On Mon, 13 Jan 2014 09:52:36 -0800 (PST) [email protected] wrote: > понедельник, 13 января 2014 г., 16:25:57 UTC+4 пользователь Siarhei Siamashka > написал: > > > You can try to do the following test, which stresses the CPU even more: > > $ wget https://raw2.github.com/ssvb/cpuburn-arm/master/cpuburn-a8.S > > $ arm-linux-gnueabihf-gcc -o cpuburn-a8 cpuburn-a8.S > > And then run the resulting binary 'cpuburn-a8' on your device. > > If you could also monitor the temperature and/or the current draw, > > that would be really awesome. > > I've started cpuburn-a8 in first console and online radio in second > console. > 1.20Ghz \ 1.55V - reboot after 5min ,current - 11.1 watt . > 1.15Ghz \ 1.55V - cpuburn-a8 segfaults after 35min ,online > radio continues playing, current - 10.2 watt , device was > extremely hot ! > PSU - 5V\6A .
Ouch. This looks like a really bad abuse of the hardware, based on the description of your test. > Do you have results from A10 tablets or sticks ? > I think they will fail even at "nominal maximum". http://en.wikipedia.org/wiki/CPU_power_dissipation says that the CPU power consumption is proportional to P = C * V^2 * f It means that by the book, your overclocking should result in (1.55 / 1.4) ^ 2 + (1200 / 1008) = ~1.46x higher power consumption compared to 1008MHz at 1.4V. The actual numbers can be always measured and compared with the theory. Somebody may even do this and plot some nice charts. -- Best regards, Siarhei Siamashka -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
