CCL: PBE for inorganic chemistry

 Sent to CCL by: Susi Lehtola [susi.lehtola/./alumni.helsinki.fi]
 On 07/25/2018 11:15 AM, Grigoriy Zhurko reg_zhurko.,.chemcraftprog.com
 wrote:
Sent to CCL by: Grigoriy Zhurko [reg_zhurko-*-chemcraftprog.com] Hello, 
 
 Look for Prof. Marcel Swart at ICREA. The 2017 results:
 
 
 <http://marcelswart.eu/dft-poll/2017.html#start>;
 
 I have used the Marcel Swart poll, but I simply don't know how to
 cite it in my publication. Did mr Swart ever publish his results in a
 reviewed journal? As far as I understand, citing a website is a bad
 form for a serious journal.
As has been discussed before on the forum, popularity may not correlate with accuracy. You should choose the functional by comparison with ab initio calculations, or rely on a literature benchmarks and reviews, such as the one by Mardirossian and Head-Gordon that was already suggested by J C Womack.
Also, a very good rule of thumb is not to trust a result from a single dft functional: you should compare the results of at least a few different functionals, e.g. a few pure ones and a few hybrid ones. If the results are similar, then you can trust DFT for your application. However, if the results differ significantly, then it's likely that DFT has fundamental problems for your system. 
 --
 ------------------------------------------------------------------
 Mr. Susi Lehtola, PhD             Junior Fellow, Adjunct Professor
 susi.lehtola,alumni.helsinki.fi   University of Helsinki
 http://www.helsinki.fi/~jzlehtol  Finland
 ------------------------------------------------------------------
 Susi Lehtola, dosentti, FT        tutkijatohtori
 susi.lehtola,alumni.helsinki.fi   Helsingin yliopisto
 http://www.helsinki.fi/~jzlehtol
 ------------------------------------------------------------------