News for hackers. Stuff that matters.
Re:An optimization guide for assembly programmers and compil... by heyjudy on Monday February 11, 2019 @09:23PM (#19140471)
Have you ever implemented optimized microcode for a pipelined, branch-predicting processor that you also had to build? MICMAC is the architecture we used once upon a time. I won all of the course's extra credit for both smallest microcode program and fastest (fewest micro ops) because I Huffman-encoded the macro ops of the test programs and used progressive decoding.

How about a Java to MIPS compiler? In C++ and then again in Java just to be self-hosting.

Or an all software OpenGL-compatible pipeline with painter-algorithm Z-buffering and trapezoid engine? Quaternions and all that state... fun-fun.

Another mandatory project in college was a non-pre-fork()ed select() caching HTTP/1.0 proxy in POSIX C89. IPC was a PITA back then.

There are still bigger fish, but one can be that much less of dilettante by digging deeper and mastering the why, not the what or the current fashion.

There are no comments on this comment yet.