diff --git a/handouts/src/18-cachot.md b/handouts/src/18-cachot.md index de6223b25aaa1590914ec459ef7289cabe75acac..0ec1538836ae2525daa0b1c0d01a8c171b7449e5 100644 --- a/handouts/src/18-cachot.md +++ b/handouts/src/18-cachot.md @@ -57,8 +57,8 @@ CPUs truly loaded or stored one single SIMD vector whenever we ask them to do so. In the case of input data, the pattern looks like this, where pink represents -data that is newly brought into the CPU cache, and pink represents data that can -be reused from previous iterations: +data that is newly brought into the CPU cache, and green represents data that +can be reused from previous iterations: 