| author |  Niles Salter <Validark@pm.me> 2023-06-20 17:32:07-06:00 |
| committer |  Andrew Kelley <andrew@ziglang.org> 2023-06-22 01:57:28-07:00 |
| log | ff5850183eee854fdfe0d3f7b7242b9ff56c2116 |
| tree | 715ed346487b0f221f92e26280ea5e5b52a9254a |
| parent | 991e00c2702b31dffb7b5ada26c7b19a552502a7 |
[priority_deque] simplify & optimize isMinLayer
LLVM has trouble compiling the old implementation, (presumably) because `leading_zeros` is thought to be a `u7` rather than a `u6`, which means `63 - clz` is not equivalent to `63 ^ clz`, which means it can't deduce that the final condition can simply be flipped. (I am assuming `usize` is a `u64` here for ease of understanding, but it's the same for any power of 2)
https://zig.godbolt.org/z/Pbj4P7ob3
The new version is slightly better too because `isMinLayer(maxInt(usize))` is now well-defined behavior.1 files changed, 1 insertions(+), 3 deletions(-)
lib
lib/std/priority_dequeue.zig+1-3
| ... | ... | @@ -69,9 +69,7 @@ pub fn PriorityDequeue(comptime T: type, comptime Context: type, comptime compar |
| 69 | 69 | // The first element is on a min layer; |
| 70 | 70 | // next two are on a max layer; |
| 71 | 71 | // next four are on a min layer, and so on. |
| 72 | const leading_zeros = @clz(index + 1); | |
| 73 | const highest_set_bit = @bitSizeOf(usize) - 1 - leading_zeros; | |
| 74 | return (highest_set_bit & 1) == 0; | |
| 72 | return 1 == @clz(index +% 1) & 1; | |
| 75 | 73 | } |
| 76 | 74 | |
| 77 | 75 | fn nextIsMinLayer(self: Self) bool { |