 C even stress testing it, but as this changes occur very fast, I could be wrong on that statement.

Any ways, good first release, hope to at least comprehend what's cooking inside.

Originally Posted by runner50783

I tried the new Kernel and is working somewhat stable atm, a few FC at 1.544+ . Installed with no issues on my B60 + Prime 1.7.

As for the performance, is kind of a tough call. Even thus I'm reaching new high "stable" speeds it seems like they don't really follow the governors set. For instance, If I set 1.544 @ Performance, speed drops constantly below 1.4. Benchmarks also show a lower scores than what I used to achieve with Clensym-Blades v7a for intance on Antutu I used to reach scores of 6000+ at 1.4 Ghz now, even thus I'm running at a "higher" clock I get sub -5800.

I tried to enforce the 1.544 setting it as the min speed as well with the same result.

This all makes me think that there is some kind enforcement to throttle down using trigger, which I guess will be CPU temp, however, my CPU did not go higher than 32 C even stress testing it, but as this changes occur very fast, I could be wrong on that statement.

Any ways, good first release, hope to at least comprehend what's cooking inside.

I'll up the 1.544 voltage. Amazing temps. It downclocks at 49degC.
Read from sysfs
cat t /sys/devices/platform/tegra-i2c.3/i2c-4/4-004c/ext_temperature

Edit: Tried that antutu bench, 7397 @1656mhz. Chart topper for this device.

edit: antutu pissing contest motivated me to try out 1680.. and that works great.. I really don't like these single threaded benchmarks.. Its luck of the draw.. whatever cpu the process is assigned to really factors in on the score.. If most of the android processes are on cpu0 - then cpu1 is obviously the desirable core to use.. because android just gleefully rapes cpu cycles whenever it pleases.

antutu: 7500 @ 1.680ghz .. lost to the damn xoom by 3pts.. I am #3 for all devices.. cuz I'm ****in awesome.. I'll be telling that to all the lovely ladys at the bar.. #3 in antutu bitch.. can you buy me a drink.

Originally Posted by jwilliamf

no ntfs compatibility, at least with any of my drives.

Was the /system/lib/modules/ntfs.ko module loaded?

/dev/block/sda4 on /Removable/USBDisk1 type ntfs (rw,relatime,uid=0,gid=0,fmask=0177,dmask=077,nls= iso8859-1,errors=continue,mft_zone_multiplier=1)

I had to load the module and make the directory.. for some reason it wasn't automatically mounted.. but it works.. Ok.. so writing to the drive is not working.. I have the module compiled with RW access enabled.. Must have something to do with the way I'm mounting it. Honestly I don't use it enough to care too much.. NTFS has been found to be slower on things like flash drives.. But it does piss me off..

/system/xbin/tcpsvd -vE 0.0.0.0 21 ftpd -t 30000 -T 30000 -w / &

ftp server works wonders 4.0-4.2 MB/sec wowie!

I hate android.. I'll bet ntfs rw works on ubuntu (on tf101)

Tuxera | Leading NTFS, exFAT Interoperability on Android, Linux ...
Industrial Grade. Industry standard Tuxera NTFS, exFAT and HFS+ deliver the most reliable and highest-performing Windows and Mac interoperability.

www.tuxera.com

this is the maker of the (stock) tntfs.ko module found in the stock ramdisk.. it probably doesn't load on modded 2.6.36.4 kernels..

http://www.tuxera.com/news/asus-sele...84%a2-tablets/

You are Amazing Blades, keep it up. Will try the new updated Kernel and post results back.