diff --git a/FAQ.md b/FAQ.md index 8fd17a960eea5bb13f916f12bc751ac94cb472c0..5d41f3f0ac1e2a68ba69e5b26e7ede6202107c0d 100644 --- a/FAQ.md +++ b/FAQ.md @@ -6,18 +6,17 @@ There are two parameters, _KBLOCK_ and _stripes_. You can vary and play with the _Why do you use AVX-512 instructions?_ If we apply our Hilbert-curve in Intel or GNU compilers, auto-vectorization will get eliminated. Writing code with AVX instructions simulates the behaviour of having an implemented auto-vectorized approach. Nevertheless, we belive that future compilers will profit from the locality assumptions of the Hilbert curve. -_ What are the Parameters KBLOCK and STRIPES used for?_ +_What are the Parameters KBLOCK and STRIPES used for?_ -KBLOCK: check after KBLOCK dimensions, whether the $\varepsilon$ distance is already exeeded. -STRIPES: How many EGO-Stripes are used. See Section 2.2 Determining the bounds''. +- KBLOCK: check after KBLOCK dimensions, whether the $\varepsilon$ distance is already exeeded. +- STRIPES: How many EGO-Stripes are used. See Section 2.2 Determining the bounds''. -There are two different parameters KBLOCK and STRIPES, where we tried out different strategies. -In our experiments (see paper) we _always_ use the following setting: +_How to set KBLOCK and STRIPES?_ +In our experiments (see paper) we _always_ use the following setting: - KBLOCK=4 - STRIPES=14 For uniform data we suggest to use the following parameter settings: - - KBLOCK=16 - STRIPES=1 \ No newline at end of file