diff --git a/Cargo.toml b/Cargo.toml
index df49da2..19443ab 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -6,10 +6,6 @@ authors = [ "Raph Levien <raph@google.com>" ]
 keywords = ["font", "truetype", "ttf"]
 description = "A font renderer written (mostly) in pure, safe Rust"
 repository = "https://github.com/google/font-rs"
-build = "build.rs"
 
 [features]
 sse = []
-
-[target.'cfg(target_feature = "sse")'.build-dependencies]
-gcc = "0.3"
diff --git a/build.rs b/build.rs
deleted file mode 100644
index 9dbf570..0000000
--- a/build.rs
+++ /dev/null
@@ -1,11 +0,0 @@
-#[cfg(feature="sse")]
-extern crate gcc;
-
-fn main() {
-    #[cfg(feature="sse")]
-    gcc::Build::new()
-        .file("src/accumulate.c")
-        .flag("-march=native")
-        .flag("-std=c99")
-        .compile("libaccumulate.a");
-}
diff --git a/src/accumulate.c b/src/accumulate.c
deleted file mode 100644
index 2551e36..0000000
--- a/src/accumulate.c
+++ /dev/null
@@ -1,37 +0,0 @@
-// Copyright 2015 Google Inc. All rights reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-// SSE3 instrinsics for cumulative sum and conversion to pixels
-
-#include <stdint.h>
-#include <tmmintrin.h>
-
-void accumulate_sse(const float *in, uint8_t *out, uint32_t n) {
-  __m128 offset = _mm_setzero_ps();
-  __m128i mask = _mm_set1_epi32(0x0c080400);
-  __m128 sign_mask = _mm_set1_ps(-0.f);
-  for (uint32_t i = 0; i < n; i += 4) {
-    __m128 x = _mm_load_ps(&in[i]);
-    x = _mm_add_ps(x, _mm_castsi128_ps(_mm_slli_si128(_mm_castps_si128(x), 4)));
-    x = _mm_add_ps(x, _mm_shuffle_ps(_mm_setzero_ps(), x, 0x40));
-    x = _mm_add_ps(x, offset);
-    __m128 y = _mm_andnot_ps(sign_mask, x);  // fabs(x)
-    y = _mm_min_ps(y, _mm_set1_ps(1.0f));
-    y = _mm_mul_ps(y, _mm_set1_ps(255.0f));
-    __m128i z = _mm_cvttps_epi32(y);
-    z = _mm_shuffle_epi8(z, mask);
-    _mm_store_ss((float *)&out[i], _mm_castsi128_ps(z));
-    offset = _mm_shuffle_ps(x, x, _MM_SHUFFLE(3, 3, 3, 3));
-  }
-}
diff --git a/src/accumulate.rs b/src/accumulate.rs
index 8766aa2..d41c536 100644
--- a/src/accumulate.rs
+++ b/src/accumulate.rs
@@ -12,10 +12,18 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#[cfg(feature = "sse")]
-#[link(name = "accumulate")]
-extern "C" {
-    fn accumulate_sse(src: *const f32, dst: *mut u8, n: u32);
+use std::mem;
+
+#[cfg(target_arch = "x86_64")]
+use std::arch::x86_64::*;
+
+#[cfg(target_arch = "x86")]
+use std::arch::x86::*;
+
+macro_rules! _mm_shuffle {
+    ($z:expr, $y:expr, $x:expr, $w:expr) => {
+        ($z << 6) | ($y << 4) | ($x << 2) | $w
+    };
 }
 
 #[cfg(feature = "sse")]
@@ -28,11 +36,33 @@ pub fn accumulate(src: &[f32]) -> Vec<u8> {
     // and so on
     let len = src.len();
     let n = (len + 3) & !3; // align data
-    let mut dst: Vec<u8> = Vec::with_capacity(n);
+    let mut dst: Vec<u8> = vec![0; n];
+
     unsafe {
-        accumulate_sse(src.as_ptr(), dst.as_mut_ptr(), n as u32);
+        let mut offset = _mm_setzero_ps();
+        let sign_mask = _mm_set1_ps(-0.);
+        let mask = _mm_set1_epi32(0x0c080400);
+
+        for i in (0..n).step_by(4) {
+            let mut x = _mm_loadu_ps(&src[i]);
+            x = _mm_add_ps(x, _mm_castsi128_ps(_mm_slli_si128(_mm_castps_si128(x), 4)));
+            x = _mm_add_ps(x, _mm_shuffle_ps(_mm_setzero_ps(), x, 0x40));
+            x = _mm_add_ps(x, offset);
+
+            let mut y = _mm_andnot_ps(sign_mask, x); // fabs(x)
+            y = _mm_min_ps(y, _mm_set1_ps(1.0));
+            y = _mm_mul_ps(y, _mm_set1_ps(255.0));
+
+            let mut z = _mm_cvttps_epi32(y);
+            z = _mm_shuffle_epi8(z, mask);
+
+            _mm_store_ss(mem::transmute(&dst[i]), _mm_castsi128_ps(z));
+            offset = _mm_shuffle_ps(x, x, _mm_shuffle!(3, 3, 3, 3));
+        }
+
         dst.set_len(len); // we must return vec of the same length as src.len()
     }
+
     dst
 }