80386 Barrel Shifter

80386 Barrel Shifter(nand2mario.github.io)

84 points by jamesbowman 145 days ago | 12 comments

rep_lodsb 143 days ago |

Implementing rotate through carry like that was a really bad decision IMO - it's almost never by more than one bit left or right at a time, and this could be done much more efficiently than with the constant-time code which is only faster when the count is > 6.

Is the full microcode available anywhere?

ajenner 143 days ago | |

I haven't published it yet as there are still some rough edges to clear up, but if you email me (andrew@reenigne.org) I'll send you the current work-in-progress (the same one that nand2mario is working from).

kjs3 143 days ago | |

Since the shifter is also used for bit tests, the 'most things are a 1-bit shift' might not be the case. Perhaps they did the analysis and it made sense.

rep_lodsb 143 days ago | | |

There are separate opcodes for shift/rotate by 1, by CL, or by an immediate operand. Those are decoded to separate microcode entry points, so they could have at least optimized the "RCL/RCR x,1" case.

And the microcode for bit test has to be different anyway.

cbsmith 141 days ago | |

Except that there are tremendous advantages to constant-time execution, not the least of which is protection from timing security attacks/information leakage (which admittedly were less of a concern back then). Sure you can get the one instruction executed for the <6 case faster, but the transistor budget for that isn't worth it, particularly if you pipeline the execution into stages. It makes optimization far more complex...

cmovq 142 days ago |

> For memory operands, there's an additional twist: the bit index is a signed offset that can address bits outside the nominal operand. A bit index of 35 on a dword accesses bit 3 of the next dword in memory.

I wonder what is the use case for testing a bit outside of the memory address given.

rep_lodsb 142 days ago | |

So you can have bit arrays of any length in memory, rather than just 32 bits in a register.

cmovq 142 days ago | | |

That makes sense. LLVM could probably do better here by using the memory operand version:

https://godbolt.org/z/jeqbaPsMz

juancn 142 days ago | |

It was probably easier to just implement it that way, given that the barrel shifter is 64 bits wide.