PGI 18.10 Compiler Benchmarks Against GCC 8.2, LLVM Clang 7.0
In the very basic C benchmarks of SciMark 2.0, the PGI compiler was yielding slower binaries than the GCC and Clang stable releases tested.
The PGI 18.10 compiler was also the slowest for the TSCP chess benchmark.
But for the C-Ray multi-threaded ray-tracing, PGI 18.10 was the fastest on this Core i9 7980XE Linux system.
PGI 18.10 picked up another win with the basic AOBench ambient occlusion renderer.
These initial tests of PGI 18.10 Community Edition were quite mixed. The PGI compiler won a number of the CPU benchmarks but also lost its fair share of them as well, really meaning you need to profile the compiler on the workloads relevant to you for determining if this NVIDIA code compiler is worthwhile for your CPU or GPU computing needs. Those wishing to try out this latest PGI compiler can do so at PGroup.com.
If you enjoyed this article consider joining Phoronix Premium to view this site ad-free, multi-page articles on a single page, and other benefits. PayPal or Stripe tips are also graciously accepted. Thanks for your support.