"History has shown us that semiconductor manufacturers have thrown transistors at the problem, and up until now, things have worked out, but we are obviously hitting new walls, in terms of power dissipation (discussed at IPF), limits of ILP (also discussed during MPF), and system design challenges (again, discussed at IPF). Going the way of frequency directly affects power, EMI levels, system design levels, and others, while going the way of IPC increases die size, adds to thermal problems, and affects things in different ways."
Yep. There were power problems back in the bipolar days. That was fixed by using a different process. maybe CMOS has run out of steam. As far as the limits of ILP, we will have to see. Changes in programming techniques and algorythms may be the answer here. Or even compiler techniques.
"Up until now, there hasn't been a lot of desire to use the compiler to increase CPU performance,"
Now this is totally wrong. Compiler improvements and better algorythms have always been an active area of investigation. It just has always taken so much longer to actually show results.
Is EPIC the architecture of the future? I dunno, I don't think it really solves the basic VLIW problem of upward compatibility, it just pushes back so that it only becomes a problem every few generations. |
