3.4Ghz slower than baseline 3.0Ghz

[pdecker]

I put together a new pc and ran the benchmark included with the Intel motherboard. I chose the 3.0Ghz HT baseline since mine is a 3.4Ghz HT processor. My pc is slower in almost all of the tests. Video is understandable because I didn't spend a bunch on the card. The memory and processor though should be better. Here is a screen shot of the test.



Let me know if you think I may have a defective product. I did update the BIOS and all drivers before running the tests. In the BIOS I chose the "optimal" settings. I also closed all the programs I could before running the benchmark (anti-virus, anti-spyware, etc.).

Thanks Guys

[Tahoe]

I have observed this same situation with a 3.4G extreme processor; I noticed that if I compare to the 3.06G with HT to my 3.4G EE, the MMX scores and other floating point based scores are lower. If I compare to a 3.06G none HT cpu, all the scores are higher on the 3.4GEE cpu. In reviewing the report; I noticed passmark is only detecting 1 CPU for the 3.4GEE cpu. That is a likely source for this reporting issue. The Desktop control center software (which is where I got the passmark software) reports 2 CPUs like it should. This could be a bug in the passmark detection code. 3.4G P4 extreme processors are based on the Gallatin core not the Northwood or prescott cores. Any ideas how we get a passmark person to look into this?

[Ian (PassMark)]

This does look a bit unusual. To help us investigate, could you please send us an email at our "help" email address and we can send you some software to determine whats going on.

Thanks.
Ian
PassMark

[Tahoe]

Be sure to edit preferences when testing HT CPUs w v1.0

PDecker et al, I improved the cpu scores for my 3.4 P4 by taking Ian's advice by editing the preferences for processes to 2. the Version 1.0 test software enables you to test one or both pipes in a P4. It doesn't automatically test both. Future versions of performance index will probably detect and test both pipelines

[David (PassMark)]

PerformanceIndex V1.0 CPU scores low

The default preference settings of PerformanceIndex V1.0 are such that the benefits from Hyper Threading are not measured. To compare results that take the benefits of Hyper Threading into account, users should change the Number of Processes for the CPU tests in the Preferences window to be equal to the number of logical processors (eg from 1 to 2, or 2 to 4). A future version will automatically default this value in preferences.

Regards,
Ian

[JeremyVnc]

I'm having the exact same problem in the exact same situation. I too noticed that it only saw one processor instead of two. I have a p4 550 (3.4 GHz) prescott core. I went all out with this computer except for the graphics card (radeon x300) but I'm saving up for a x850 xt. So anyone know whats up with these?

[JeremyVnc]

That helped integer math, floating decreased, MMX went way up, SSE went way down, compression went way up, encryption was a bit better, and image rotation went way up. It only went up by about 5 points though the CPU score went up 30 points which is good enough for me.