Hacking processors for better performance has long been a hobbyist activity. A long time ago, it was possible to enable the math coprocessor on 386SX chips through the careful use of a drill.1 Then came the era of overclocking which at first involved hobbyists messing around with jumpers and voltages in order to change clock frequencies to obtain faster speeds on cheaper hardware. That practice became easier over time with motherboards with special BIOS’s designed to aid in overclocking. But the clockspeed age has largely passed and now the core hacking age has arrived. Rather than making separate manufacturing processes for different numbers of cores, most chip makers use a single die and simply test the produced cores. If only two of the four on-die cores work, for example, the two “lesser” cores are disabled and the processor is dubbed a dual core whereas a fully functioning die is dubbed a quad-core. Simple? Well, processor hacking has now entered the phase of re-enabling the inert cores specifically on AMD Phenom chips. Just as overclocking resulted in potentially unstable performance, no guarantees are made as to the stability of an activated core, but it’s certainly a cheaper way to extend performance when it works.
1 For those that don’t remember or weren’t alive, the SX versus DX notation indicated whether a chip included the floating point unit which could perform math functions much faster than the emulation modes resulting in better games.
Similarly tagged OmniNerd content: