Matrix4f Mapped optimized

public void mul4f(Matrix4fMapped left, Matrix4fMapped right) {
    float rmX3 = right.m03;
    float rmX2 = right.m02;
    float rmX1 = right.m01;
    float rmX0 = right.m00;

    final float lm03 = left.m03;
    this.m03 = lm03 * rmX0 + left.m13 * rmX1 + left.m23 * rmX2 + left.m33 * rmX3;
    final float lm02 = left.m02;
    this.m02 = lm02 * rmX0 + left.m12 * rmX1 + left.m22 * rmX2 + left.m32 * rmX3;
    final float lm01 = left.m01;
    this.m01 = lm01 * rmX0 + left.m11 * rmX1 + left.m21 * rmX2 + left.m31 * rmX3;
    final float lm00 = left.m00;
    this.m00 = lm00 * rmX0 + left.m10 * rmX1 + left.m20 * rmX2 + left.m30 * rmX3;

    rmX3 = right.m13;
    rmX2 = right.m12;
    rmX1 = right.m11;
    rmX0 = right.m10;

    final float lm13 = left.m13;
    final float lm12 = left.m12;
    final float lm11 = left.m11;
    final float lm10 = left.m10;

    this.m13 = lm03 * rmX0 + lm13 * rmX1 + left.m23 * rmX2 + left.m33 * rmX3;
    this.m12 = lm02 * rmX0 + lm12 * rmX1 + left.m22 * rmX2 + left.m32 * rmX3;
    this.m11 = lm01 * rmX0 + lm11 * rmX1 + left.m21 * rmX2 + left.m31 * rmX3;
    this.m10 = lm00 * rmX0 + lm10 * rmX1 + left.m20 * rmX2 + left.m30 * rmX3;

    rmX3 = right.m23;
    rmX2 = right.m22;
    rmX1 = right.m21;
    rmX0 = right.m20;

    final float lm23 = left.m23;
    final float lm22 = left.m22;
    final float lm21 = left.m21;
    final float lm20 = left.m20;

    this.m23 = lm03 * rmX0 + lm13 * rmX1 + lm23 * rmX2 + left.m33 * rmX3;
    this.m22 = lm02 * rmX0 + lm12 * rmX1 + lm22 * rmX2 + left.m32 * rmX3;
    this.m21 = lm01 * rmX0 + lm11 * rmX1 + lm21 * rmX2 + left.m31 * rmX3;
    this.m20 = lm00 * rmX0 + lm10 * rmX1 + lm20 * rmX2 + left.m30 * rmX3;

    rmX3 = right.m33;
    rmX2 = right.m32;
    rmX1 = right.m31;
    rmX0 = right.m30;

    this.m33 = lm03 * rmX0 + lm13 * rmX1 + lm23 * rmX2 + left.m33 * rmX3;
    this.m32 = lm02 * rmX0 + lm12 * rmX1 + lm22 * rmX2 + left.m32 * rmX3;
    this.m31 = lm01 * rmX0 + lm11 * rmX1 + lm21 * rmX2 + left.m31 * rmX3;
    this.m30 = lm00 * rmX0 + lm10 * rmX1 + lm20 * rmX2 + left.m30 * rmX3;
}