From owner-chemistry@ccl.net Wed Jul 27 12:26:01 2022
From: "Andrew DeYoung andrewdaviddeyoung*o*gmail.com" <owner-chemistry^server.ccl.net>
To: CCL
Subject: CCL:G: Comparing the computational cost of BLYP and B3LYP
Message-Id: <-54776-220727122410-27687-IH+vapOhCu6OTPSzpaz0Xg^server.ccl.net>
X-Original-From: Andrew DeYoung <andrewdaviddeyoung---gmail.com>
Content-Type: multipart/alternative; boundary="00000000000044c76b05e4cbd804"
Date: Wed, 27 Jul 2022 12:23:51 -0400
MIME-Version: 1.0


Sent to CCL by: Andrew DeYoung [andrewdaviddeyoung(a)gmail.com]
--00000000000044c76b05e4cbd804
Content-Type: text/plain; charset="UTF-8"

Hi,

I am rather new to quantum chemical calculations, and I am wondering if you
can help me make sense of some benchmark timing results I obtained from
Gaussian 16.

I ran Gaussian 16 calculations of an ion pair with a total of 31 atoms (18
atom cation, 13 atom anion).  I obtained the following wall-clock timings
for various combinations of method and basis set:

HF/6-31G, 6.47 sec
HF/6-31G(d), 14.21 sec
HF/6-31G(d,p), 18.93 sec

BLYP/6-31G, 20.35 sec
BLYP/6-31G(d), 46.43 sec
BLYP/6-31G(d,p), 56.93 sec
BLYP+D3/6-31G(d,p), 57.68 sec

B3LYP/6-31G, 18.21 sec
B3LYP/6-31G(d), 32.50 sec
B3LYP/6-31G(d,p), 41.17 sec
B3LYP+D3/6-31G(d,p), 41.24 sec

I am not surprised that HF is the least computationally expensive method of
the three, but I am surprised that, according to these results, BLYP is
more expensive than B3LYP.  Should I be surprised by this?

I was under the impression that BLYP is rather popular in ab initio MD
precisely because it is less expensive than B3LYP.  (Of course, Gaussian
does not, as far as I know, do ab initio MD, so perhaps BLYP's greater
expense in Gaussian is just due to implementation; I'm guessing BLYP is not
that popular for electronic structure these days, so perhaps its
implementation in Gaussian is not as optimized as the immensely popular
B3LYP?)

B3LYP is a hybrid functional, while BLYP is not, but I realize that this
does not, by itself, mean that BLYP should be less expensive than B3LYP.

Thanks,
Andrew

Andrew DeYoung, PhD
Department of Chemistry
Carnegie Mellon University

--00000000000044c76b05e4cbd804
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Hi,</div><div><br></div><div>I am rather new to quant=
um chemical=20
calculations, and I am wondering if you can help me make sense of some=20
benchmark timing results I obtained from Gaussian 16. <br></div><div><br></=
div><div>I
 ran Gaussian 16 calculations of an ion pair with a total of 31 atoms=20
(18 atom cation, 13 atom anion).=C2=A0 I obtained the following wall-clock=
=20
timings for various combinations of method and basis set:</div><div><br></d=
iv><div>HF/6-31G, 6.47 sec</div><div>
HF/6-31G(d), 14.21 sec <br></div><div>
HF/6-31G(d,p), 18.93 sec <br></div><div><br></div><div>
BLYP/6-31G, 20.35 sec</div><div>
BLYP/6-31G(d), 46.43 sec <br></div><div>
BLYP/6-31G(d,p), 56.93 sec <br></div><div>
BLYP+D3/6-31G(d,p), 57.68 sec=20



</div><div><br></div><div>B3LYP/6-31G, 18.21 sec</div><div>
B3LYP/6-31G(d), 32.50 sec <br></div><div>
B3LYP/6-31G(d,p), 41.17 sec <br></div><div>
B3LYP+D3/6-31G(d,p), 41.24 sec <br></div><div><br></div><div>I am not=20
surprised that HF is the least computationally expensive method of the=20
three, but I am surprised that, according to these results, BLYP is more
 expensive than B3LYP.=C2=A0 Should I be surprised by this?=C2=A0=C2=A0</di=
v><div><br></div><div>I was under the impression that BLYP is rather popula=
r in ab initio MD precisely because it is less expensive than B3LYP.=C2=A0 =
(Of course, Gaussian does not, as far as I know, do ab initio MD, so perhap=
s BLYP&#39;s=C2=A0greater expense in Gaussian is just due to implementation=
; I&#39;m guessing BLYP is not that popular for electronic structure these =
days, so perhaps its implementation in Gaussian is not as optimized as the =
immensely popular B3LYP?)</div><div><br></div><div>B3LYP
 is a hybrid functional, while BLYP is not, but I realize that this does no=
t, by=20
itself, mean that BLYP should be less expensive than B3LYP.<br></div><div><=
br></div><div>Thanks,</div><div>Andrew</div><div><br></div><div>
<div>Andrew DeYoung, PhD</div><div>Department of Chemistry</div><div>Carneg=
ie Mellon University</div>

</div></div>

--00000000000044c76b05e4cbd804--