Johan van Ace's Hardware denkt dat Intel's Foster, de serverversie van de Pentium 4, uitgerust zal zijn met twee processors op één chip, zoals AMD van plan is met de Sledgehammer. Dit na aanleiding van een Intel presentatie over Foster waarin over 'Jackson Technology - on-chip multithreading support' wordt gesproken.
Paul DeMone denkt echter dat er iets anders aan de hand is, in deze thread vraagt hij zich af of de huidige Pentium 4 ook uitrust is met deze eigenschap. Aangezien Foster en Willamette dezelfde, onverklaard enorme, core gebruiken is het volgens hem waarschijnlijk dat SMT (Simultaneous Multithreading) wel in de Pentium 4 zit, maar om een of andere reden niet werkt:
Is this a form of simultaneous multithreading? Perhaps Intel has beat Compaq/DEC to the SMT punch by more than two years (EV8 doesn't ship until ~2003) The implications are interesting. If the Foster and Willamette share the same die then the Pentium 4 also potentially implements SMT. If P4 Willy does have SMT, is it:An x86 processor like the Pentium III, K7, and P4 typically only average around 1 native x86 instruction or 1.5 uOPs per clock cycle on most code or about half or less of their peak performance. If SMT can increase throughput of a decoupled execution superscalar x86 machine as well as it seems to be capable of for wide issue RISC type designs then this could represent a significant pool of currently untapped performance in the Willamette/Foster family. Perhaps this also help explains the large number of transistors and amount of die area in P4 that seemingly couldn't be accounted for on the basis on all the other known features and functions. Time will tell.
- broken in the current version of the chip
- disabled in the current version of the chip
- functional but not currently used, or
- currently used by some programs but not others