Commit 31d2ab1f by Martin Perdacher

Merge branch 'master' of gitlab.cs.univie.ac.at:martinp16cs/hilbertJoin

parents a53c88e8 be2acaf7
 ### AVX-512 instructions _What is the correct setting of parameters for uniform distributed data?_ There are two parameters, _KBLOCK_ and _stripes_. You can vary and play with them, but throughout our experiments we have always used _KBLOCK=4_ and _stripes=14_ which are the default settings for any kind of distribution. For uniform distributed data _KBLOCK=16_ is faster, but the variation of _stripes_ has no effect. There are two parameters, _KBLOCK_ and _stripes_. Throughout our experiments we have always used _KBLOCK=4_ and _stripes=14_ which are the default settings for any kind of distribution. For uniform distributed data _KBLOCK=16_ is faster, but the variation of _stripes_ has no effect. _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. ... ... @@ -13,11 +13,13 @@ _What are the Parameters KBLOCK and STRIPES?_ _How to set KBLOCK and STRIPES?_ KBLOCK should be smaller then the dimension of the dataset. Best fitting values for stripes are: $1,2,5,14,41 (=((3^j)+1)/2)$. KBLOCK should be smaller then the dimension of the dataset. Within our distance calculation, we check after KBLOCK dimensions whether we have exceeded $\varepsilon^2$ or not. Best fitting values for active dimesnions are $0,1,2,3,4,5$ which corresponds to $1,2,5,14,41 (=((3^j)+1)/2)$ stripes, (for more details see paper Section 3.1 "Determination of the Bounds") In our experiments (see paper) we _always_ use the following setting: - KBLOCK=4 - STRIPES=14 - active dimensions=3, which are exactly 14 stripes For uniform data we suggest to use the following parameter settings: - KBLOCK=16 ... ...
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment