[ltp] [ANN] tp_smapi 0.19 adds dates, improves stability

David A. Desrosiers linux-thinkpad@linux-thinkpad.org
Sat, 8 Apr 2006 10:57:18 -0400 (EDT)


> Can you please post a detailed, reproducible description of the 
> problem?

 	# Make the machine drop to its lowest CPU speed, 600Mhz in my
 	# case:
 	cd /sys/devices/system/cpu/cpu0/
 	echo powersave > cpufreq/scaling_governor

 	# Check the current frequency
 	cat cpufreqd/cpuinfo_cur_freq
 	600000

 	# Check the maximum frequency at this point
 	cat cpufreq/cpuinfo_max_freq
 	2100000

 	# Crank the machine to its maximum speed:
 	echo performance > cpufreq/scaling_governor

 	cat cpufreqd/cpuinfo_cur_freq
 	2100000

 	modprobe tp_smapi

 	cat cpufreqd/cpuinfo_cur_freq
 	1200000

 	cat cpufreqd/cpuinfo_max_freq
 	1200000

 	The governor was not changed, and still reports "performance", 
but tp_smapi upon load, immediately halves the maximum processor 
speed, and changing any of those values by echo'ing the proper values 
to their counterparts, does not change the behavior.

> BTW, my understanding is that "exercising" the battery by cycling it 
> will wear it down badly. Working on the last few percents of 
> capacity (i.e., low voltage operation) is especially bad for Li-Ion 
> batteries.

 	In my case, I'm having exactly the opposite results. I 
actually have added about 15% of life back into 1 of the batteries by 
repeatedly draining and recharging the battery, while the laptop is 
unattended and powered off.

 	I'm sure I won't get back to the full 71280 mWh, but its 
better than it was before, and I was about to buy new batteries 
anyway, since these are now 4 days out of warrantee, and don't hold 
more than 2 hours charge.

 	I wish *ANY* battery in the T42p would hold more than 2 hours 
of charge. The battery ratings I've been reading are completely false, 
and are fictitiously 3-4x the actual real-world results I'm getting on 
maximum power savings, in either Windows or Linux.


David A. Desrosiers
desrod@gnu-designs.com
http://gnu-designs.com