Auto merge of #3387 - RalfJung:bench, r=RalfJung

bors · bors · commit 850447de75ec · 2024-03-18T08:52:41.000Z
make 'invalidate' benchmark shorter This is currently by far the slowest benchmark in our suite, taking >9s, when the second slowest takes 2.7s. So let's speed this up to 2.3s, making it still the second-slowest in the benchmark suite. `@saethlin` any objections? Also, why is this called "invalidate"? It got added in #3083 but I can't figure out the point of that name even after looking at the PR.^^ There should be a comment in the benchmark explaining what it is testing.
diff --git a/bench-cargo-miri/range-iteration/Cargo.lock b/bench-cargo-miri/range-iteration/Cargo.lock
diff --git a/bench-cargo-miri/range-iteration/Cargo.toml b/bench-cargo-miri/range-iteration/Cargo.toml
@@ -1,5 +1,5 @@
 [package]
-name = "invalidate"
+name = "range-iteration"
 version = "0.1.0"
 edition = "2021"
 
diff --git a/bench-cargo-miri/range-iteration/src/main.rs b/bench-cargo-miri/range-iteration/src/main.rs
@@ -1,4 +1,5 @@
+//! This generates a lot of work for the AllocId part of the GC.
 fn main() {
     // The end of the range is just chosen to make the benchmark run for a few seconds.
-    for _ in 0..200_000 {}
+    for _ in 0..50_000 {}
 }

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,5 @@`
	`1`	`+//! This generates a lot of work for the AllocId part of the GC.`
`1`	`2`	`fn main() {`
`2`	`3`	`// The end of the range is just chosen to make the benchmark run for a few seconds.`
`3`		`- for _ in 0..200_000 {}`
	`4`	`+ for _ in 0..50_000 {}`
`4`	`5`	`}`