G09

go back to Main Page, Computational Resources, Chemistry & More, Computational Codes, GAUSSIAN

Gaussian09 vs Gaussian03

Scaling New Functional Testing

Tests calculations performed at the CESCA supercomputer on parallel8 queue (prades) with 14000MB ram.

Calcfc, opt to TS, and frequency.

Job has 687 basis functions with B3LYP method.

     Step	           G03	          G09
calcfc to 2nd l103	7h 54m 08s     3h 42m 10s
2nd opt step	            8m 00s	   9m 16s
3rd opt step	            6m 06s	   7m 06s
freq	                8h 30m 25s	4h 6m 15s
Total calc time        16h 38m 39s	8h 4m 47s

So the frequency calculation is about 2x faster in Gaussian09.

The SCF takes longer, but it is using the new GEDIIS algorithm (vs GDIIS in G03 which now doesn't exist). They say GEDIIS should give better performance, especially for calculations not so close to convergence as this example.

Testing new functionals with G09

Job has 687 basis functions, starting structure from B3LYP method in G03.

     Step	              Functional	
   measured              M06	         B97D
calcfc to 2nd l103    4h 49m 34s      1h 59m 42s
2nd opt step	         13m 36s          7m 42s
3rd opt step	         13m 37s          7m 15s
Total opt time	     19h 05m 32s      3h 36m 35s
No. steps	             89	             32
avg opt step time	    772s	    406s
freq	              5h 27m 59s      2h 11m 48s
Total calc time      29h 50m 18s      7h 48m 05s

The M06 functional is 33% slower than B3LYP (based on the frequency times), but still faster than B3LYP in G03. Grimme's B97D functional is very fast, and also takes less steps to optimize in this case too.

Solvation models, Polarizable Continuum Model (PCM)

As in Gaussian03, the SCRF keyword requests that a calculation be performed in the presence of a solvent by placing the solute in a cavity within the solvent reaction field.

• The integral equation formalism variant IEFPCM, is the default SCRF method. It has not changed from Gaussian03, BUT the formalism used and its implementation has changed. That is: you will get different results, using the same method in G03 and G09.

• The default RADII used in G03 was UAO , whilst now is UFF with spheres placed by default on all Hydrogen atoms. No need to use the SPHEREONH= keyword, except if you use UAO or another radii model that does not have them explicitly.

• It seems to give less "convergence failure" problems, but maybe its too early to say so.

Calculation of solvent effects

Optimization in solvent