Implement vectorized matmul

1 files changed, 10 insertions, 0 deletions

diff --git a/README.md b/README.md
index 6f49640..1cd42ec 100644
--- a/README.md
+++ b/README.md
@@ -2,6 +2,8 @@
 
 This repo contains a small C++ dense numerical linear algebra library for `double`, with a companion experiments directory for evaluating performance.
 
+The current matmul uses a vectorized dot-product kernel. The implementation supports compile-time SIMD backends for AVX, AVX2, AVX512, NEON (AArch64/ARM64).
+
 ## Build
 
 ```bash
@@ -9,6 +11,14 @@ cmake -S . -B build
 cmake --build build
 ```
 
+On x86, you can explicitly choose a matmul SIMD target at configure time:
+
+```bash
+cmake -S . -B build -DLINEAR_ALGEBRA_SIMD=AVX2
+```
+
+Valid values are `AUTO`, `NONE`, `AVX`, `AVX2`, and `AVX512`. `AUTO` uses the compiler's current target. `NONE` forces the scalar fallback.
+
 ## Run tests
 
 ```bash