<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body text="#000000" bgcolor="#FFFFFF">
I am using the following command to fully saturate one core:<br>
<br>
<b>dd if=/dev/zero of=/dev/null bs=4M count=500000<br>
<br>
</b>The following results were achieved on different systems<br>
<b><br>
</b><b>System A : <br>
</b>Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz<br>
6-cores - single chip<br>
Turbo up to 3.80Ghz<br>
OS: Gentoo 64bit 3.2.1-gentoo-r2<br>
2097152000000 bytes (2.1 TB) copied, 100.911 s, <b>20.8 GB/s</b><br>
<br>
<b>System B :</b><br>
Intel(R) Xeon(R) CPU X3470 @ 2.93GHz<br>
4-cores - single chip<br>
OS: Redhat 5.7 64 bit 2.6.18-308.el5 <br>
2097152000000 bytes (2.1 TB) copied, 80.6255 seconds, <b>26.0 GB/s</b><br>
<br>
To my understanding this test is not dependent on IO so disk
performance is out of the picture. My question is how can a 2-3 year
old chip Xeon X3470 (System B) running almost a whole Ghz slower
then the i7 (System A) produce a result which is so much better in
comparison. Does this indicate a problem on the i7 system ? Is there
anything I can configure / check to improve the performance ?<br>
<br>
Any thoughts <br>
Alex<br>
</body>
</html>