Hacker News Reader

I think for x86-64, the 32-bit immediate operands totally qualify as CISC-style because they contribute to the variable length nature of the instruction encoding. And GCC uses them all the time, of course. GCC has a surprising tendency to use the string instructions for memset and memcpy, too. In the other direction, without -march=x86-64-v3, GCC cannot even use most of the more RISC-style three-operand instructions.

View on HN

The 32-bit operands are actually hurting code density most of the time. An instruction that would take 4 bytes in a RISC ISA takes at least 5 bytes in a CISC ISA.

Same thing with the 32-bit offsets for x86_64 conditional branches.

Edit: With "CISC style" I meant those instructions that use implicit registers instead of GPRs and/or do fancy multi-operation stuff, like LOOP or POPA/PUSHA etc.

Reply on news.ycombinator.com