Auto merge of #41418 - hirschenberger:prefetch-intrinsic, r=nagisa

Adding support for the llvm `prefetch` intrinsic

Optimize `slice::binary_search` by using prefetching.

1 files changed, 47 insertions, 0 deletions

diff --git a/src/libcore/intrinsics.rs b/src/libcore/intrinsics.rs
index 9f1870e56d3..3566bbdebc2 100644
--- a/src/libcore/intrinsics.rs
+++ b/src/libcore/intrinsics.rs
@@ -564,8 +564,55 @@ extern "rust-intrinsic" {
     pub fn atomic_umax_rel<T>(dst: *mut T, src: T) -> T;
     pub fn atomic_umax_acqrel<T>(dst: *mut T, src: T) -> T;
     pub fn atomic_umax_relaxed<T>(dst: *mut T, src: T) -> T;
+
+    /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
+    /// if supported; otherwise, it is a noop.
+    /// Prefetches have no effect on the behavior of the program but can change its performance
+    /// characteristics.
+    ///
+    /// The `locality` argument must be a constant integer and is a temporal locality specifier
+    /// ranging from (0) - no locality, to (3) - extremely local keep in cache
+    #[cfg(not(stage0))]
+    pub fn prefetch_read_data<T>(data: *const T, locality: i32);
+    /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
+    /// if supported; otherwise, it is a noop.
+    /// Prefetches have no effect on the behavior of the program but can change its performance
+    /// characteristics.
+    ///
+    /// The `locality` argument must be a constant integer and is a temporal locality specifier
+    /// ranging from (0) - no locality, to (3) - extremely local keep in cache
+    #[cfg(not(stage0))]
+    pub fn prefetch_write_data<T>(data: *const T, locality: i32);
+    /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
+    /// if supported; otherwise, it is a noop.
+    /// Prefetches have no effect on the behavior of the program but can change its performance
+    /// characteristics.
+    ///
+    /// The `locality` argument must be a constant integer and is a temporal locality specifier
+    /// ranging from (0) - no locality, to (3) - extremely local keep in cache
+    #[cfg(not(stage0))]
+    pub fn prefetch_read_instruction<T>(data: *const T, locality: i32);
+    /// The `prefetch` intrinsic is a hint to the code generator to insert a prefetch instruction
+    /// if supported; otherwise, it is a noop.
+    /// Prefetches have no effect on the behavior of the program but can change its performance
+    /// characteristics.
+    ///
+    /// The `locality` argument must be a constant integer and is a temporal locality specifier
+    /// ranging from (0) - no locality, to (3) - extremely local keep in cache
+    #[cfg(not(stage0))]
+    pub fn prefetch_write_instruction<T>(data: *const T, locality: i32);
 }
 
+// Empty bootstrap implementations for stage0 compilation
+#[cfg(stage0)]
+pub fn prefetch_read_data<T>(_data: *const T, _locality: i32) { /* EMPTY */ }
+#[cfg(stage0)]
+pub fn prefetch_write_data<T>(_data: *const T, _locality: i32) { /* EMPTY */ }
+#[cfg(stage0)]
+pub fn prefetch_read_instruction<T>(_data: *const T, _locality: i32) { /* EMPTY */ }
+#[cfg(stage0)]
+pub fn prefetch_write_instruction<T>(_data: *const T, _locality: i32) { /* EMPTY */ }
+
 extern "rust-intrinsic" {
 
     pub fn atomic_fence();