dirk, you need to connect a suitable usb device. run the command when it says waiting, then hold "u" on the keyboard of your n900 while you plug in the USB cable to the device. only after the USB logo pops up in the upper right corner should you let go of the "u" and if the drivers have been installed properly the flasher will place the new kernel on the device, should take >10s. i would wait and see if you can disconnect the device through the manager safely before just pulling the plug out. let the device boot or do what it needs to do after.
Now slide open keyboard, press and hold U. Put USB in while holding U... and then put battery in while still holding the U button. Hold U button for about 5 seconds or so and it should be back in flashing mode... now repair it!
I was talking about the N900. And the people who wrote the datasheet did design and implement it.
Lets clarify, Nokia designed the N900 and implemented the TI chip inside it, TI designed the chip and have the data sheet, who were you originally refering too?
MIPS is a lie. Also, when you tap the touchscreen or the keyboard the device bumps the cpu frequency to the maximum temporarily. This explains why what you're doing is hardly a test.
Use the above or even watch cat /sys/devices/system/cpu/cpu0/cpufreq/stats/scaling_cur_freq .
The general consensus seems the MIPS aren't exactly accurate but they can't be far off since it seems to scale with whatever the specs of the N900 state (min and max speeds), kernal changes or not.
Either way I'll submit to your point of it not being accurate since I don't have any better arguements against it.
However, I don't seem to have scaling_cur_freq, only time_in_stat and total_trans in that path....
I ran Conky on my stock N900, watched it idle at 250MHz, plugged in the cable, and after it jumped up to 600MHz until I dismissed the connection choice menu it settled on 500MHz. After I disconnected the cable it dropped right back down to 250MHz.
This seems like it could be troublesome with OC, since I usually tend to charge my phone overnight, probably for longer than it needs.
Conky shows my N900 running at about 20% of 125 mhz when charging (800 mhz kernel).
Someone tried Enna after flashing to 900mhz ? should be interesting to know if this app is really running faster.
What should be the "fastest" 900mhz with stock DSP or 900mhz with 500DSP ?
I am running the 800mhz flash , and using enna a little just to try it out, apart from 1 sec of "greenscreen" when loading a video file its running smooth for me so far
I am running the 800mhz flash , and using enna a little just to try it out, apart from 1 sec of "greenscreen" when loading a video file its running smooth for me so far
hope that helps a bit
James
i am running the 900mhz w/o dsp and it is idle at 125 on the wall charger and 500 on the usb-to-pc cable.
I have a question regarding cpu vs dsp speed. Looking at the stock version of the omap3-opp.h header file in the kernel source there seems to be a factor between the dsp and cpu frequency thats almost the same for every opp.
OPP CPU / DSP
1 125 / 90 = 1.38888889 (even though not used)
2 250 / 180 = 1.38888889
3 500 / 360 = 1.38888889
4 550 / 400 = 1.375
5 600 / 430 = 1.39534
I wonder if this is a preferable difference between the two?
If DSP is overclocked to 470 MHz the CPU should be set to 652 MHz? Whats a good dsp rate when cpu is running in 800 or 900 MHz? Going above 500 would be pushing it a little bit to far I guess :-) Or maybe this doesn't matter at all? (Sorry if someone already has Q&A this)
