hpc/roofline/report/inputs/discussion.tex

According to the definition used the arithmetic intensity is measured by operations per byte. This might not be adequat for haswell processors (and later). Due to the fused multiply-add\footnote{although called multiply-add there are 36 different slightly instructions} extension two floating point operations can be performed with a single instruction. 

- worse results for 4 threads @ NUMA-STREAM not necessarily expected
- better results for triad possibly due to combined storage in FMA
- striding for arrays

%%% Local Variables:
%%% mode: latex
%%% TeX-master: "../report"
%%% End:
report (roofline) begonnen, kernels implementiert, komische ergebnisse - irgendwo wird noch immer was wegoptimiert glaub ich 2016-06-19 18:30:49 +00:00			`According to the definition used the arithmetic intensity is measured by operations per byte. This might not be adequat for haswell processors (and later). Due to the fused multiply-add\footnote{although called multiply-add there are 36 different slightly instructions} extension two floating point operations can be performed with a single instruction.`

hälfte vom roofline report fertig 2016-06-23 00:40:48 +00:00			`- worse results for 4 threads @ NUMA-STREAM not necessarily expected`
report kernels+8er kernel bug behoben 2016-06-23 19:22:30 +00:00			`- better results for triad possibly due to combined storage in FMA`
			`- striding for arrays`
hälfte vom roofline report fertig 2016-06-23 00:40:48 +00:00
report (roofline) begonnen, kernels implementiert, komische ergebnisse - irgendwo wird noch immer was wegoptimiert glaub ich 2016-06-19 18:30:49 +00:00			`%%% Local Variables:`
			`%%% mode: latex`
			`%%% TeX-master: "../report"`
			`%%% End:`