To huge acclaim, Apple unveiled two MacBook Pro laptops featuring the high-performance and power-efficient GeForce GT 650M GPU. Among their many virtues, it turns out that they’re also excellent CUDA development platforms.

The NVIDIA GPUs provide the processing horsepower to drive the beautiful new 15” Retina Display at 2880×1800 resolution. To support all those pixels – twice what you’d see on a traditional 1080p display – each GT 650M, based on the NVIDIA Kepler Architecture, uses two energy-efficient SMX streaming multiprocessors (each comprised of 192 cores) that work with 512MB to 1GB of GDDR5 GPU memory at up to 80GB per second access speeds.

The new MacBook Pro laptops put all that GPU horsepower in a sleek UNIX-based system, making them truly elegant software development platforms for CUDA developers – especially researchers and scientists running applications on Linux clusters.

The latest CUDA 4 production release works well on the GT 650M-powered MacBook Pros. Check out the Getting Started Guide for Mac OS X for step-by-step instructions on setting up your CUDA development environment.

Similar Stories

  • kaellar

    I’ll ask one more time:
    Will nvidia support Kepler graphics found in new MBP’s? Shall we see full-feature drivers for Mac? 

  • Basti

    Pretty cool.
    But there’s more than Macs out there:

  • Will Ramey

    @kaellar, NVIDIA supports Apple in developing OpenGL drivers for NVIDIA GPUs, and Apple decides on the features and user experience to deliver for their products.

    Are there specific features you’d like to see in Apple products enabled with NVIDIA GPUs?

  • frortega

    How likely will be to use the new retina MBP as a CUDA capable device?. I am a scientist, and I am developing some CUDA applications to solve inverse problems using montecarlo methods. I know the retina MBP graphic card supports cuda, but will I be able to go beyond developing and testing using that graphic card or I will be able to run full blown jobs there?. I know it can’t be compared with the TESLA M20 i have in the lab, but i will like to have a mobile workstation that is able to deal with CUDA. 
    Thanks a lot.