commit | 66eaf20f3e2ce0623bd25934c3917056c0b5d848 | [log] [tgz] |
---|---|---|
author | Austin Clements <austin@google.com> | Tue Jul 10 11:25:43 2018 -0400 |
committer | Austin Clements <austin@google.com> | Tue Jul 10 15:26:53 2018 +0000 |
tree | fafdb0ecd3307c5c827b2f5320b35b291b75a538 | |
parent | faa4b5713e7266163a9e77ceabb1494e50891fe6 [diff] |
design/14951: fix minor typo Change-Id: I40c99e135f665ecce3e5c5d122ffb2b8696acd5b Reviewed-on: https://go-review.googlesource.com/123035 Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
diff --git a/design/14951-soft-heap-limit.md b/design/14951-soft-heap-limit.md index 10f0c20..a3f82f8 100644 --- a/design/14951-soft-heap-limit.md +++ b/design/14951-soft-heap-limit.md
@@ -173,7 +173,7 @@ Interestingly, while most of the variance in the original design was around GC CPU usage, that variance has been traded to the heap ratio -in this new design[o][p]. +in this new design. This is because the scheduler *does not allow* GC CPU usage to drop below 25%. Hence, the controller saturates and the inherent variance shifts to