Slashdot Mirror
David Patterson Says It's Time for New Computer Architectures and Software Languages (ieee.org)
Tekla S. Perry, writing for IEEE Spectrum: David Patterson -- University of California professor, Google engineer, and RISC pioneer -- says there's no better time than now to be a computer architect. That's because Moore's Law really is over, he says : "We are now a factor of 15 behind where we should be if Moore's Law were still operative. We are in the post -- Moore's Law era." This means, Patterson told engineers attending the 2018 @Scale Conference held in San Jose last week, that "we're at the end of the performance scaling that we are used to. When performance doubled every 18 months, people would throw out their desktop computers that were working fine because a friend's new computer was so much faster." But last year, he said, "single program performance only grew 3 percent, so it's doubling every 20 years. If you are just sitting there waiting for chips to get faster, you are going to have to wait a long time."
As an SGI collector and fan of Unix hardware it pains me to say this, but: get real RISC is shit and so are it's engineers. This guy didn't get enough of being completely wrong in the 1990's he's gotta pile some more bullshit on top. First off, look around you, see any RISC processors that are worth a shit? I'm sure many will say ARM because they don't realize that the compute performance of an ARM chip is utter shit when seen in the light of what CISC has done. Yes, Intel and others have done silly-shit with execution pipelines and other now-dangerous-tricks. However, that still doesn't detract from the big flaw in RISC - the ISA. RISC was supposed to be just fine because everyone would just write everything in C and never get horrified at how idiotic the ISA was underneath. Compilers don't tend to use as many tricks as ASM coders do/can because any tricks they use have to be absolutely rigidly repeatable. A human ASM coder knows when he can use certain techniques and when they are better avoided. So, compilers did get better, and they sorta kinda partially made up for the shitty internal design of their RISC chips. Trouble was, RISC solved a problem that was temporary - lack of space on the silicon. The advances in lithography since the 1990's has made RISC seem pretty damn ridiculous. In fact, chip makers had so much space from the smaller litho, they didn't even know what to do with it! Heaven forbid they hire more engineers or do anything interesting - so they just stamped out more cores on the same silicon. These days, we are drowning in so much geography on the chips that the problems RISC aimed to solve are totally irrelevant. Anyone who cut their teeth making RISC chips is already extremely suspect to me. Though I can see why they'd call for new architectures, they are used to ones that suck balls.