On July 31, 2000, exactly four weeks ago from now, Intel released their latest Pentium III processor, which happened to be specified for a clock speed of 1.13 GHz, thus being the x86 processor with the highest clock speed ever released.
Business As Usual?
A few days before that I had received my very test sample from Intel and I was getting ready to have another benchmarking and writing weekend to get the article about this CPU ready for the launch day. I considered the article as being a rather boring and straight forward piece, which was merely supposed to state that Intel's new 1.13 GHz Pentium III was now the fastest x86 processor in the world, as usual backed up with the quality testing and benchmarking my readers are used to.
The New Processor Is Making Problems
It turned out pretty soon that the test sample that I had received was unable to run benchmarks reliably on any of the platforms that I had available. I tried a motherboard with Intel's i820 chipset, the Asus P3C-L, with VIA's Apollo Pro 133A chipset, the Asus P3V4X, with Intel's i840 chipset, Intel's very own OR840, and finally with Intel's good old 440BX chipset, the Asus P3B-F. The results were the same on each platform. The processor was so unstable that I had to call off the benchmarking completely.
Was The Missing Micro Code Update To Blame?
In the first instance I believed that Intel's new processor required a new micro code update, which is the little software that is loaded inside the CPU right after the system starts. This software is supposed to cover up for bugs in the processor core. I assumed that the only reason for the failure of a brand new Intel processor had to be a software issue and the only explanation was indeed the micro code.
I was pretty upset about this situation, because I blamed Intel for the failure in supplying me with their latest micro code update. I hadn't received the motherboard Intel was supplying with the processor sample and I assumed that the micro code update was at least available on this very motherboard. Facing the fact that I was unable to supply my readers with actual benchmark data, I sat down and wrote the