And much more!

Golden Jubilee Conclave

Sparsh

Shaastra Nights

Golden Design Challenges

Shop Floor To Top Floor

Innovator's Challenge

Shaastra Mobile

Smriti

Shaastra Jewels

Newsletter

The Defense Exhibition

Workshops

GPGPU (Parallel Computing) Workshop

Are you fed up because your simulations take a lot of time. Do you want your codes to run at amazing speeds without having to understand complicated clusters and bearing their exorbitant costs? Wouldn't you like to know if there is an easy way out?

This workshop aims at acquainting people with one of the most exciting developments in computation technology in recent times. GPGPU or General Purpose processing using Graphics Processor Units has gained a lot of popularity both in academics and industry. The prime motivator behind this revolution is NVIDIA with it's new disparate API called CUDA (Compute Unified Device Architecture). The massively parallel graphics processors which had been used only for computing graphics operations till recently are increasingly being used in a whole new spectrum of fields like finance, scientific and engineering applications, especially in simulations, weather forecasts etc.

The workshop will be handled by experts from NVIDIA,Bangalore and Pune. It will kick start with a brief introduction to parallel programming. It shall proceed with the comparison of architecture and organization of traditional CPUs with GPUs. The notion of GPGPU and introduction to CUDA shall be given in detail next. The various range of NVIDIA graphics cards (like Geforce, Quadro and Tesla) shall be demonstrated and their features explained. It shall conclude with the juxtaposition of a few sample codes in CUDA with their usual sequential versions and the discussion of the speed ups obtained.

Watch this space for more updates.

Eligibility : Preferably participants with CS / ECE backgrounds with decent knowledge about computer architecture and C/C++ programming are encouraged to attend. Others with similar knowledge can also attend.
The workshop is on 5th oct and there shall be two sessions of about 2-3 hours (Morning and Afternoon)

FAQ

  • Do I need to know parallel programming for attending the workshop ?
    No. Basics of parallel programming shall be dealt with initially in the workshop which shall be sufficient to understand the latter part of it.
  • Will the workshop also explain graphics processing ?
    No, the workshop does not intend to explain graphics processing. We shall focus on applications of graphics cards in everything but graphics.
  • Do I need to know very advanced computer architecture ?
    No, you need to understand basic computer architecture and organization which is generally taught as a course at most colleges.
  • What areas in C/C++ do I need to be good at?
    Nothing much. You just need to know the basics like functions, classes, branching, looping, pointers etc.
  • What use will it be to me because I cannot afford a graphics card ?
    CUDA is freely available for download and also comes with a free Emulator which can be used to do parallel computing without the actual GPUs.

Register

To register for the workshop, drop an email to dhruvrastogi@gmail.com or rathnakr@gmail.com.

Send a mail only if you are very serious and sure about attending the workshop and have the sufficient prior knowledge as we had mentioned earlier.