At 09:54 p.m. 13/01/2004, you wrote:
Hi all,
I'm just collating some results from a test run I did using a Monte Carlo code on two different computers, one was a 1.5GHz P4 running Windows 2000 and the other was a 2.0GHz P4 running RH8.
The Windows machine was faster than the RH machine. I ran the test again and
got the same result. The difference was only 10%, but it was a surprise. Has
anyone got any idea why there was so much of a difference?
The test case didn't use much memory and the version of the program was the
same in both cases (apart from running on different OSs of course).
You have not really given enough information to explain the difference.
If this is anything but a troll then you would need to provide a lot more information for the question to be able to be answered but is probably rather off topic.
For a real comparison, it would be advisable to install the OS on the same hardware and then document what installation you preformed, the tests and the results and then opinions may be able to be offered after the hardware differences have been eliminated.
Also of note, for some things windows will preform "better" than other OS's (Linux only being one) - especially in an untuned environment. Some of the advantages of the unix type operating systems is the ability to cut back any un-necessary cruft to gain more CPU cycles, but these tend to require you to know how the OS works.
Just a side note - You CANNOT test something like this on two different PCs. I'm not advocating Linux as being faster or whatever, but the difference between PCs is usually HUGE.
The machines we build here are on average 20% faster than any other white box PC we get in here with the same CPU. Unfortunately most people always think that a 2.6GHz CPU PC is faster than a 2.4GHz CPU PC but there are SO MANY different factors which make this not so. Look at the amount of possible configurations:
Motherboard (Some are simply faster than other due to better IO etc)
RAM (Dual Channel -vs- Single Channel, DDR SDRAM -vs- RDRAM, Speed of the DDR SDRAM (266, 333, 400MHz), CL2 -vs- CL2.5 -vs CL3)
Hard disk (5400RPM -vs- 7200RPM -vs- 10000RPM -vs- 15000RPM, 2Mb Cache -vs- 8Mb Cache, ATA33 -vs- ATA66 -vs ATA100 -vs- ATA133 -vs ATA150 (SerialATA))
CPU (400, 533, or 800MHz FSB with or without HyperThreading support and enabled in the Motherboard & OS)
If you had given me results on two 100% identical PCs, with only the CPU speed differering, fair enough, but you haven't specified, and I know for a fact that a 1.5GHz CPU has a VERY different supporting architecture required than a 2.0GHz one.
Steve is right, for some things Windows IS faster, but in this case I'd say the test is very tainted anyway.
Regards, Ed.
-- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
