Thanks for the compliments.. Yeah, I don't like to use any android apps unless I have to - most always run in the background.. I just got emc working (well, booting - and not bootlooping) - I think I have to screw around with the timings a little before it actually boots into android.. I define booting as the audible click of the camera at start up.. that tells you the system at least got somewhere..
After something like 20 days of uptime I decided it was too stable.. So I 'advanced' the timings again and no crash.. I also decided that sdmmc is awful slow.. I have 2x RAID0 vertex 3s that just benchmarked at 1025 MB/sec read speeds.. yet this tablet ****s itself when trying to copy a file at 10MB/sec.. So yeah.. trying to boost the speeds of that.. Probably won't go too well..
sdmmc tweak just booted.. My objective is to have it report a max clock of 52000000(/1000/1000).. and its a no go.. here I was thinking it was pll_u.. at least thats at 52000000..
sdmmc is pll_p divided 4.5 - soo uhh.. why.. must be something awful like emc has been.. all I'm asking for is 4 mhz dammit..
I'll post the 09-20-2011 (small increase in perf) soon.
2.0GHz.. not on any of these -
(embedded image de-embedded due to size)
http://home.comcast.net/~ibladesi/tableoftablets.jpeg
The regulator won't allow anything past 1.5 volts for vdd and vdd cpu must be 50mv less than vdd core... I can try 1.8 - if there is anyone that can't do that at max vcore then i can start bumping up the freq past that..
There may be some type of 'fix' to alter/remove this limitation.. but that would require some homework to say the least.. It probably won't work.. theres a usually a ceiling or a point of diminishing returns.. in other words. bumping up the vcore past 1.5V might not even return any gain in possible freq.. then theres heat..
I'd like to see what happens when I disable the second core (not in kernel, via sysfs).. and bump up the freq to a normally crashable level. It'll probably work..
The strange thing is that my TF101 needs more vcore than the G-Tab (~1.375V @ 1704) and the Xoom ~1400mv at 1.7GHz.
That is all
Hmm, is anyone else having issues where their transformer will just freeze up with this kernel? It's happend to me enough, that I'm back on the stock kernel for now.... I'm thinking my TF just cant handle the default of 1540Mhz - it that a typical result of being too-overclocked?
When the TF freezed, I can still turn the screen off and on with the power button, but that's it.. I have to hold in the power button to reboot to get it back at that point....
Thanks!
Sent from my Transformer TF101 using Tapatalk
http://home.comcast.net/~ibladesi/TF101-09-20-2011.zip
Hold up, I think I should be able to address both of your issues, all of them.
If not, it will tell me that I've gotten a bit edgy with the ram timings. But, this is not the case for me, as I've played plants vs zombies HD (ram hog, closes on anything with less than 1GB of ram) - I played this silly game during a 3.5 hour plane trip because I had empty seats next to me (so no one could see how much of a tool I looked like).
Anyways, I'll bump up the voltage for 1544mhz and set things up so it hits all freqs (as much as it can). This is called go_maxspeed_load. This will help with battery slightly, but until stability is good, battery takes a backseat (in terms of vcore)..
I'll try to do something about that today and I'll put in a script I use to change frequencies using terminal emulator - its like 2K at most and is really easy to use. For example, stock freqs would be set by typing:
chngfrq min
chngfrq high ---- Will set to 1.424ghz
chngfrq min full ---Will set the min and max to be 1.0ghz
chngfrq higher =1.544
chngfrq higherer = 1.592
Also displays battery, dock battery (if dock is connected), temperature, current core, current set freq (min/max), current freq..
There will be a new version. This 09-20 version has quicker memory timings.. if its 'crashier' for you, then that could be evidence that its not the cpu freq but the memory timings causing instability - I doubt it tho. Probably need more voltage at 1544.