laptop with nvidia and intel graphics?

Accelerated Computing CUDA CUDA Programming and Performance
#1

Hello,

as I can see ,most laptops which have nvidia (cuda capable) have intel graphics also.

My question is , in order to run gpu code.

The setup (using linux) is not as simple as if nvidia was alone?Do I have to install bumblebee or something?Am I going to have problems?

Thanks!

#2

Can someone tell me about this please?

Also, someone told me that processors which have “U” means that they use the graphics card at low capability.Is that right?For example if a laptop has Intel Core i3 3217U and a nvida GeForce GT 740M , they say that it uses a more ‘slow’ edition of nvida card.

But I can’t understand.The nvida card is the 740M.It is capable for certain things (speed etc…).
Does the “U” plays role?

Thanks!

#3

To your first post, yes, you’ll have to use bumblebee or the new native support on newer kernels, although I believe those lack power management (for now, anyway).

Actually bumblebee is not a requirement to use CUDA at all. You can let the Intel GPU drive your display all the time and just install the nvidia drivers without creating an xorg configuration file and just modprobe the driver whenever you want to use cuda… see: https://wiki.ubuntu.com/Bumblebee

In regards to your other question, see:

This is from a laptop with an i7-4700MQ processor, and the bandwidth of the card is PCI-E 3.0 x8: https://devtalk.nvidia.com/default/topic/546357/cuda-programming-and-performance/sounds-like-gk208-laptops-cards-will-support-most-sm_35-features/post/3875737/#3875737

This is from an ASUS laptop with an i5-4500U processor:

(The video card’s bandwidth in that case is capped at PCI-E 2.0 x4)

My own Toshiba P50-ABT2G22 laptop with an i5-4200U processor also is capped at PCI-E 2.0 x4 speeds. So it is possible that (most?) U processors with NVIDIA chipsets are connected to a lower bandwidth PCI-E lane. Other than that, the clocks (and pipelines) on identical GPUs should remain the same across a same model card (e.g. GK208 GT740s should all have a clock around ~980 MHz and 384 cores)

#4

Thanks a lot for your help!

So , as I understand

  1. I can let intel chip to be responsible for the display and when I want to code on cuda ,I just code!(having installed the necessary packages first).

  2. So , when I see the “U” in a processor maybe it means lower bandwidth speeds.If the processor has “M” or “Q” ,this doesn’t happen.

Thanks!

#5

(1) In a high level, yes. Of course you can install bumblebee anyway and it will work with CUDA fine and have the advantage of being able to running any GPU code that is graphics-aware (e.g. OpenGL, etc) You just have to be aware to launch anything that uses the GPU with the optirun syntax for it to behave correctly.

Or for that matter if you configure the NVIDIA native xrandr support correctly:
http://us.download.nvidia.com/XFree86/Linux-x86/331.20/README/randr14.html

then graphics-aware GPU programs will work as well (with the caveats mentioned in the link above). If you simply install the drivers but do not use them as part of your X server config (what I did years ago when I wasn’t aware of bumblebee) CUDA apps will run, but any that attempt to draw to the screen will not work. Just depends what approach you take.

(2) Can depend on how the manufacturer decides to design the motherboard to use different PCI-E lanes, but from my experience it seems that the low voltage U processors tend to be using lower-bandwitdth PCI-E lanes.

#6

Thank you very much!Nice information.

#7

Hello,

I have an optimus laptop and I just installed 13.10. In order to make cuda to work I installed bumblebee and CUDA by adding extra ppa and it went smooth. Everything seems ok so far.