Authors: Jacob Villemagne and Curt Bridgers
[TODO: Write the rest of this]
This section contains a log of the amount of time spent working on this project. Each author has a subsection in which their time was logged.
June 15, 2022: 9am-11am (Got the sequential bitonic sort working)
June 15, 2022: 2pm-4pm (Class Time: Started cuda implementation)
June 16, 2022: 2pm-4pm (Class Time: Thinking...)
June 17, 2022: 10am-11am (Setup the cuda code with makefiles, helpers, etc. Can't get any print outs from the kernels...)
June 17, 2022: 11:30am-12:30pm (Got the indexing math working during the initial kernel calls for each stage of the algorithm. The values retrieved from arr are wrong, though. Probably some issue with the copy from host to device.)
June 17, 2022: 4pm-5pm (Fixed the weird memory issue. Added the memcpy from device to host after the sort is complete. Added swap code for the top level of each stage. Need to figure out how I want to handle the intermediary swapping steps of each stage. Another loop that calls more kernels? Recursive kernel launches? Decisions decisions...)
June 20, 2022: 2pm-4pm (Added a convenient way of toggling debug output/behavior. Need to add an inner loop that launches kernels for each intermediary step for the stages. Made the algorithm use more than a single CUDA block allowing us to sort meaningfully large arrays. I.E. Arrays of size greater than 2^21. I think this should be completely working by tomorrow.)
June 20, 2022: 9pm-10pm (Colorized output to reduce eye strain trying to parse it when debugging. Quality of life improvement that I should have done ages ago...)
June 21, 2022: 10am-12noon (The implementation works for array sizes up to 2^11. Beyond that, it very nearly works, but I'll have to do some more digging to figure out why. At least we have proof of concept at this point...)
June 21, 2022: 2pm-4pm (The implementation works for array sizes that are quite large. TODO: Calculate the exact number of cuda blocks we need instead of hardcoding the value in and CLEAN UP THE CODE :))
June 15, 2022: 2pm-4pm (Class Time: Started OpenMPI implementation)