Muhammad Habib Rosyad was born in Lombok, Indonesia on November 1989. He obtained Bachelor of Computer Science from Gadjah Mada University in 2012. Since then, he becomes an assistant researcher in CTech Labs, EDWAR Technology’s Center for High Performance Computing. His research interests includes information system development, parallel programming, image processing, high performance computing and software engineering in general.
HPC evaluation in CTech Labs
Abstract—Target of computation power to be achieved in the newly installed HPC system in CTech Labs is 1 TFlops. Therefor, an evaluation needed to get the real measured performance. Evaluation was done by running HPL and HPCC as benchmarking software. Both performance of physical and virtual (cloud) cluster that build in the system were tested. The results shows that 1.2851 TFlops as highest performance that can be achieved.
Linear Tomograpy – Image Processing
Abstract—AHM project provides projection images to be reconstructed. The first solution was scripted in MATLAB using FBP as image reconstruction algorithm. This script run in serial manner by reconstructing the projections into slices of image (in axial) with average processing time for each was approximately 5 minutes. Total slices of image to be reconstructed is more than 200, therefore a parallel method needed to speed up the process. By developing the parallel solution in Python, the results shows that for total of 1024 slices to be reconstructed, 170 times of speed up can be achieved.