http://www.compal.com/asp/driver_dnd/index.htm
What do you think about it? Is it worth to flash my EL80 currently with 116B?
thanks
-
What is your computer not doing that you think this BIOS would fix? Anything in the release notes seem useful to you? If not, then I'd say no. Don't flash unless something isn't working.
-
All the bios images I've downloaded didn't have anything in the release notes about what it fixed/changed... I was pretty disappointed with that.
-
they repackaged the encrypted 119B released a few days ago.
but still, they DID NOT include release notes in this one as it was customary.
I don't know whether they forgot about it or was done on purpose (which purpose??)
anyway, I'm not reflashing without knowing what it's got inside -
I flashed up to 119B and I don't "see" any changed functionality...
Something I'm wondering if anyone can confirm is that somewhere along the lines (I went from 114B to 118B) the nvidia driver for linux stopped working from both my installed copy of Kubuntu 7.04, and Kubunt 7.04 / SimplyMEPIS 6.5 live CD's. -
nvidia driver wors fine for me on kubuntu and archlinux, 118b here.
-
How strange.. I just noticed since it's been about two months since I've been in my kubuntu, but it wont load X now using the 'nvidia' driver. Same for SimplyMEPIS..
-
Nvidia drivers working fine here too, 119B.
-
I flashed with 119B in the hope that the occasional ACPIEC Event 1 errors would disappear. The jury is still out on that one, since the error was not occurring all the time.
The one thing I can say about 119B is that it stops programmable button no. 2 (above the keyboard) from working. -
Flashed back down to 116B and both my custom buttons work, and I can also use the NVIDIA linux driver to load X windows again.. YAY!
My Vote is that 116 is the best they have made so far. -
yes, as for linux usage, is the 116B the best? also, does it support C2D? that's pretty much my only concern.... i have a T5500 and i want it to work fine.
cheers,
//m. -
C2D support was in my original Bios 108B -
My custom buttons work with 114B and 116B, but 119B breaks button no. 2.
I know I've seen some posts reporting Event Log entries from ACPIEC (event id 1), where the description says "...possibly a poorly designed BIOS". I've had 10 reboots since flashing 119B, and I haven't had the error once. So, although I may be jinxing things by saying this, it's possible that 119B at least takes care of that problem. -
Okay! I finally got my act together today and decided to flash to 119B. Oh the glory! Finally, my EL80 generates all the ACPI events properly - power button, lid opened/closed, AC On/Off-line.
Plus, the coretemp temperature sensors are closer to truth (finally) and! ta-da, I get PCM control of Alsa (tho that could be something else as I have also moved my ~/.asoundrc out of the way before rebooting into the new bios, something to investigate) PLUS the machine feels faster as well! (YMMV).
Also I get an acpi battery meter interface that I have not had before
All my buttons work dunno about you guys.
atkbd.c: Use 'setkeycodes e074 <keycode>' to make it known.
atkbd.c: Use 'setkeycodes e039 <keycode>' to make it known.
This is dmesg output for button 1 and 2, respectively. Ofcourse, nvidia doesn't cease to work. That's ubuntu's problem imho, it's just a broken distro.
ArchLinux ftw.
Anyway, to sum up, if anyone uses Linux out there, and has suffered due to the issues I mentioned got fixed on my EL80, give 119B a go! YMMV as written above, but you never know. Mine's a PM2. The full specs is: CHEL80945PM2.84.
Cheers,
//m. -
really? they didn't publish changelog for 119B so I didn't want to mess with it.
What do you mean by "coretemp temperature sensors are closer to truth"?
I never got ANY temp measurement with all the other bioses and with like 5 different distros and kernels. Do you mean that /proc/acpi/thermal_zone now is not empty anymore?
acpi battery, ac on/off, etc all used to work...
anyway, I guess I'll give it a try just for fun -
ok ok.. I'm confused. I downloaded 119B from the Compal site.
Apart from the mentioned lack of "changelog" (so what changed/improved is a guess).
There are two files now: EL80119B.WPH and EL80119B.ROM.
The older BIOSes had only a ROM file. What is the WPH?
And then, the HEL.BAT file that you are supposed to use says:
phlash16 EL80119B.wph /x /bbl /mode=3
completely ignoring the existence of the .ROM file
I thought that the WPH would be there to use it with the WinPhlash utility... but alas, this ZIP does not include WinPhlash, only phlash16!!
It's really really confusing. What did you guys use? The .bat? The .wph or the .rom directly? From within Windows (never!) or from within DOS?
OK, I used the WPH since it's the same size as the older ROM's
It seems to be working although I see no improvement!!
I'll investigate further -
I retried 119b and I still can't boot into X when using it... so I went back to 116b, where everything for my system works..
I'm tempted to load 119B and try a liveCD for my linux distro though, just to see if maybe something would be detected differently.. -
well. I don;t know about /proc/acpi/thermal_zone, it seems empty over here, but who needs that rubbish anyway? modprobe coretemp, sensors-detect, /etc/rc.d/sensors start, sensors. resulting in:
coretemp-isa-0000
Adapter: ISA adapter
temp1: +45°C (high = +100°C)
coretemp-isa-0001
Adapter: ISA adapter
temp1: +47°C (high = +100°C)
reasonably okey I guess?unless there is some other crucial programs dependant upon /proc/acpi/thermal_zone ;]
cheers
ps/ i used WPH, like the bat file said -
later last night i found the coretemp module in my Sabayon that I'm testing (distro whore here)
Sad that most distributions don't include coretemp by default.
Anyway, the "raw" reading from the hwmon in /sys/module/coretemp/blablabla seemed low for me... both cores idled between 28000 and 35000
If you consider those values absolute (celsius), it looks low compared to the 40somethings that I get in Windows (and that you mention in your post)
So I'm not sure if those raw values should be the offset from some fixed unknown value... on the other hand, siince I wasn't using X, just regular console mode, and almost no services running, and it's winter here, it could be the case that in fact my CPU was at an average of 30 C... which is pretty cool
(I think in the HEL80, the CPU and GPU share the heat sink, so having no graphics working would mean a cooler-than-usual GPU..) -
-
As for the heatsink, no, each has its own heatpipe, the CPU has one and the GPU has another. Deliberately, I have actually HAD something along 30 degrees on the HEL.. but I don't remember exactly how and when. I think it could've been just after removing dust off of the heatsink and using still Celeron M, on powersave (~145Mhz).
I'll actually be posting maybe some total guide to cooling/powerconsumption in notebooks on Linux, H/EL80 exactly. That would include PHC usage, PowerTOP suggested solutions, configuring acpid, and whatever comes to my mind.
Stay tuned! -
what would be insteresting is to install the custom DSDT table as explained in some other thread (I'll try it in a few days when I find the time), and see what acpi tells you and compare it to the values of coretemp....
You are right the pipes are separate but they get very close together (air coming out of the back-left corner) and sharing the same fan. So I guess that a hot GPU would have some inffluence over the CPU temp as well.
Since I wasn't using any kind of graphics (text mode linux with minimal services running), my GPU would be pretty cool, and maybe the general temp within the system was lower than when using windows graphicallly. That *could* explain unusually low CPU temperatures. On the other hand, the raw values could be some offset from a fixed value.. like.. 20 or whatever.. so when the coretemp says 28000 it actually means 48 C
I'm just wildly conjecturing here -
no, according to my settings, 28000 means 28C. Weird. That's REAL low
Not that I'm jealous or anything - congrats for having such a version of laptop
I suspect there could be something tossed with my heatpipes.... I've also unfortunately messed up the heatsink a bit and I probably should get a new one; not to mention a new fan. I'll try and convince some local reseller to sort me out with these, and if they're not too pricey, I'll go for them. Although I actually want to sell this EL80 and get an FL90. So I'm still thinking if it's worth throwing money into it ;]
-
try it with text mode linux and turn off everything and see what you get
mine has arctic silver by powernotebooks, if that means anything -
I've tried just today. About 40-42. Not everything's off tho I guess. But most is
I have configured my acpid to turn off hald polling of cdroms so SATA powersaving kicks in, CONFIG_USB_SUSPEND for unused usb ports/devices, proper laptop_mode is coming. Usb itself is supposed to give a ~1W lower momentary consumption, so composed with the rest, I reckon a 2-3W lowering. If i get to disable the programs that kill my CPU, mostly the timer is the one which makes the most wake-ups of the cpu, perhaphs CONFIG_HPET_EMULATE_RTC will make that better, those values should improve even more. Aiming for 2h 30-45minutes with EL80 6-cell. I'll have to look into the results
Oh and I use liquidmetal on my box, which is suposedly better than arctic silver. Indeed works very well on my graphics (about 41C in 2D, very nice) but not so great on cpus. I think somethings busted with the heatpipe/heatsink over the cpu. -
Do the new drivers let you change you resolution again? I've not seen a machine yet that can change resolution / refresh rate through the GUI tools using the "NVIDIA" driver... only the "NV"... -
-
I have an issue where I seem to be stuck at 50hz refresh on this system, so many games under wine and cedega won't run.
BTW.. I just realized I already have the new drivers. oh well.. -
huh? Mine seems at 50Hz too, but it's not really 50 it looks more 60tish. What's cedega to do with that. I had 50Hz before on another box and cedega went fine, afaik.
-
NEW BIOS for EL80
Discussion in 'Other Manufacturers' started by aldarion, May 25, 2007.