The included source code implements a parallel Radix Exchange Sort algorithm, as described in the included problem description text file. Build environments for both Windows and Linux are included. Intel Threading Building Blocks (TBB) and OpenMP are both used to implement the parallelism. Utilization of Intel Parallel Studio to correct and refine the performance is documented.
Target Audience: Undergraduate with 3D math and physics background (senior)
Description: Medium-sized project that simulates the n-body problem (multiple non-colliding celestial objects influencing each other through gravity). Uses OpenGL and FreeGlut for visualization. Includes three versions for both Windows and Linux: serial, OpenMP, and CILK.
Target Audience: Graduate with 3D math and Linear Algebra background
Large project with feature-rich ray tracer (including anti-aliasing, glossy reflections, translucency, depth of field, soft shadows, and median-split volume hierarchy acceleration structure). Does not include visualization, but can easily be converted to support OpenCV. Reads Pixar Renderman (*.RIB) files and outputs *.PPM image files. Includes three versions for both Windows and Linux: serial, OpenMP, and CILK.
This workshop is inspired by Dan Grossman’s SIGCSE 2011 workshop on Data Abstractions. We review C/C++ conversions of the original Java-based materials and will include material from the Parallel Algorithms course at Kent State. The workshop will appeal to data-structure and algorithms course instructors. Workshop topics will include divide and conquer approaches, work sharing concepts, and a scoped locking scheme in OpenMP for C++ classes.