Double-double

Accelerated Computing HPC Compilers Legacy PGI Compilers
1

Has anyone used the PGI compilers with the latest version (qd-2.3.14) of DD and QD package from Berkeley?

If so, would someone kindly share with me their Makefile.

Thanks.

Malcolm

2

Hi Mat. In a paper on using Cuda 5.0, Brent Leback used as an example the generation of DD values. Would you mind asking him for a copy of his Makefile - particularly any parts related to invoking qd-2.3.14, and then either post it or email it to me.

Thanks.

Malcolm

3

Hi Malcolm,

I asked Brent about this. He just use a few functions from the ddfun90 package since they are written in Fortran, making it easier to convert to CUDA Fortran. He didn’t use QD since it’s written in C++. Also, he didn’t use a makefile, though all the code examples are available from links in the article.

  • Mat
4

For those interested in using DD precision the following commands are useful for generating compiler flags and linker flags:

“dd-config --fcflags”
“qd-config --fclibs”

Malcolm