A missing benchmark.

[azimafroozeh]

In your paper, there is a particular experiment that benchmarks GPU-SIMDBP128 and compares its performance with GPU-FOR on the same microbenchmark. According to your results, GPU-FOR (with 𝐷 = 16) achieved a runtime of 1.55ms, whereas GPU-SIMDBP128 took 4.3ms. Consequently, GPU-SIMDBP128 was reported to be 2.7 times slower than GPU-FOR.

However, the source code for this specific experiment is not available on repo. Would it be possible for you to share the source code for this experiment