This is to request that PGI provide support for Fortran quad precision. In the past, such requests have received the response “no hardware is available”. Today, we have the Power9 family that does support true 128 bit operations. PGI compilers are supported on Power9 with real(8)/complex(8). It would appear that implementing this request should be straight forward and a natural feature-extension.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| quad precision, real(16) | 0 | 8358 | November 14, 2018 | |
| Quad Precision now in Hardware | 0 | 7250 | June 20, 2019 | |
| Floating point precision on 64bit arch | 3 | 2013 | March 2, 2010 | |
| Intel Xeon Phi | 1 | 2739 | January 14, 2017 | |
| How to compile with real 16 type data? | 1 | 4933 | October 23, 2006 | |
| Fortran Coarray support | 1 | 3074 | August 9, 2010 | |
| Fortran / CUDA Fortran on Nvidia's Jetson Xavier | 1 | 1437 | February 28, 2019 | |
| Support for cc61 (Pascal) | 2 | 2621 | July 9, 2016 | |
| Any plan to implement Fortran 2008 BLOCK construct? | 3 | 6538 | March 17, 2020 | |
| choose 16/48KB vs. 48/16KB | 1 | 2960 | December 4, 2010 |