Coursework

FZJ-2026-02359

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png HPC CUDA Compiler Performance Comparison

Ctistis, T. (Corresponding author) ; Herten, A. (Thesis advisor)FZJ*

2025

This record in other databases:  

Please use a persistent id in citations: doi:10.34734/FZJ-2026-02359

Abstract: This seminar thesis compares the performance of selected CUDA samples and an real-world application between the compilers Clang and NVCC combined with GCC run on the HPC system JURECA-DC with five optimization flags -O0 up to -O3, -Ofast and -ffast-math. An empirical experiment was run to benchmark compilation and execution time as well as binary size. The benchmarks were automated with scripts written for compilation and execution with Clang and NVCC for all five optimization stages. The analysis of the results reveals that no compiler can be clearly recommended and the performance is highly dependent of the compiler, its optimization flag and the problem

Note: PDF bitte veröffentlichen.
Note: Course work, Unviersity of Cologne, 2025

---

Contributing Institute(s):

1. Jülich Supercomputing Center (JSC)

Research Program(s):

1. 5112 - Cross-Domain Algorithms, Tools, Methods Labs (ATMLs) and Research Groups (POF4-511) (POF4-511)
2. ATML-X-DEV - ATML Accelerating Devices (ATML-X-DEV) (ATML-X-DEV)

---

Database coverage:

---