The Ancient one beats the Core i7 hehehe

[rimzi]

Seriously though, I just benchmarked my old PC and received a ridiculously high CPUMark, which is above 5000. The CPU is Athlon 1.33GHz. I think it may be a bug

The results:

PassMark(TM) PerformanceTest 7.0 Evaluation Version (http://www.passmark.com)
Results generated on: Friday, March 20, 2009


Benchmark Results

Test Name: This Computer
CPU - Integer Math: 42.0
CPU - Floating Point Math: 241.1
CPU - Find Prime Numbers: 68.9
CPU - Multimedia Instructions: 72.3
CPU - Compression: 470.2
CPU - Encryption: 2.2
CPU - Physics: 13.2
CPU - String Sorting: 239.1
Graphics 2D - Solid Vectors: 3.8
Graphics 2D - Transparent Vectors: 0.5
Graphics 2D - Complex Vectors: 48.5
Graphics 2D - Fonts and Text: 37.4
Graphics 2D - Windows Interface: 77.7
Graphics 2D - Image Filters: 38.7
2D Graphics - Image Rendering: 71.3
Graphics 3D - Simple: 194.9
Graphics 3D - Medium: 27.1
Graphics 3D - Complex: 3.5
Memory - Allocate Small Block: 790.1
Memory - Read Cached: 636.2
Memory - Read Uncached: 267.2
Memory - Write: 204.5
Memory - Large RAM: 41.2
Disk - Sequential Read: 27.6
Disk - Sequential Write: 27.2
Disk - Random Seek + RW: 2.4
CD - Read: 3.6
CPU Mark: 5536.2
2D Graphics Mark: 197.5
Memory Mark: 121.3
Disk Mark: 206.8
CD Mark: 436.0
3D Graphics Mark: 99.8
PassMark Rating: 397.7

System information: This Computer
CPU Manufacturer: AuthenticAMD
Number of CPU: 1
Cores per CPU: 1
CPU Type: AMD Athlon
CPU Speed: 1334.4 MHz
Cache size: 256KB
O/S: Windows XP (32-bit)
Total RAM: 767.5 MB.
Available RAM: 507.8 MB.
Video settings: 1280x1024x32
Video driver:
DESCRIPTION: NVIDIA GeForce4 MX 4000
MANUFACTURER: NVIDIA
BIOS: Version 4.18.20.38.00
DATE: 10-6-2003
Drive Letter: C
Total Disk Space: 38.3 GBytes
Cluster Size: 4.0 KBytes
File system: NTFS

[Michael (Passmark)]

Thanks for letting us know about this. After looking into this it seems that there is indeed a bug. The problem lies in your multimedia score which is extremely high for your CPU (2-3x that of a new core i7)

For a long time now PT has tried to run an SSE test first if the CPU supports it but otherwise falls back onto a 3dNow test. However for quite a while now pretty much any CPU on the market has supported SSE instructions meaning the 3Dnow test never really got used. During this time the 3dnow test results have become somewhat out of sync with the SSE results for some reason, returning much higher results.

The end result of this is that a few very old AMD cpu's like yours that actually do make use of the 3dNow test are scoring much higher than they should be. The next release of PT will bring the 3dNow test results back in line with the SSE ones to rectify this problem.