Bench CFX/Fluent with xeon5520

bredberg · September 25, 2009, 17:04

I've run two different case (one Fluent with 1M cells and one with CFX with 2M cells) on single CPU's.
The results are:

Testcase: Fluent 6.2, MRF, ca 1M cells
Start with 200iter, measure from 150iter
Time CPU

23m30s core2 3.16GHz
24m36s xeon 3GHz
27m50s core2 3.0GHz
28m02s xeon5520(nehalem) 2.26GHz

29m43s core2 2.8GHz

Testcase: CFX 12, ca 2M cells 20iterations

Time CPU

1517s xeon5520 (nehalem) 2.26GHz

1296s xeon 3GHz
1603s core2 3.0GHz

Thus it seems that nehalem at 2.26GHz is similar to a core2 at 3GHz, I was expecting more from this CPU. Has anyone similar experiance?

/Jonas

Ps
core2 at 2.8Ghz is a oc e6300 (1.86GHz) with 800MHz DDR2

Ds

mystic_cfd · October 3, 2009, 16:32

The strength of the Nehalem becomes evident when using more than one core on the CPU. Try the benchmarks by running a 2-way parallel run on each machine, or by running the benchmark on a machine where another job is running at the same time.

regards
hp

phiBe · October 9, 2009, 01:36

Thanks for the numbers! Interesting results.

:-)

September 25, 2009, 17:04	Bench CFX/Fluent with xeon5520	#1
bredberg New Member Jonas Bredberg Join Date: Sep 2009 Posts: 1 Rep Power: 0	I've run two different case (one Fluent with 1M cells and one with CFX with 2M cells) on single CPU's. The results are: Testcase: Fluent 6.2, MRF, ca 1M cells Start with 200iter, measure from 150iter Time CPU 23m30s core2 3.16GHz 24m36s xeon 3GHz 27m50s core2 3.0GHz 28m02s xeon5520(nehalem) 2.26GHz 29m43s core2 2.8GHz Testcase: CFX 12, ca 2M cells 20iterations Time CPU 1517s xeon5520 (nehalem) 2.26GHz 1296s xeon 3GHz 1603s core2 3.0GHz Thus it seems that nehalem at 2.26GHz is similar to a core2 at 3GHz, I was expecting more from this CPU. Has anyone similar experiance? /Jonas Ps core2 at 2.8Ghz is a oc e6300 (1.86GHz) with 800MHz DDR2 Ds

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
BENCH MARKING ADVECTION-DIFFUSION	student	Main CFD Forum	0	November 12, 2008 10:09
bench marking of 3dconvection-diffusion subroutine	INAM	Main CFD Forum	0	October 31, 2008 09:06
meshing in work bench	sach	CFX	0	October 31, 2006 02:11
fluent bench mark casting problem	prashanth	FLUENT	0	July 17, 2003 02:35
Bench Mark - Vaporizing Drop Trajectory	R. Sureshkumar	Main CFD Forum	5	December 9, 1999 10:00

October 3, 2009, 16:32		#2
mystic_cfd New Member HP Kritzinger Join Date: May 2009 Location: South Africa Posts: 25 Rep Power: 16	The strength of the Nehalem becomes evident when using more than one core on the CPU. Try the benchmarks by running a 2-way parallel run on each machine, or by running the benchmark on a machine where another job is running at the same time. regards hp

October 9, 2009, 01:36		#3
phiBe New Member Philipp Beierer Join Date: Mar 2009 Posts: 5 Rep Power: 17	Thanks for the numbers! Interesting results. :-)