Unique serial number exists in all 25 micron Intel chips
Celerons, PIIs and Xeons as well, now Intel responds
Update Two An architect who currently works for Intel US has now confirmed what we were beginning to suspect all along -- every .25 micron Intel chip has the processor serial number (PSN). The chip designer, who revealed the news under strict conditions of anyonymity, said: "Any .25 micron core including the PII, all Celerons and all current Xeons have had serialisation ever since .25 micron technology started." The exception is the .25 Tillamook P5 MMX, he said. Since we posted our original story on Saturday, however, more information has come to light. Intel's own document, The Pentium II Xeon Processor Server Platform System Management Guide which is order no. #243835-001 confirms the existence of the PSN in the Xeon. See the processor's PIROM, table one. It's called Processor Electronic Signature. Table One is on page 20. Also see 184.108.40.206 on page 19. This morning (Monday) we placed a call to Intel and it promised it would get back to us. We now have that statement. A representative said: "A prototype of the circuitry has been present in some PII and Celeron parts, purely for validation and testing purposes. "It wasn't on all, by any means. It was disabled during the manufacturing process." He would not say which proportion of the PII core had the serial number, nor would he say whether the feature is disabled in BIOS. But he did confirm that the "erratum" we wrote about last week was essentially a batch of Dixon core chips which escaped this process. This means that Intel is and was shipping products with the serialisation number switched off. The only exception is the Tillamook P5 .25 micron parts, the Intel chip architect, who originally tipped us off, said. But why didn't Intel tell us and the whole world what it was doing then, and why? It seems that maybe it just thought it would lump in the PSN announcement to co-incide with the introduction of Katmai-PIII. So when we first broke the story about the Mobile PII with Dixon core, the only "erratum" was that Intel forgot to turn it off. Intriguing. How US bodies pushing for a boycott of Intel parts will react to all of this is even more interesting. ® Related Stories Dixon Intel PIIs have serial numbers -- it's a bug Intel stealth shipping Coppermine on Dixon parts? Unique ID customer numbers on mobile PIIs no accident Pentium II Xeons have serial numbers too Gelsinger "brains" behind shipping chips with PSN switched on Opinion: Intel the Chipzilla just naive
Sponsored: What next after Netezza?