Petz & WATSONYOUTH, re: <WATSONYOUTH, thanks for your help on guard bands in spec-ing processor speeds. I expect that Athlon overclocks better than the normal guard band % if a very good cooling solution is implemented. Heat, not switching speed is the limiting factor with the Athlon; if the temp can be kept below 50 degC, the switching performance improves also.>
WATSONYOUTH, let me also thank you for pointing the guard bands. Technically, a cherry picked 800MHz production part, (10% guard banding) could have a raw speed of 880MHZ or even higher. To explain the 'higher' part, a part with a raw speed of 879MHZ could barely fail 800MHZ speed bin and could be binned as 750MHz production part. So a cherry picked 750MHZ prod part could have a max raw speed of 879MHZ, which is about 16% higher than the prod spec. Extending the same logic to 800MHz, one could conceivably have an 800MHz prod part with a raw speed of (if the next speed bin is 850MHz) 934MHz. In real testing, some tests could impose double gaurd banding on the DUT - ex: Inputs may be supplied with clocks of low raising/falling edges than the spec (guard banding) while the output pins are tested for faster raising/falling clock edges than the spec limits(gaurd banding). If you take this into consideration, an 800MHz production part can technically have more than 1GHz of raw clock speed, if(as Petz mentioned) the chip architecture permits the high speeds and the temperature is kept within the limits.
regards,
Goutama |
