From cc7d8014d720ac800f868a0b7ec854f4a5036e3f Mon Sep 17 00:00:00 2001
From: Scott McMurray <scottmcm@users.noreply.github.com>
Date: Tue, 9 Nov 2021 21:43:20 -0800
Subject: Specialize array cloning for Copy types

Because after PR 86041, the optimizer no longer load-merges at the LLVM IR level, which might be part of the perf loss.  (I'll run perf and see if this makes a difference.)

Also I added a codegen test so this hopefully won't regress in future -- it passes on stable and with my change here, but not on the 2021-11-09 nightly.
---
 library/core/src/array/mod.rs | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

(limited to 'library/core/src/array')

diff --git a/library/core/src/array/mod.rs b/library/core/src/array/mod.rs
index 089f1f36390..e8e0cbf84b3 100644
--- a/library/core/src/array/mod.rs
+++ b/library/core/src/array/mod.rs
@@ -339,9 +339,7 @@ impl<T: Copy, const N: usize> Copy for [T; N] {}
 impl<T: Clone, const N: usize> Clone for [T; N] {
     #[inline]
     fn clone(&self) -> Self {
-        // SAFETY: we know for certain that this iterator will yield exactly `N`
-        // items.
-        unsafe { collect_into_array_unchecked(&mut self.iter().cloned()) }
+        SpecArrayClone::clone(self)
     }
 
     #[inline]
@@ -350,6 +348,27 @@ impl<T: Clone, const N: usize> Clone for [T; N] {
     }
 }
 
+#[cfg(not(bootstrap))]
+trait SpecArrayClone: Clone {
+    fn clone<const N: usize>(array: &[Self; N]) -> [Self; N];
+}
+
+#[cfg(not(bootstrap))]
+impl<T: Clone> SpecArrayClone for T {
+    default fn clone<const N: usize>(array: &[T; N]) -> [T; N] {
+        // SAFETY: we know for certain that this iterator will yield exactly `N`
+        // items.
+        unsafe { collect_into_array_unchecked(&mut array.iter().cloned()) }
+    }
+}
+
+#[cfg(not(bootstrap))]
+impl<T: Copy> SpecArrayClone for T {
+    fn clone<const N: usize>(array: &[T; N]) -> [T; N] {
+        *array
+    }
+}
+
 // The Default impls cannot be done with const generics because `[T; 0]` doesn't
 // require Default to be implemented, and having different impl blocks for
 // different numbers isn't supported yet.
-- 
cgit 1.4.1-3-g733a5


From 5b115fcb6839b94a14fb77acdd738281b4bb3fae Mon Sep 17 00:00:00 2001
From: Scott McMurray <scottmcm@users.noreply.github.com>
Date: Wed, 10 Nov 2021 11:57:14 -0800
Subject: Moar #[inline]

---
 library/core/src/array/mod.rs | 2 ++
 1 file changed, 2 insertions(+)

(limited to 'library/core/src/array')

diff --git a/library/core/src/array/mod.rs b/library/core/src/array/mod.rs
index e8e0cbf84b3..09bb4519170 100644
--- a/library/core/src/array/mod.rs
+++ b/library/core/src/array/mod.rs
@@ -355,6 +355,7 @@ trait SpecArrayClone: Clone {
 
 #[cfg(not(bootstrap))]
 impl<T: Clone> SpecArrayClone for T {
+    #[inline]
     default fn clone<const N: usize>(array: &[T; N]) -> [T; N] {
         // SAFETY: we know for certain that this iterator will yield exactly `N`
         // items.
@@ -364,6 +365,7 @@ impl<T: Clone> SpecArrayClone for T {
 
 #[cfg(not(bootstrap))]
 impl<T: Copy> SpecArrayClone for T {
+    #[inline]
     fn clone<const N: usize>(array: &[T; N]) -> [T; N] {
         *array
     }
-- 
cgit 1.4.1-3-g733a5