Nvidia Privative Drivers installation


I got a new test machine for it: eeepc 1015pn: Atom N550, Nvidia Ion2 512Mo GRAM, Optimus & CUDA (for now with 1Gb RAM & no hdd)...
Starts & runs well (3.7.8, 64bits, E16, live), using nvidia legacy 340xx... (but can't figure out to see datas about the graphic ram).
A thing at start, after the nvidia driver choice, a black screen, apparently all suspending, things coming back when touching -seems any- key (usually Ctrl)...


1 Like

I have found a bug on the tool which has been fixed, you can try:

  1. plug your LAN cable
  2. boot in "console mode", with the boot option added (press TAB) of "modprobe.blacklist=nouveau"
  3. say "no" when asks for installing the drivers
  4. when console is available, run "apui elive-tools"
  5. run "sudo nvidia-privative-drivers-install -l" and install now the driver
  6. run "startx /usr/bin/e16" or similar to start the graphical system with drivers running

for the installed system that wont work, but you can reinstall them later in the installed system with similar steps (updating the tool and installing them manually)

next build should include the drivers correctly working (tested with 390-xx which didn't worked before, not tested with "current")

1 Like

So far two annoyance issues
en_SV is not in the extended list of locales ;-/
when ejecting a USB storage the message "Unfinished writes" (or some such) remains on screen, presumably forever, long after the kernel is done releasing the device.

1 Like

Considered. :rofl:(remember?)
Seriously, will forward to the head of devs.
Thanks for report :typing:

@Thanatermesis please:
Swedish / English language/locales issue;
and dbus (?) is not responding to ejected USB drive (or what, we had this bug earlier, what was it again?):

yes same bug, not possible to solve it I pressume, since looks like to work as a simple notification message, which is not the optimal, the issue is that thunar is using this system to show the message

About the language, @Hammer459 I see that there's a special option added called "Other", have you tried it to see if works? (maybe I should put it on the first of the list)


On the eeepc 1015pn, now with 2Gb RAM: starting (3.7.8 x64 live) with the 1 to 6 points, a bit long to wait but it goes to the Nvidia splash screen & e16 launch...
Little things as usual (stopping sounds....);
One point (for long) about the system monitor in the menu: why two ones (a "gnome" + "gnomeless")?
How to fit Conky prefs (on this little screen, a part goes down the limit & I didn't figure out to resize it)?
System comes quickly about freeze and hot (full vent)...
Is there a way to monitor the graphic RAM (didn't in Conky (or need extended one), nor in System Monitor)?
I presume the 512Mo video ram should help? CUDA idem?


1 Like

They're both gnome system monitor i.e exactly the same. I hadn't seen that yet ...... thank you. :kissing_heart:

  • Removed from first "gnome-system-monitor" from utilities menu in "Elive-GTK-theme-light" :innocent:
1 Like

New fresh...
Soundjuicer does not seem to work at all as installed. Does not recognize optical drive, no settings, nada ....

So I assume nvidia worked well? :slight_smile:

not sure to understand this issue, @triantares?

conky cannot be resized, but you can edit your own conf and remove un-needed entries to make it smaller in .conkyrc

try "sudo htop", i use it pretty much

cooling problems? this system should not use much cpu (so no hot)

i have see it act strangely too (suggestions for a better tool?), try to open and close the application and also the cdrom / dvd device, and also try to run "thunar" before to run it

BTW, these are questions not related to the thread, we should put them in their specific places :slight_smile: or in the betatesting sections at least :offtopic:


It seems, perhaps a problem: when I launch Elive 3.7.8 x64 live on the eeepc 1015pn with the last 1-6 process (so for privative nidia drivers, 340xx here), I have, for the best, time to arrive in e16 & do little things before the machine stops cause heat...
I change the thermal paste on the cpu+nvidia gpu, verify if it's not an issue from the the new ram of 2Gb (going back to the original 1Gb one)....
Re-test with the 1-6 ones >> same hot stop (even not arrived to e16)...
And right now, booting same conditions but strait forward (first option in the boot lines, so supposedly with "nouveau", it runs without heat stop (current uptime about 1h)...

Could you please remember me how to verify / test nvidia nouveau parameters, by the way?

When you tell about "htop", isn't it more "nvtop"? (that said here "Impossible to initialize nvidia nvml : Driver Not Loaded" (== works only with privative's?))

.conkyrc: viewed! Replacing all letters sizes from 8 to 6 and it's ok! Thanks! Now should be cool to have nvidia's datas too in it (mainly gpu%, T°, graphic ram, eventually CUDA state), but even by extend it via synaptics, I got nothing more...


1 Like


Back with the eeepc 1015pn, 2Gb ram, 3.7.8 x64, but this time, installed on a little but performing ssd...
As I still have some heat problems, I made the install by night, under a fan... :hot_face: And it runs quite well, (about 2 hours of install process), and I in fine have THE nvidia preferences tools (here with legacy 340xxx)...
I made the 6 points process for live, and made the install as it, and it works...
Well, already somewhat broken by this morning attempts with CUDA... ;-p


I doubt that your hardware, NVIDIA® ION N11M-PT1 (512MB VRAM), will support CUDA...

@Franc your heat problem seems like to be caused by a dusty cpu, but i strongly suggest you to try the "mprime" tool which will stop if happens a wrong calculation in a hot moment, more details in: How to check and verify that your computer doesn't have hardware problems

If your computer shuts down arriving to 100C', Elive should tell you about this too in your next boot (notification message)

I have also see nvidia cards that "blocks" a computer (freeze), my solution was stopping to use nvidia and switch to the internal Intel card, I don't remember now if this was solved by upgrading the entire system with the new alphas on which the kernel / drivers / xorg are much newer

About Nvidia, they should be much more easy to install on the next release 3.7.9 which i still working on, Im having some difficulties because many misc changes

1 Like


@Thanatermesis: I can ensure you that there's no dust in this computer, thermal past changed (with a supposedly good one)... Only one not changed is a little thermal pad (I haven't some), but it's usually not necessary to renew that...
I will made hardware check you point to, asap...

@Rebel450 : It's an Ion2, with 8 CUDA... :slight_smile:


Try live mode, if it's overheating with it, too.
(may be also another distro) :hooray:
I suspect a kernel issue on the system installed... :eyepopping: :gun:

Still this cute machine would be run better
with the 32bit variant, though.... :nod:

Atom & CUDA, seriously ?


Eeepc 1015pn finaly running with 3.7.8 x64 installed, again from the "6 points process" live start: nvidia privative 340 ok, nvidia X server settings running ok, showing all including the 8 cuda's, just can't figure out yet to use CUDA, but nevertheless it works likely (see 3.7.8 alpha release )..!

Surprisingly, if the eee stays a bit hot, it (for now) didn't heat crash even after about an hour uptime... The only thing I made in between is an installation of CAELinux... Can't figure out what's happen... Or it's "just" that the max temperature where I am has a bit fall (5-10°) from last week due to windy conditions.....


Have a look at



I have best experiences on many diferent NoteBooks with it,
it also reduces heat of the processor



An other thing: (eeepc 1015pn, 3.7.8 x64, installed) when playing around CUDA stuff, I install via synaptic, the nvidia cuda toolkit... The synaptic system doesn't tells anything wrong about it so I launch it... Going right until a window saying that the nvidia 418 (current) driver has been installed in place of the necessary for this computer legacy-340xx, with no possibility to go back or force to 340...
So, having to reboot (no window manager), connect via rj45, re launch the nvidia-privative-drivers-install -l ... Once done, it re install 340 ok, but tells that there's an incompatibility with 418 so turns back to 418...
I made it two or tree times, before stopping the process preventing it to retro to 418 (by hard reboot)... And it went okay..!?
Strange somewhere for me...

Beside that, the three libnvidia-legacy-340xx- (compi, cuda1, nvcuv ) 340.107.4 installs ok, but I can't figure out to evaluate it for now...



On the Dell M4400 nvidia Quadro FX770M, privative 340xx: OK!
With adaptive mode OK...

Good job again, Mr Thanat'! :+1::ok_hand:


1 Like


For the Eeepc 1015pn, 3.7.9 (upgrade from 3.7.8) rocks with nvidia 340xx, adaptive mode ok!

Next soon!



Now live from the Dell M6700/nvidia M5000M: first start choosing "current"; seems ok... asking for opengl in e16... but if the nvidia x server settings line is here, clicking on it gives nothing...
Trying to install tools with synaptic (nvtop...) to try see more...

Once done with or ~out results, I will re-try but with 390xx...

To be continued..!


4: So, on the M6700/nvidia M5000M (synaptic a bit long to dowmload) still on "current" , I installed nvidia-cuda-toolkit besides other thing (bumblebee...) AND IT ROCKS!!!
Verified with Blender 2.79 from the iso, and 2.80: Got CUDA for rendering!

@Thanatermesis: You "make my day"! HURRAY! :+1::ok_hand::vulcan_salute:

Blender 2.80 with this nvidia + Elive & CUDA is impressively fast! Having now to find or do a big work to experiment further!!!

It seems that Bumblebee isn't here after "live" install... Appears nowhere in menus or even terminology, but synaptic...
Nvidia x server settings still not here...

To be continued 2....! :slight_smile:

1 Like