Simulation run times should depend upon the processor speed, not the hard drive, right? The hard drive is being referenced, particularly towards the end when the output is written, but the majority of time is spent doing calculations. So, I would imagine it is dependent upon processor speed.
---In EnergyPlus_Support@xxxxxxxxxxxxxxx, <jglazer@...> wrote:
I recently purchased a solid state drive in hopes that it
would increase the speed of EnergyPlus 8.1 simulations. I
ran some benchmarks of the simulation files on either the
solid state drive or a disk drive and the EnergyPlus program
on the solid state drive or disk drive. These were run from
EP-Launch as a group and the temporary files were located
where the input files are. I set EP-Launch to use one
simulation process at a time. The operating system was on
the disk drive. I took the elapsed time from the ERR file
and also reported the timestamp of the first and last file
to be created by the group of simulations. See the first
four columns of results in the attached spreadsheet.
The conclusion surprised me: the solid state drive made
almost no difference.
While I was at it, I ran the same simulations on some other
computers I have.
To do the benchmarking, I just used the 16 reference
buildings that come with EnergyPlus 8.1 but made them all
annual simulations, also attached.
Feel free to add your own test results to the spreadsheet
and redistribute but please describe the computer
configuration. Can anyone's computer beat 1415 seconds?
Jason
--
Jason Glazer, P.E., GARD Analytics, 90.1 ECB chair
Admin for onebuilding.org building performance mailing lists