was the major use-case, and to update the dep-graph. Other kinds of
predicates are now excluded from the cache because there is no easy way
to make a good dep-graph node for them, and because they are not
believed to be that useful. :)
Fixes#30741. (However, the test still gives wrong result for trans,
for an independent reason which is fixed in the next commit.)
Fix spacing style of `T: Bound` in docs
The space between `T` and `Bound` is the typical style used in code and
produced by rustdoc's rendering. Fixed first in Reflect's docs and then
I fixed all occurrences in docs I could find.
Sorry for nitpicking, but I think the example of the expanded macro should be wrapped inside a pair of curly braces to match the macro definition. Also the current example triggers a variable redefinition error.
We currently pass generic as the CPU to LLVM. This results in worse
than required code generation. On little endian, which is only POWER8,
we avoid many POWER4 and newer instructions.
Pass ppc64 and ppc64le instead.
Restore indexed formulation of clone_from_slice
For good codegen here, we need a lock step iteration where the loop
bound is only checked once per iteration; .zip() unfortunately does not
optimize this way.
If we use a counted loop, and make sure that llvm sees that the bounds
check condition is the same as the loop bound condition, the bounds
checks are optimized out. For this reason we need to slice `from`
(apparently) redundantly.
This commit restores the old formulation of clone_from_slice. In this
shape, clone_from_slice will again optimize into calling memcpy where possible
(for example for &[u8] or &[i32]).
For good codegen here, we need a lock step iteration where the loop
bound is only checked once per iteration; .zip() unfortunately does not
optimize this way.
If we use a counted loop, and make sure that llvm sees that the bounds
check condition is the same as the loop bound condition, the bounds
checks are optimized out. For this reason we need to slice `from`
(apparently) redundantly.
This commit restores the old formulation of clone_from_slice. In this
shape, clone_from_slice will again optimize into calling memcpy where possible
(for example for &[u8] or &[i32]).
The space between `T` and `Bound` is the typical style used in code and
produced by rustdoc's rendering. Fixed first in Reflect's docs and then
I fixed all occurrences in docs I could find.
We currently pass generic as the CPU to LLVM. This results in worse
than required code generation. On little endian, which is only POWER8,
we avoid many POWER4 and newer instructions.
Pass ppc64 and ppc64le instead.
test: Increase resolution of MB/s stat for bench runs close to 1 second
MB/s was based on the number of iterations performed in a second, when
the iteration duration nears 1 second (1e9 ns), the resolution of the
MB/s stat decreases.
* Moved semicolon to the right place in the `let` statement in the ZST section.
* Fixed the missing ZST additions for `RawValIter<T>` from this section in the final code section.
