Core i7 920 2.66ghz vs. 2x Xeon e5420 2.5 ghz

CaglarCoskun · January 31, 2013, 08:50

Hi
I have a computer having a Core i7 920 processor (4 cores and 8 threads) with 16gb of RAM, running a FLUENT case file having about 14 million meshes.
I have another computer, which is a Dell workstation having dual Xeon e5420 processors (4 cores and 4 threads per processor) with 24gb RAM and this computer runs another FLUENT case having about 16 million meshes.
These case files are physically very similar. The problem is the second computer (Dell) is considerably slower than the first one. I measured the iteration times, and the first computer (i7) calculates an iteration in about 40 seconds and the other one calculates about 190 seconds! The parallel processing is enabled in both case files to 8 processors.
Is this normal? or what might be the cause of this slowness?

Note: The operating systems and ANSYS versions are also the same (Windows 7 Professional and ANSYS 14.0)

scipy · January 31, 2013, 13:36

Old E5420 Xeon = single channel DDR2 memory.
Newer i7 920 = triple channel DDR3 memory.

That's the difference.

And before you ask, yes - it can be that drastic/high/huge. However, a newer i5 (22 nm) with 4 cores and dual channel DDR3 would be faster than both of these (and more power efficient), while a new i7 3930K (32 nm) with quad channel DDR3 would literally be several times faster than both of your PCs combined.

Time to trade in the old hardware, it only looks impressive on paper.

CaglarCoskun · January 31, 2013, 13:57

Hi

I know the i7 is newer and introduce some new technologies, but the performance gap between them in CFD calculations is really surprising for me =)

Another strange point is, according to the Windows experience index, while the subscore of 2x Xeon is 7.6, the subscore of i7 920 is 7.4 =) It may not be a reliable tool for benchmarking, but the result is intresting.

scipy · January 31, 2013, 14:56

Windows experience index is determined by the lowest subscore in any category, so it's not that trustworthy.

As far as CFD performance goes - the memory speed and bandwith is the most prominent factor, and this is where the newer stuff is mostly different and where most of the improvements come from.

January 31, 2013, 08:50	Core i7 920 2.66ghz vs. 2x Xeon e5420 2.5 ghz	#1
CaglarCoskun New Member Çağlar Coşkun Join Date: Feb 2012 Posts: 20 Rep Power: 14	Hi I have a computer having a Core i7 920 processor (4 cores and 8 threads) with 16gb of RAM, running a FLUENT case file having about 14 million meshes. I have another computer, which is a Dell workstation having dual Xeon e5420 processors (4 cores and 4 threads per processor) with 24gb RAM and this computer runs another FLUENT case having about 16 million meshes. These case files are physically very similar. The problem is the second computer (Dell) is considerably slower than the first one. I measured the iteration times, and the first computer (i7) calculates an iteration in about 40 seconds and the other one calculates about 190 seconds! The parallel processing is enabled in both case files to 8 processors. Is this normal? or what might be the cause of this slowness? Note: The operating systems and ANSYS versions are also the same (Windows 7 Professional and ANSYS 14.0) Last edited by CaglarCoskun; January 31, 2013 at 13:26.

January 31, 2013, 13:36		#2
scipy Senior Member Alex Pasic Join Date: Aug 2011 Location: Croatia Posts: 199 Rep Power: 16	Old E5420 Xeon = single channel DDR2 memory. Newer i7 920 = triple channel DDR3 memory. That's the difference. And before you ask, yes - it can be that drastic/high/huge. However, a newer i5 (22 nm) with 4 cores and dual channel DDR3 would be faster than both of these (and more power efficient), while a new i7 3930K (32 nm) with quad channel DDR3 would literally be several times faster than both of your PCs combined. Time to trade in the old hardware, it only looks impressive on paper. Ravindra Shende and Teja1687 like this.

January 31, 2013, 14:56		#4
scipy Senior Member Alex Pasic Join Date: Aug 2011 Location: Croatia Posts: 199 Rep Power: 16	Windows experience index is determined by the lowest subscore in any category, so it's not that trustworthy. As far as CFD performance goes - the memory speed and bandwith is the most prominent factor, and this is where the newer stuff is mostly different and where most of the improvements come from. Teja1687 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
solving a conduction problem in FLUENT using UDF	Avin2407	Fluent UDF and Scheme Programming	1	March 13, 2015 03:02
Superlinear speedup in OpenFOAM 13	msrinath80	OpenFOAM Running, Solving & CFD	18	March 3, 2015 06:36
4 core & 8 core: same 64bit XP, Xeon, x5570, 2.93GHZ	leeruoyun	CFX	4	August 19, 2009 01:47
Dual Xeon PIV 3.8Ghz vs 2x Dual Core E5130 2.0 GHz	Michiel	Hardware	4	July 31, 2009 07:06
Core 2 Duo vc. Dual Core Xeon 3060	Jan	Main CFD Forum	6	February 20, 2007 07:21

January 31, 2013, 13:57		#3
CaglarCoskun New Member Çağlar Coşkun Join Date: Feb 2012 Posts: 20 Rep Power: 14	Hi I know the i7 is newer and introduce some new technologies, but the performance gap between them in CFD calculations is really surprising for me =) Another strange point is, according to the Windows experience index, while the subscore of 2x Xeon is 7.6, the subscore of i7 920 is 7.4 =) It may not be a reliable tool for benchmarking, but the result is intresting.