AMD-ZD battle ramps up a notch
3D Winbench mark "discredited"
A reader of The Register has now sent us a copy of the email AMD is sending its customers rubbishing Ziff-Davis benchmarks (see ZD rejigs benchmarks after AMD complaint). The text is below, in full. "Ziff Davis has issued a statement revealing that the 3D Processing Tests used in the 3D Winbench 99 benchmark are misleading and inaccurate. They specifically point out that the tests significantly underrate the performance of AMD K6-2 and AMD K6-III processors. "Intel has used this test suite to point out potential advantages in the Pentium III and SSE technology. In addition, some of our OEM customers, industry partners, and press have used this portion of the benchmark in their testing of AMD processors. "Note that this relates to the 3D Processing Tests which include 'transformation' and the 'lighting and transformation' tests as well as the use of the Null driver. "It does not refer to the overall 3D Winmark results designed to test graphics adapters. Please review the attached detailed bulletin and communicate this to our customers and partners as appropriate. (We used this bulletin in our story yesterday, Ed) "The message: The 3D Processing Tests in the 3D Winbench 99 benchmark have been discredited by Ziff Davis, the developer of the benchmark. Any results used to negatively portray AMD K6-2 or AMD K6-III processor performance should be disregarded as inaccurate." ®