define NDEBUG, and simplify casts.

Makefile

@@ -11,7 +11,7 @@ ifdef MAKEDIR:
 CXX = cl
 # C++20 for designated initializers
 # TODO: Detect AVX.
-CXXFLAGS = /MD /Gy /Z7 /EHsc /O2 /arch:AVX2 /std:c++20 /GL
+CXXFLAGS = /MD /Gy /Z7 /EHsc /O2 /arch:AVX2 /std:c++20 /GL /DNDEBUG
 CC = $(CXX)
 CFLAGS = $(CXXFLAGS)
 LDFLAGS=/incremental:no

ggml.c

@@ -408,8 +408,8 @@ void quantize_row_q4_0(const float * restrict x, void * restrict y, int k) {
     const int nb = k / QK;
     const size_t bs = sizeof(float) + QK/2;
 
-    uint8_t * restrict pd = (uint8_t *) ((char*)y + 0*bs);
-    uint8_t * restrict pb = (uint8_t *) ((char*)y + 0*bs + sizeof(float));
+    uint8_t * restrict pd = (uint8_t *)y + 0*bs;
+    uint8_t * restrict pb = (uint8_t *)y + 0*bs + sizeof(float);
 
     uint8_t pp[QK/2];
 
@@ -655,8 +655,8 @@ void dequantize_row_q4_0(const void * restrict x, float * restrict y, int k) {
     const int nb = k / QK;
     const size_t bs = sizeof(float) + QK/2;
 
-    const uint8_t * restrict pd = (const uint8_t *) ((char*)x + 0*bs);
-    const uint8_t * restrict pb = (const uint8_t *) ((char*)x + 0*bs + sizeof(float));
+    const uint8_t * restrict pd = (const uint8_t *)x + 0*bs;
+    const uint8_t * restrict pb = (const uint8_t *)x + 0*bs + sizeof(float);
 
     // scalar
     for (int i = 0; i < nb; i++) {
@@ -1302,11 +1302,11 @@ inline static void ggml_vec_dot_q4_0(const int n, float * restrict s, const void
 
     const size_t bs = sizeof(float) + QK/2;
 
-    const uint8_t * restrict pd0 = (const uint8_t *) ((char*)x + 0*bs);
-    const uint8_t * restrict pd1 = (const uint8_t *) ((char*)y + 0*bs);
+    const uint8_t * restrict pd0 = (const uint8_t *)x + 0*bs;
+    const uint8_t * restrict pd1 = (const uint8_t *)y + 0*bs;
 
-    const uint8_t * restrict pb0 = (const uint8_t *) ((char*)x + 0*bs + sizeof(float));
-    const uint8_t * restrict pb1 = (const uint8_t *) ((char*)y + 0*bs + sizeof(float));
+    const uint8_t * restrict pb0 = (const uint8_t *)x + 0*bs + sizeof(float);
+    const uint8_t * restrict pb1 = (const uint8_t *)y + 0*bs + sizeof(float);
 
     float sumf = 0.0;
 
@@ -1732,8 +1732,8 @@ inline static void ggml_vec_mad_q4_0(const int n, float * restrict y, void * res
     const int nb = n / QK;
     const size_t bs = sizeof(float) + QK/2;
 
-    const uint8_t * restrict pd = (const uint8_t *) ((char*)x + 0*bs);
-    const uint8_t * restrict pb = (const uint8_t *) ((char*)x + 0*bs + sizeof(float));
+    const uint8_t * restrict pd = (const uint8_t *)x + 0*bs;
+    const uint8_t * restrict pb = (const uint8_t *)x + 0*bs + sizeof(float);
 
 #if __ARM_NEON
 #if QK == 32