Skip to content


Subversion checkout URL

You can clone with
Download ZIP


Cannot access secondary GPU (Geforce GT650M) #216

throgh opened this Issue · 57 comments

9 participants


After setting up a fresh install of Ubuntu 12.04 LTS 64bit I used to install bumblebee. So adding the repository and install the stable release as also the current nvidia drivers (295.59, nvidia-current-package). Then after a restart I got the following error message when trying to test "optirun glxgears" in the terminal:

[ 75.833050] [ERROR]Cannot access secondary GPU - error: You need to change the ConnectedMonitor setting in /etc/bumblebee/xorg.conf.nvidia to

[ 75.833076] [ERROR]Aborting because fallback start is disabled.

After a deeper look in the logs:

75.069 Loading /usr/lib/nvidia-current/xorg/
75.069 Loading /usr/lib/xorg/modules/
75.069 Loading /usr/lib/xorg/modules/
75.082 NVIDIA(0): Creating default Display subsection in Screen section
"Default Screen Section" for depth/fbbpp 24/32
75.082 NVIDIA(0): Depth 24, (==) framebuffer bpp 32
75.082 NVIDIA(0): RGB weight 888
75.082 NVIDIA(0): Default visual is TrueColor
75.082 NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
75.082 NVIDIA(0): Option "NoLogo" "true"
75.082 NVIDIA(0): Option "UseEDID" "false"
75.082 NVIDIA(0): Option "ConnectedMonitor" "DFP"
75.082 NVIDIA(0): Enabling 2D acceleration
75.082 NVIDIA(0): ConnectedMonitor string: "DFP"
75.082 NVIDIA(0): Ignoring EDIDs
75.809 NVIDIA(0): Failed to enable display hotplug notification
75.809 NVIDIA(GPU-0): Invalid ConnectedMonitor request; request was for '', but the
75.809 NVIDIA(GPU-0): valid display devices are ''.
75.810 NVIDIA(0): NVIDIA GPU GeForce GT 650M (GK107) at PCI:1:0:0 (GPU-0)
75.810 NVIDIA(0): Memory: 2097152 kBytes
75.810 NVIDIA(0): VideoBIOS:
75.810 NVIDIA(0): Detected PCI Express Link width: 16X
75.810 NVIDIA(0): Interlaced video modes are supported on this GPU
75.813 NVIDIA(0): Connected display device(s) on GeForce GT 650M at PCI:1:0:0
75.813 NVIDIA(0): none
75.813 NVIDIA(0): No display devices found for this X screen.
75.831 UnloadModule: "nvidia"
75.831 Unloading nvidia
75.831 UnloadModule: "wfb"
75.831 Unloading wfb
75.831 UnloadModule: "fb"
75.831 Unloading fb
75.831 Screen(s) found, but none have a usable configuration.

I know this GPU is already new to the market, but my hope was that this issue was solved with #172. But not for me at the moment. So is there something to get rid of this and using my notebook with his full architecture? For me this is a new topic, so when some more log-information is needed please give me perhabs a little hint where to find it.

Thanks for all help or tips / tricks.


You need to use the latest nvidia beta driver, 304.22


Okay, I'll give this a try. Thanks for reply!


So after a fresh reinstall I've tried out this, resulting in another error message:

[ 67.729891] [ERROR]Cannot access secondary GPU - error: XORG NVIDIA(0): Failed to assign any connected display devices to X screen 0

[ 67.729917] [ERROR]Aborting because fallback start is disabled.

I've added the nvidia beta driver 302.17: sudo apt-add-repository ppa:ubuntu-x-swat/x-updates
Or is there another package to install?


I've found the package:

sudo add-apt-repository ppa:upubuntu-com/nvidia-unstable
sudo apt-get update
sudo apt-get install nvidia-current

So trying to install above the corresponding files from 302.17. Perhabs this one helps!

So far, so good:

Loading new nvidia-current-304.22 DKMS files...
Building only for 3.2.0-27-generic
Building for architecture x86_64
Building initial module for 3.2.0-27-generic


Sorry guys! Another error message after that:

[ 34.049500] [ERROR]Cannot access secondary GPU - error: Could not load GPU driver

[ 34.049526] [ERROR]Aborting because fallback start is disabled.

A little view in the logs:

Jul 29 20:25:09 medion-notebook bumblebeed[1007]: Module nouveau could not be loaded (timeout?)
Jul 29 20:25:09 medion-notebook bumblebeed[1007]: Could not load GPU driver

It seems so my system doesn't find the correct driver. Or I've got something missed? Any clues for this?
Thanks so far!


It seems this could also be a problem with the current version of the kernel I'm using. For now this is the stable release After reading through some further information on this the support for the nvidia kepler architecture is not the best within 3.2. So I'm trying to move on now, keeping this issue updated after another try with the newest kernel!

UPDATE #1: After the update I've got the error message again. Bumblebee was not able to find the driver for my gpu. Trying to reinstall bumblebee in combination with the last beta drivers now!

UPDATE #2: No success on this after all.
Again "optirun" fails on loading when using a command in terminal like "optirun glxgears".

[ 64.927927] [ERROR]Cannot access secondary GPU - error: Could not load GPU driver

[ 64.927953] [ERROR]Aborting because fallback start is disabled.


You need to adjust /etc/bumblebee/bumblebee.conf to reference proper kernel module and library paths. They are different for nvidia-current package.


Ah, thanks for this hint. So I've found the concurrent values within configuration - I guess. But the paths are existing and the module file is also within the right place:



Sorry for this seems to be a very unknowledgeable based question from my side, I'm trying to find my way through the system. The files also got the correct version number as for example Got a big question mark above my head for now! But trying to keep up the search.


Did you set KernelDriver=nvidia-current?


Yes this was done at first by me!
Perhabs it would be helpful to quote the whole conf-file:

Configuration file for Bumblebee. Values should not be put between quotes

Server options. Any change made in this section will need a server restart
to take effect.
The secondary Xorg server DISPLAY number
Should the unused Xorg server be kept running? Set this to true if waiting
for X to be ready is too long and don't need power management at all.
The name of the Bumbleblee server group name (GID name)
Card power state at exit. Set to false if the card shoud be ON when Bumblebee
server exits.
The default behavior of '-f' option on optirun. If set to "true", '-f' will
be ignored.
The Driver used by Bumblebee server. If this value is not set (or empty),
auto-detection is performed. The available drivers are nvidia and nouveau
(See also the driver-specific sections below)

Client options. Will take effect on the next optirun executed.
The method used for VirtualGL to transport frames between X servers.
Possible values are proxy, jpeg, rgb, xv and yuv.
Should the program run under optirun even if Bumblebee server or nvidia card
is not available?

Driver-specific settings are grouped under [driver-NAME]. The sections are
parsed if the Driver setting in [bumblebeed] is set to NAME (or if auto-
detection resolves to NAME).
PMMethod: method to use for saving power by disabling the nvidia card, valid
values are: auto - automatically detect which PM method to use
bbswitch - new in BB 3, recommended if available
switcheroo - vga_switcheroo method, use at your own risk
none - disable PM completely

Section with nvidia driver specific options, only parsed if Driver=nvidia
Module name to load, defaults to Driver if empty or unset
colon-separated path to the nvidia libraries
comma-separated path of the directory containing and the
default Xorg modules path

Section with nouveau driver specific options, only parsed if Driver=nouveau


Okay, after a short DEBUG with "bumblebeed" got this output:

[ 1559.034620] [INFO]Loading driver bbswitch (module bbswitch)
[ 1559.034848] [DEBUG]Process modprobe started, PID 2538.
FATAL: Module bbswitch not found.
[ 1559.036691] [DEBUG]Process with PID 2538 returned code 1
[ 1559.036963] [ERROR]Module bbswitch could not be loaded (timeout?)
[ 1559.037020] [DEBUG]bbswitch is not available, perhaps you need to insmod it?
[ 1559.037062] [INFO]Skipping switcheroo PM method because it is not explicitly selected in the configuration.
[ 1559.037117] [WARN]No switching method available. The dedicated card will always be on.
[ 1559.037153] [DEBUG]Active configuration:
[ 1559.037187] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf
[ 1559.037230] [DEBUG] X display: :8
[ 1559.037264] [DEBUG] LD_LIBRARY_PATH:
[ 1559.037303] [DEBUG] Socket path: /var/run/bumblebee.socket
[ 1559.037340] [DEBUG] pidfile: /var/run/
[ 1559.037376] [DEBUG] xorg.conf file: /etc/bumblebee/xorg.conf.nouveau
[ 1559.037414] [DEBUG] ModulePath:
[ 1559.037451] [DEBUG] GID name: bumblebee
[ 1559.037488] [DEBUG] Power method: auto
[ 1559.037525] [DEBUG] Stop X on exit: 1
[ 1559.037562] [DEBUG] Driver: nouveau
[ 1559.037600] [DEBUG] Driver module: nouveau
[ 1559.037638] [DEBUG] Card shutdown state: 1
[ 1559.037807] [DEBUG]Process /sbin/modinfo started, PID 2539.
[ 1559.037880] [DEBUG]Hiding stderr for execution of /sbin/modinfo
[ 1559.040360] [DEBUG]SIGCHILD received, but wait failed with No child processes
[ 1559.040444] [DEBUG]Configuration test passed.
[ 1559.040571] [ERROR]Daemon already running, pid 1122

So optirun and bbswitch seems to have the same problem as mentioned in issue #215 on my machine. The driver used is nouveau?


Reading here some more facts and answers brings up more facts, which I'm trying out tomorrow evening! So in issue #212 there are some answers. But the central problem keeps up that the driver is not really found for now. Hands up who is next in trying to bash nvidia for some kind of bad support of the "Optimus Platform"?

Going to be serious again and hoping to solve the problem in some way. Any clues I've missed? The fun fact here is: On my older notebook (manufactured by Lenovo) "bumblebee" works like a charm. No problem at all running an "NVidia Geforce GT410M" even this one is an exotic device, but the native support of the stable drivers is given. Makes me sad and even willingly to solve the problem with the new one!

The checklist:

-> tried to install newer beta drivers 304.22 worked but not loaded for now
-> tried to upgrade to newer kernel version 3.5 for better / native support of kepler architecture, no influence at the behaviour
-> tried to modify the configuration, for now the drivers seems to be at the correct place but not really loaded from bumblebee or found by the system


Try to set these two things in the bumblebee.conf


Doing that resulted in me being able to use the 3d software.


@throgh Indeed, has fractaluser says it, you have to made such a change because the 304.22 driver has changed it's name for nvidia instead of nvidia-current (the module and the driver, not the package), which leads to no autodetection of it and bumblebee trying nouveau instead.

Also, I think that you're currently running a different kernel than the one you originall installed the modules on, and DKMS maye have not handle it, depending on how you installed the new kernel.

Please try the following :
sudo apt-get install --reinstall bbswitch-dkms nvidia-current
And paste the install log.



So far the log, which makes me sure of an existing failure on my side installing the current kernel. So the kernel sources seems to be missing.

Ersatz für bbswitch-dkms wird entpackt ...
Trigger für man-db werden verarbeitet ...
nvidia-current (304.22-beta-precise) wird eingerichtet ...
INFO:Enable nvidia-current
DEBUG:Parsing /usr/share/nvidia-common/quirks/dell_latitude
DEBUG:Parsing /usr/share/nvidia-common/quirks/put_your_quirks_here
DEBUG:Parsing /usr/share/nvidia-common/quirks/lenovo_thinkpad
DEBUG:Processing quirk Latitude E6530
DEBUG:Failure to match Medion with Dell Inc.
DEBUG:Quirk doesn't match
DEBUG:Processing quirk ThinkPad T420s
DEBUG:Failure to match Medion with LENOVO
DEBUG:Quirk doesn't match
Loading new nvidia-current-304.22 DKMS files...
Building only for 3.5.0-030500-generic
Building for architecture x86_64
Module build for the currently running kernel was skipped since the
kernel source for this kernel does not seem to be installed.
bbswitch-dkms (0.4.2-2~preciseppa1) wird eingerichtet ...
Loading new bbswitch-0.4.2 DKMS files...
Building only for 3.5.0-030500-generic
Module build for the currently running kernel was skipped since the
kernel source for this kernel does not seem to be installed.
Trigger für bamfdaemon werden verarbeitet ...
Rebuilding /usr/share/applications/bamf.index...

I'm such a beginner on this, I think for now ... am I wrong with the missing source files? Sorry some logs are in german.


Ok, so you've installed Kernel 3.5 by hand ? Or using a PPA maybe ? But it lacks the kernel header indeed. And where is your 304.22 from ?

Please lead me to where you find it, so that I can say you what packages to install. For german, that don't worry me a lot (those are basic outputs), but you can put LANG=C before the command you're running so that it's displayed in english.


Ok, same issue than #215, could you please ppa-purge Up-Ubuntu repos and try again on stock kernel.


So back again with a fresh installation and getting it up. Sorry for requesting only within an hour!

For the information: I took the kernel information from And the driver itself was from ppa:upubuntu-com/nvidia-unstable.


Ok, that's the problem, their kernels and drivers are wrong. So, please avoid using these PPAs.

If I understood well, everything is fine now ?


Oh not so far. I just admitted that I setup the system new for better testing purpose and the hope to get it right. Bumblebee is not installed yet! The system itself is new. Sorry for missunderstanding here. Are the ppas not correct? I'm not sure about this because getting a very official way doing it "correct" is sometimes very hard regarding this. That is on one point very good regarding in freedom of working, but also not the best when reaching a stable state for the system to use.


Okay I'll keep up to update my steps for now, perhabs it also helps out others when searching for solutions. So after installation getting the current updates and the kernel revision for Ubuntu Linux 12.04 x64. After this I'm usual at correcting the display for reading name of the generalized graphics adapter (Intel HD 4000) with "sudo apt-get install mesa-utils".

My next step would be adding the repositories for the unstable nvidia-drivers and the kernel. Perhabs somebody got other sources? I'm not sure about the sources. I think I better wait at this step or should I try out again?

UPDATE #1: Worked so far and the Intel chipset driver is shown correctly. Trying now getting the ppa for the nvidia-drivers, only searching for now.

UPDATE #2: I've found the following sources and ppas:

sudo apt-add-repository ppa:xorg-edgers/ppa
sudo add-apt-repository ppa:ubuntu-x-swat/x-updates

So which is the right one? I've got the feeling that there could be some incomplete data perhabs or something is missing in explanation so the user has to do something and this is not written up. Got some question marks above my head now!


For the kernel I've found: sudo add-apt-repository ppa:xorg-edgers/ppa
As for giving this a chance because it is mentioned within #172?!

Now trying it! Perhabs this helps.

UPDATE: No definitely not helping so don't try this at home, kids. My whole desktop environment is messed up and icons are gone for now. At little sigh and starting again by zero.


Hey, I'm running Debian with the GeForce 640m LE and have had some similar issues. First of all, not sure if you're experiencing freezes periodically, but if you are, you might want to try upgrading your kernel to at least 3.3. If you want a simple solution that has headers, try the liquorix kernel, which is packaged for Debian but seems to work for Ubuntu as well: Install the headers too, so you should install all 4 packages from the liquorix repo. I've had good success with it. If you have an otherwise stable kernel, you may want to avoid the confusion of doing this.

Secondly, make sure you are using the correct beta driver, the 304.22 one, not the 302.17 one: I think you're doing this, but the packages seem to be named confusingly. Make sure that if you install something from ubuntu-unstable that you install the unstable version of its dependencies as well (not sure how this is handled with Ubuntu). Installing using the .run scripts will make the drivers a lot harder to install and might prevent them from surviving kernel changes. Either way, install your kernel before you install your graphics driver.

Third, make sure you have the LibraryPath and ModulePath variables set in the bumblebee.conf file and the xorg.conf.nvidia file. That is, make sure these are accurate:


You can always do a ldd `which glxinfo` to check where your integrated graphics card is getting its libraries, and verify that the paths account for all of the required links (but the discrete card's versions, ideally).

Also, I'm not sure the Ubuntu Bumblebee repos have been updated, so you might want to download directly from here.

Hope this helps!


Thanks for the tips! So my next step solving this is to add again the repository for getting the 3.5-kernel, but ONLY this one with its dependencies. After this: Would it be helpful getting the driver directly from NVidia and install it? You mentioned the run-scripts so I admitted that for now.

Only for side note: The problem from xorg-edgers is that there are some more installations done such as some drivers and I'm not sure Ubuntu or better the x-server likes them very well for the moment.
Bumblebee got the version number 3.0.1 at last downloaded. After a little lockup here I think this is the actual version, isn't it?


Ok, you should better have wait for me to come back, but don't matters.

Do not use UpUbuntu PPAs, I've took a look at their code, both kernel and drivers are not packaged correctly. Please remain on the 3.2 sotck kernel for now on, and to 302.17 nvidia driver, we will first see if they work, and if not, what to do.

Also, do not use xorg-edgers unless you really need it, this is a PPA with bleeding edge drivers and, compiled almost daily from git.

And Ubuntu Bumbebee repos I've been updated on Friday by myself.


I suggest getting the nvidia drivers from an ubuntu ppa (maybe unstable has them), unless you know FOR SURE that it won't work. If you install directly from nvidia, they'll give you the .run script and you have to use it. I'm telling you right now, on Debian the .run script did not work. It might work for Ubuntu. The problem I had was that I could not divert the glx/3D stuff through the Intel chip without using the packages from the Debian experimental repos...which was causing me to lose 3D completely and boot into the fallback classic gnome desktop. Maybe this problem won't exist for Ubuntu, but I would suggest running something like

sudo apt-get build-dep nvidia-current

so that you have all the necessary dependencies for installing the driver manually. This may be nothing, or it may save you.

ArchangeGabriel, have the Debian repos been updated? And also, the 3.2 stock kernel is known to have problems with IvyBridge processors, but maybe they've been fixed. Also, zer01ne did not have success with the 302.17 driver as per the referenced thread. But maybe this case will be different. I do know that the GeForce 650m and 640m LE are almost identical (the latter might be an underclocked version of the former).

#201 (comment) (Sorry, I don't know how to reference a thread.)


Ubuntu doesn't work like Debian for a lot of things, but indeed using the .run for installing nvidia-current is an huge error unless you know exactly what you do.

I know that there are issues with kernel and driver, but I would first attempt to get things working with stocks packages before trying some others solutions, but I won't tell anyone to use UpUbuntu repo, I discovered them today because of #215 and #216, and their kernel and nvidia packages are wrong. So avoid using them.

Also, 640M and 650M are very different, the first one is in fact a 530M if I remember quite well, so a Fermi generation card, while the 650M is a true Kepler card. And both are suffering big mess of nVidia, particularly under Linux.


Understood, thanks for the clarification.

Actually though, two things about the graphics cards:

1) the 640m and 640m LE are distinct, but are essentially the same card (the LE is underclocked version of regular).
2) the 640m LE has two versions, a Fermi and a Kepler version. The one in the new Sony Vaio S machines (that me and zer01ne have) are the Kepler ones. This is confirmed by a number of sources including myself. Here's some info: I don't know if these cards are similar to the 650m cards, but I do know that you can overclock the 640m/640m LE cards to reach speeds near the 650m cards without too much of an overheating problem, suggesting that they might be similar.

Finally, can you let me know if/when the Debian repos will be updated? No worries if they won't be, but it would be nice to know if they do.


Ok, thanks for the card infos, this proves that nvidia policy on namming cards is even worse than what I was thinking.

Also, sorry for Debian, I've forgot that part of your message. In fact, we're not responsible of Debian repos (neither me nor an other Bumblebee project member), so that this don't rely on us. We only provide direct support for Ubuntu and Arch (because thoses are the distros we use), so that's why they're updated fastly, but for other distro, it depends on third-party packagers on which we have no control. So, juste wait & see.


Up, up and here again with the next fresh installation (new install for now again). Knowing what guys? It's a little bit cool to see that everywhere IT is the same as always: Sometimes a little bit wired, but all the time cool with trying out new ways ... so this is no problem to wait for. In fact it is more about learning and then being ready with new information. A very very big "THANK YOU" for all the things I've learned until now and that is more when working with "Windows" all day at the company. ;-)

Let me sum this up: For now I'll do only until my regular "mesa-utils" installation. Then trying to find a good source! I assume that some of the packages are very experimental for now. Not the big problem, but not the best for a stable system. The BETA-drivers are not the problem, for more missing the kernel sources in some kind. Before I forget this: The newest issue #220 seems to have even same problems. So it is time also for NVidia helping here - only for something at least comparable to a wish.

Ah, not to forget: When somebody got more information about good ppa to use, please feel free to post this here. It could only help and at the moment it is no problem for me to try out and test.


Okay, thank you ArchangeGabriel.


Please first try to see if Bumblebee works on a fresh installation, before trying to get newer driver - you may still update them after. Also, for 304.22 from UpUbuntu aren't packaged the regular way for this driver, and the kernel from the same PPA group is missing headers (the "kernel sources"). In fact they are present, but the kernel is stock upstream one, with no Ubuntu configuration, and they aren't building meta packages to ensure proper installation and upgrade.

If a fresh install don't work, we may consider a 12.10 alpha installation, but first try with stock system - it worked for #215.


Ok, doing so. Wait a moment, doing a reboot after the first updates. Then adding bumblebee as normal way with the stock kernel 3.2 delivered.

UPDATE #1: Repo added! Doing now the installation.

UPDATE #2: Installation in progress. bbswitch and bumblebee throw no errors. But this was always the same behaviour.


And again the error message when trying out "optirun glxgears":

[ 87.159194] [ERROR]Cannot access secondary GPU - error: You need to change the ConnectedMonitor setting in /etc/bumblebee/xorg.conf.nvidia to

[ 87.159223] [ERROR]Aborting because fallback start is disabled.

Looking concrete at the Xorg.8.log:

[ 85.073]
X.Org X Server 1.11.3
Release Date: 2011-12-16
[ 85.073] X Protocol Version 11, Revision 0
[ 85.073] Build Operating System: Linux 2.6.42-26-generic x86_64 Ubuntu
[ 85.073] Current Operating System: Linux medion-notebook 3.2.0-27-generic #43-Ubuntu SMP Fri Jul 6 14:25:57 UTC 2012 x86_64
[ 85.073] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-3.2.0-27-generic root=UUID=49858e1d-7ca3-4cc7-b102-c2c1ef3e39a7 ro quiet splash vt.handoff=7
[ 85.073] Build Date: 16 July 2012 08:06:31PM
[ 85.073] xorg-server 2:1.11.4-0ubuntu10.6 (For technical support please see
[ 85.073] Current version of pixman: 0.24.4
[ 85.073] Before reporting problems, check
to make sure that you have the latest version.
[ 85.073] Markers: (--) probed, (*) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
85.073 Log file: "/var/log/Xorg.8.log", Time: Mon Jul 30 22:35:47 2012
85.074 Using config file: "/etc/bumblebee/xorg.conf.nvidia"
85.074 Using system config directory "/usr/share/X11/xorg.conf.d"
85.094 ServerLayout "Layout0"
85.094 No screen section available. Using defaults.
85.094 |-->Screen "Default Screen Section" (0)
85.094 | |-->Monitor ""
85.094 No device specified for screen "Default Screen Section".
Using the first device section listed.
85.094 | |-->Device "Device1"
85.094 No monitor specified for screen "Default Screen Section".
Using a default monitor configuration.
85.094 Option "AutoAddDevices" "false"
85.094 Not automatically adding devices
85.094 Automatically enabling devices
85.094 The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[ 85.094] Entry deleted from font path.
85.094 The directory "/usr/share/fonts/X11/100dpi/" does not exist.
[ 85.094] Entry deleted from font path.
85.094 The directory "/usr/share/fonts/X11/75dpi/" does not exist.
[ 85.094] Entry deleted from font path.
85.094 The directory "/usr/share/fonts/X11/100dpi" does not exist.
[ 85.094] Entry deleted from font path.
85.094 The directory "/usr/share/fonts/X11/75dpi" does not exist.
[ 85.094] Entry deleted from font path.
85.094 The directory "/var/lib/defoma/x-ttcidfont-conf.d/dirs/TrueType" does not exist.
[ 85.094] Entry deleted from font path.
85.094 FontPath set to:
85.094 ModulePath set to "/usr/lib/nvidia-current/xorg,/usr/lib/xorg/modules"
85.094 |-->Input Device ""
85.094 |-->Input Device ""
85.094 The core pointer device wasn't specified explicitly in the layout.
Using the default mouse configuration.
85.094 The core keyboard device wasn't specified explicitly in the layout.
Using the default keyboard configuration.
85.094 Loader magic: 0x7f0bf07e8b00
85.094 Module ABI versions:
[ 85.094] X.Org ANSI C Emulation: 0.4
[ 85.094] X.Org Video Driver: 11.0
[ 85.094] X.Org XInput driver : 16.0
[ 85.094] X.Org Server Extension : 6.0
85.095 PCI:
(0:1:0:0) 10de:0fd1:1b0a:20dd rev 161, Mem @ 0xf5000000/16777216, 0xe0000000/268435456, 0xf0000000/33554432, I/O @ 0x0000e000/128, BIOS @ 0x????????/524288
85.095 Open ACPI successful (/var/run/acpid.socket)
85.095 LoadModule: "extmod"
85.118 Loading /usr/lib/xorg/modules/extensions/
85.118 Module extmod: vendor="X.Org Foundation"
[ 85.118] compiled for 1.11.3, module version = 1.0.0
[ 85.118] Module class: X.Org Server Extension
[ 85.118] ABI class: X.Org Server Extension, version 6.0
85.118 Loading extension MIT-SCREEN-SAVER
85.118 Loading extension XFree86-VidModeExtension
85.118 Loading extension XFree86-DGA
85.118 Loading extension DPMS
85.118 Loading extension XVideo
85.118 Loading extension XVideo-MotionCompensation
85.118 Loading extension X-Resource
85.118 LoadModule: "dbe"
85.118 Loading /usr/lib/xorg/modules/extensions/
85.118 Module dbe: vendor="X.Org Foundation"
[ 85.118] compiled for 1.11.3, module version = 1.0.0
[ 85.118] Module class: X.Org Server Extension
[ 85.118] ABI class: X.Org Server Extension, version 6.0
85.118 Loading extension DOUBLE-BUFFER
85.118 LoadModule: "glx"
85.118 Loading /usr/lib/nvidia-current/xorg/
86.011 Module glx: vendor="NVIDIA Corporation"
[ 86.011] compiled for 4.0.2, module version = 1.0.0
[ 86.011] Module class: X.Org Server Extension
86.012 NVIDIA GLX Module 295.40 Thu Apr 5 21:57:38 PDT 2012
86.012 Loading extension GLX
86.012 LoadModule: "record"
86.012 Loading /usr/lib/xorg/modules/extensions/
86.012 Module record: vendor="X.Org Foundation"
[ 86.012] compiled for 1.11.3, module version = 1.13.0
[ 86.012] Module class: X.Org Server Extension
[ 86.012] ABI class: X.Org Server Extension, version 6.0
86.012 Loading extension RECORD
86.012 LoadModule: "dri"
86.012 Loading /usr/lib/xorg/modules/extensions/
86.012 Module dri: vendor="X.Org Foundation"
[ 86.012] compiled for 1.11.3, module version = 1.0.0
[ 86.012] ABI class: X.Org Server Extension, version 6.0
86.012 Loading extension XFree86-DRI
86.012 LoadModule: "dri2"
86.013 Loading /usr/lib/xorg/modules/extensions/
86.013 Module dri2: vendor="X.Org Foundation"
[ 86.013] compiled for 1.11.3, module version = 1.2.0
[ 86.013] ABI class: X.Org Server Extension, version 6.0
86.013 Loading extension DRI2
86.013 LoadModule: "nvidia"
86.013 Loading /usr/lib/nvidia-current/xorg/
86.111 Module nvidia: vendor="NVIDIA Corporation"
[ 86.111] compiled for 4.0.2, module version = 1.0.0
[ 86.111] Module class: X.Org Video Driver
86.138 LoadModule: "mouse"
86.138 Loading /usr/lib/xorg/modules/input/
86.154 Module mouse: vendor="X.Org Foundation"
[ 86.154] compiled for 1.11.3, module version = 1.7.1
[ 86.154] Module class: X.Org XInput Driver
[ 86.154] ABI class: X.Org XInput driver, version 16.0
86.154 LoadModule: "kbd"
86.155 Warning, couldn't open module kbd
86.155 UnloadModule: "kbd"
86.155 Unloading kbd
86.155 Failed to load module "kbd" (module does not exist, 0)
86.155 NVIDIA dlloader X Driver 295.40 Thu Apr 5 21:38:35 PDT 2012
86.155 NVIDIA Unified Driver for all Supported NVIDIA GPUs
86.167 using VT number 7

86.168 Loading sub module "fb"
86.168 LoadModule: "fb"
86.168 Loading /usr/lib/xorg/modules/
86.168 Module fb: vendor="X.Org Foundation"
[ 86.168] compiled for 1.11.3, module version = 1.0.0
[ 86.168] ABI class: X.Org ANSI C Emulation, version 0.4
86.168 Loading sub module "wfb"
86.168 LoadModule: "wfb"
86.169 Loading /usr/lib/xorg/modules/
86.182 Module wfb: vendor="X.Org Foundation"
[ 86.182] compiled for 1.11.3, module version = 1.0.0
[ 86.182] ABI class: X.Org ANSI C Emulation, version 0.4
86.182 Loading sub module "ramdac"
86.182 LoadModule: "ramdac"
86.182 Module "ramdac" already built-in
86.184 Loading /usr/lib/nvidia-current/xorg/
86.184 Loading /usr/lib/xorg/modules/
86.184 Loading /usr/lib/xorg/modules/
86.207 NVIDIA(0): Creating default Display subsection in Screen section
"Default Screen Section" for depth/fbbpp 24/32
86.207 NVIDIA(0): Depth 24, (==) framebuffer bpp 32
86.208 NVIDIA(0): RGB weight 888
86.208 NVIDIA(0): Default visual is TrueColor
86.208 NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
86.208 NVIDIA(0): Option "NoLogo" "true"
86.208 NVIDIA(0): Option "UseEDID" "false"
86.208 NVIDIA(0): Option "ConnectedMonitor" "DFP"
86.208 NVIDIA(0): Enabling 2D acceleration
86.208 NVIDIA(0): ConnectedMonitor string: "DFP"
86.208 NVIDIA(0): Ignoring EDIDs
87.070 NVIDIA(0): Failed to enable display hotplug notification
87.071 NVIDIA(GPU-0): Invalid ConnectedMonitor request; request was for '', but the
87.071 NVIDIA(GPU-0): valid display devices are ''.
87.124 NVIDIA(0): NVIDIA GPU GeForce GT 650M (GK107) at PCI:1:0:0 (GPU-0)
87.124 NVIDIA(0): Memory: 2097152 kBytes
87.124 NVIDIA(0): VideoBIOS:
87.124 NVIDIA(0): Detected PCI Express Link width: 16X
87.124 NVIDIA(0): Interlaced video modes are supported on this GPU
87.136 NVIDIA(0): Connected display device(s) on GeForce GT 650M at PCI:1:0:0
87.136 NVIDIA(0): none
87.136 NVIDIA(0): No display devices found for this X screen.
87.157 UnloadModule: "nvidia"
87.157 Unloading nvidia
87.157 UnloadModule: "wfb"
87.157 Unloading wfb
87.157 UnloadModule: "fb"
87.157 Unloading fb
87.157 Screen(s) found, but none have a usable configuration.
[ 87.157]
Fatal server error:
[ 87.157] no screens found
[ 87.157]
Please consult the The X.Org Foundation support
for help.
[ 87.158] Please also check the log file at "/var/log/Xorg.8.log" for additional information.
[ 87.158]
[ 87.158] ddxSigGiveUp: Closing log
[ 87.158] Server terminated with error (1). Closing log file.


According to the LOG: The graphics adapter was found. Here: 87.124 NVIDIA(0): NVIDIA GPU GeForce GT 650M (GK107) at PCI:1:0:0 (GPU-0). But even that could not be addressed because there are no screens found. I must admit: My knowledge here keeps me alone with this information, because I can't translate it what this could mean in this context.

The driver version is the last certified. Hope this is not my fault the ticket is so filled up with comments or I got you even with false information. If this is so, my apologize for doing the beginner errors. At the moment I'm feeling a little bit confused about using the beta drivers or not! ;-)


Ok, so then you effectively need the 304.22 driver and replacing "ConnecteMonitor" "DFP" by "UseDisplayDevice" "none" in /etc/bumblebee/xorg.conf.nvidia.

For getting the 304.22, the best would be to wait until the stable version derived from it land into x-swat PPA, but you may use UpUbuntu ones if you really want, however I won't be able to help a lot with specific problems that may happen then.

For getting 3.5 kernel, please take a look at instructions I gave in the end of #215.


Thank you! :-)
So I'm going to try out the config-file at first. Only to be sure of.


No, the config file is only if you're using 304.22 drivers from xorg-edgers.


Understood, sorry for this. And again trying out! Hope the x-swat PPA driver is updated within time. A pity not to use the full amount of the machine. Keep the ticket updated with my results, so it could help anybody with the same problems.


Added the repository from Changed the configuration file.


[ 35.922302] [ERROR]Cannot access secondary GPU - error: Could not load GPU driver

[ 35.922348] [ERROR]Aborting because fallback start is disabled.


What give those commands:
sudo modprobe nvidia
sudo modprobe nvidia-current


Also, could you paste the installation log or update log for nvidia-current ?


Ha, exactly the same I wanted to post in time. The next try was to change up the bumblebee.config!

Exactly this I've done:

Added "driver=nvidia" and also changed "KernelDriver=nvidia". Now optirun seems to work. But please with closing the ticket until glxgears and glxspheres have got verified logs.


This is now taken from glxspheres:

optirun glxspheres
Polygons in scene: 62464
Visual ID of window: 0x21
Context is Direct
OpenGL Renderer: GeForce GT 650M/PCIe/SSE2
146.422204 frames/sec - 163.407179 Mpixels/sec
154.082256 frames/sec - 171.955797 Mpixels/sec
154.158205 frames/sec - 172.040556 Mpixels/sec
152.885180 frames/sec - 170.619861 Mpixels/sec
150.620836 frames/sec - 168.092853 Mpixels/sec
154.174214 frames/sec - 172.058423 Mpixels/sec

So in fact what was done for users with the 650M?

  1. Installation of Ubuntu (of course, including actual updates with mesa-utils for display bugfix in system window)
  2. bumblebee installation from stable
  3. bumblebee-nvidia installation from stable (nvidia-current)
  4. Adding the repository from
  5. Changing the config in /etc/bumblebee/xorg.conf.nvidia
  6. Changing name of driver (seems to be messed up sometimes?)

So? It works! Yeah and I'm glad and also surprised again by Linux, the community and everything around. In Windows this would be end up a little bit different wether there are helpful people around or not. Do you have some kind of "Donate"? Or perhabs even helping with more messages. :-)



Nice to see it working !

The name of driver is a change introduced in 304.22 in fact.

No, there isn't any donate, because we agreed with the others members not to receive any donations. We're simply happy to help the community (and ourself, because we firwt went here as concerned people obviously).


throgh, can you run your glxspheres for about 2 mins and see where the values go? I'm having a problem where mine decrease to incredibly low values (below 10fps and 10Mpps) after around 30/45 seconds, and wonder if you have the same behavior. I'll post a bug report about this later tonight when I'm at my Linux computer.


Yes, I'll do this. Erm, keep up writing here even if closed? Just for a test.

EDIT: Okay, functional. Wait: I'll have a look on.


Started at 23:18 here in germany local time.

23:19 - stable at 180 - 190 for now.
23:20 - already stable at 180 ... ah let me correct: I meant 155 - 160 frames, 176 - 180 MPixels/sec

At last: The test runs now for four minutes and the program keeps up stable framerate.


Yes you can keep writing even if closed : this is just for putting it as "solved". And also, you can re-open if facing again this issue.


Okay thanks for the data. Maybe I need to reinstall some things or point to different libraries/modules. I'll play more with this later.


@ArchangeGabriel: Thanks! Do you have also some cookies here hidden? Just a little joke but this seems so good and a dream at the moment, because of working together and helping about a problem with the goal to solve this, learning about and get more knowledge. Just to say this is the way to go! :-) Wish this could be done everywhere in this way.

@seanlaguna: Okay, when you need something more results or some tests again, keep a short update on this.


Okay, thanks for the help throgh.

I agree as well, thanks so much for supporting bumblebee/bbswitch. These are amazing packages, and I'm very grateful for you and everyone else's work on them.


I had same problem is a Sony Vaio S with the nvidia GT 640 LE.

I followed the @throgh and @ArchangeGabriel steps and it works pretty well!

$ optirun glxspheres
Polygons in scene: 62464
Visual ID of window: 0x21
Context is Direct
OpenGL Renderer: GeForce GT 640M LE/PCIe/SSE2
118.690321 frames/sec - 132.458398 Mpixels/sec
120.290725 frames/sec - 134.244450 Mpixels/sec
119.697766 frames/sec - 133.582707 Mpixels/sec

Thanks so much!!


One little offtopic: I own a sony vaio S too, and it seems to me that the hardware switch between stamina mode and speed mode is just "fake".

I mean, it doesnt actually switch anything at an hardware level, it's software that changes the mode under windows according to the state of the switch.

Right ?


hi , everyone
i have this problem as you.
my laptop is gt650m card
if you have solved this problem,please let me kown.
any way,i am expecting.
my email is x140ke@gmail


@cyder : Try this :

sudo add-apt-repository ppa:ubuntu-x-swat/x-updates
sudo apt-get update
sudo apt-get upgrade

after reading all of these comments.. first thumbsup to the community behind bunblebee.. thanks for your support..
What i have learned is that through bumblebee, only select applications (through optirun) can be run using the nvdia graphics for hardware acceleration.. and the ubuntu system as a whole solely runs on intel's h/w acceleration.. is it?? and so its not a true switch to nvdia hw to be used by default??
and how things differ if say the bios has a switch which turns off the intel h/w so that only nvdia is used by default every where??Will the same bumblebee drivers work in such a scenario too?

And.. most importantly.. now i understood why Linus Trovalds showed middle finger to nvdia few months back!
Why cant nvdia just recruit you all guys and give an official support to the greatest OS if all time - LINUX ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.