diff --git a/examples/brew-volley/main.c b/examples/brew-volley/main.c
old mode 100755
new mode 100644
index 16693116c6..82eee847fb
--- a/examples/brew-volley/main.c
+++ b/examples/brew-volley/main.c
@@ -453,7 +453,7 @@ int main()
     debug_init_isviewer();
     debug_init_usblog();
 
-    display_init(RESOLUTION_640x480, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE_FETCH_ALWAYS);
+    display_init(RESOLUTION_640x480, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_DEDITHER);
 
     controller_init();
     timer_init();
diff --git a/examples/fontdemo/fontdemo.c b/examples/fontdemo/fontdemo.c
index 136265cc40..b7454f6b16 100644
--- a/examples/fontdemo/fontdemo.c
+++ b/examples/fontdemo/fontdemo.c
@@ -85,7 +85,7 @@ int main()
     controller_init();
 
     dfs_init(DFS_DEFAULT_LOCATION);
-    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE);
+    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_RESAMPLE);
     rdpq_init();
     // rdpq_debug_start();
     // rdpq_debug_log(true);
diff --git a/examples/gldemo/gldemo.c b/examples/gldemo/gldemo.c
index 5db49e1695..000297f3a6 100644
--- a/examples/gldemo/gldemo.c
+++ b/examples/gldemo/gldemo.c
@@ -114,6 +114,8 @@ void setup()
     glFogf(GL_FOG_END, 20);
     glFogfv(GL_FOG_COLOR, environment_color);
 
+    glEnable(GL_MULTISAMPLE_ARB);
+
     glGenTextures(4, textures);
 
     #if 0
@@ -204,7 +206,7 @@ int main()
     
     dfs_init(DFS_DEFAULT_LOCATION);
 
-    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE_FETCH_ALWAYS);
+    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_RESAMPLE_ANTIALIAS_DEDITHER);
 
     rdpq_init();
     gl_init();
diff --git a/examples/overlays/actor/overlays_actor.c b/examples/overlays/actor/overlays_actor.c
index 4be19fde83..6b4c144e4f 100644
--- a/examples/overlays/actor/overlays_actor.c
+++ b/examples/overlays/actor/overlays_actor.c
@@ -100,7 +100,7 @@ int main()
     debug_init_isviewer();
     debug_init_usblog();
     //Init rendering
-    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE);
+    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_RESAMPLE);
     rdpq_init();
     rdpq_debug_start();
     scr_width = display_get_width();
diff --git a/examples/overlays/scene/overlays_scene.cpp b/examples/overlays/scene/overlays_scene.cpp
index d414de2a60..071c9289c0 100644
--- a/examples/overlays/scene/overlays_scene.cpp
+++ b/examples/overlays/scene/overlays_scene.cpp
@@ -7,7 +7,7 @@ int main()
     debug_init_isviewer();
     debug_init_usblog();
     //Init rendering
-    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE);
+    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_RESAMPLE);
     rdpq_init();
     rdpq_debug_start();
     //Init miscellaneous system
diff --git a/examples/pixelshader/pixelshader.c b/examples/pixelshader/pixelshader.c
index 6ff91efde7..ffaac0bf7c 100644
--- a/examples/pixelshader/pixelshader.c
+++ b/examples/pixelshader/pixelshader.c
@@ -62,7 +62,7 @@ void rsp_blend_process_line(surface_t *dest, int x0, int y0, int numlines) {
 int main(void) {
     debug_init_isviewer();
     debug_init_usblog();
-    display_init(RESOLUTION_640x480, DEPTH_16_BPP, 2, GAMMA_NONE, ANTIALIAS_RESAMPLE);
+    display_init(RESOLUTION_640x480, DEPTH_16_BPP, 2, GAMMA_NONE, FILTERS_RESAMPLE);
     dfs_init(DFS_DEFAULT_LOCATION);
     controller_init();
     rdpq_init();
diff --git a/examples/rdpqdemo/rdpqdemo.c b/examples/rdpqdemo/rdpqdemo.c
index f89aa21db2..8d83382f45 100644
--- a/examples/rdpqdemo/rdpqdemo.c
+++ b/examples/rdpqdemo/rdpqdemo.c
@@ -21,7 +21,7 @@ static object_t objects[NUM_OBJECTS];
 
 // Fair and fast random generation (using xorshift32, with explicit seed)
 static uint32_t rand_state = 1;
-static uint32_t rand(void) {
+static uint32_t myrand(void) {
 	uint32_t x = rand_state;
 	x ^= x << 13;
 	x ^= x >> 7;
@@ -32,8 +32,8 @@ static uint32_t rand(void) {
 // RANDN(n): generate a random number from 0 to n-1
 #define RANDN(n) ({ \
 	__builtin_constant_p((n)) ? \
-		(rand()%(n)) : \
-		(uint32_t)(((uint64_t)rand() * (n)) >> 32); \
+		(myrand()%(n)) : \
+		(uint32_t)(((uint64_t)myrand() * (n)) >> 32); \
 })
 
 static int32_t obj_max_x;
@@ -101,7 +101,7 @@ int main()
     debug_init_isviewer();
     debug_init_usblog();
 
-    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, ANTIALIAS_RESAMPLE);
+    display_init(RESOLUTION_320x240, DEPTH_16_BPP, 3, GAMMA_NONE, FILTERS_RESAMPLE);
 
     controller_init();
     timer_init();
diff --git a/examples/videoplayer/videoplayer.c b/examples/videoplayer/videoplayer.c
index e5e69614a1..2268bd63a8 100644
--- a/examples/videoplayer/videoplayer.c
+++ b/examples/videoplayer/videoplayer.c
@@ -23,7 +23,7 @@ int main(void) {
 	debug_init_isviewer();
 	debug_init_usblog();
 
-	display_init(RESOLUTION_320x240, DEPTH_32_BPP, NUM_DISPLAY, GAMMA_NONE, ANTIALIAS_OFF);
+	display_init(RESOLUTION_320x240, DEPTH_32_BPP, NUM_DISPLAY, GAMMA_NONE, FILTERS_DISABLED);
 	dfs_init(DFS_DEFAULT_LOCATION);
 	rdpq_init();
 
diff --git a/include/display.h b/include/display.h
index 19bb3e04f2..ce116f54b0 100644
--- a/include/display.h
+++ b/include/display.h
@@ -272,6 +272,13 @@ uint32_t display_get_bitdepth(void);
  */
 uint32_t display_get_num_buffers(void);
 
+/**
+ * @brief Get the current number of frames per second being rendered
+ * 
+ * @return float Frames per second
+ */
+float display_get_fps(void);
+
 
 /** @cond */
 __attribute__((deprecated("use display_get or display_try_get instead")))
diff --git a/include/fmath.h b/include/fmath.h
index 8c061cf14d..968c238b64 100644
--- a/include/fmath.h
+++ b/include/fmath.h
@@ -170,11 +170,23 @@ float fm_sinf_approx(float x, int approx);
 /**
  * @brief Faster version of cosf.
  * 
- * @see dragon_sinf for considerations on why and how to use this functions
+ * @see #fm_sinf for considerations on why and how to use this functions
  *      instead of the standard sinf.
  */
 float fm_cosf(float x);
 
+/**
+ * @brief Faster version of sincosf.
+ * 
+ * Similar to #fm_sinf and #fm_cosf, but calculates both sinf and cosf
+ * for the same angle.
+ * 
+ * @param x    The angle in radians
+ * @param sin  The pointer to the variable that will contain the sine
+ * @param cos  The pointer to the variable that will contain the cosine
+ */
+void fm_sincosf(float x, float *sin, float *cos);
+
 /**
  * @brief Faster version of atan2f.
  * 
@@ -199,10 +211,11 @@ float fm_atan2f(float y, float x);
     // we expect that to be resolved at compile-time with the maximum accuracy,
     // so it's important to generate code that calls the standard C function
     // which is then intrinsified by the compiler.
-    #define fmodf(x, y)   ((__builtin_constant_p(x) && __builtin_constant_p(y)) ? fmodf(x,y) : fm_fmodf(x,y))
-    #define sinf(x)       (__builtin_constant_p(x) ? sinf(x) : fm_sinf(x))
-    #define cosf(x)       (__builtin_constant_p(x) ? cosf(x) : fm_cosf(x))
-    #define atan2f(y, x)  ((__builtin_constant_p(x) && __builtin_constant_p(y)) ? atan2f(y, x) : fm_atan2f(y, x))
+    #define fmodf(x, y)     ((__builtin_constant_p(x) && __builtin_constant_p(y)) ? fmodf(x,y) : fm_fmodf(x,y))
+    #define sinf(x)         (__builtin_constant_p(x) ? sinf(x) : fm_sinf(x))
+    #define cosf(x)         (__builtin_constant_p(x) ? cosf(x) : fm_cosf(x))
+    #define sincosf(x,s,c)  (__builtin_constant_p(x) ? sincosf(x,s,c) : fm_sincosf(x,s,c))
+    #define atan2f(y, x)    ((__builtin_constant_p(x) && __builtin_constant_p(y)) ? atan2f(y, x) : fm_atan2f(y, x))
 #endif
 
 #ifdef __cplusplus
diff --git a/include/rsp.h b/include/rsp.h
index 01e1fd44cc..80fa19cfb3 100644
--- a/include/rsp.h
+++ b/include/rsp.h
@@ -157,6 +157,7 @@
 
 #include <stdbool.h>
 #include <stdint.h>
+#include <stdlib.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -468,9 +469,13 @@ void rsp_read_data(void* data, unsigned long size, unsigned int dmem_offset);
  * 
  * @see #rsp_crashf
  */
+#ifndef NDEBUG
 #define rsp_crash()  ({ \
     __rsp_crash(__FILE__, __LINE__, __func__, NULL); \
-}) 
+})
+#else
+#define rsp_crash() abort()
+#endif
 
 /**
  * @brief Abort the program showing a RSP crash screen with a symptom message.
@@ -487,9 +492,13 @@ void rsp_read_data(void* data, unsigned long size, unsigned int dmem_offset);
  * 
  * @see #rsp_crash
  */
+#ifndef NDEBUG
 #define rsp_crashf(msg, ...) ({ \
     __rsp_crash(__FILE__, __LINE__, __func__, msg, ##__VA_ARGS__); \
 })
+#else
+#define rsp_crashf(msg, ...) abort()
+#endif
 
 /**
  * @brief Create a loop that waits for some condition that is related to RSP,
@@ -549,9 +558,13 @@ void run_ucode(void) {
 // Internal function used by rsp_crash and rsp_crashf. These are not part
 // of the public API of rsp.h. Do not call them directly.
 /// @cond
+#ifndef NDEBUG
 void __rsp_crash(const char *file, int line, const char *func, const char *msg, ...)
    __attribute__((noreturn, format(printf, 4, 5)));
 void __rsp_check_assert(const char *file, int line, const char *func);
+#else
+static inline void __rsp_check_assert(const char *file, int line, const char *func) {}
+#endif
 /// @endcond
 
 #ifdef __cplusplus
diff --git a/src/GL/gl_internal.h b/src/GL/gl_internal.h
index 349fe89765..8dd2ea7533 100644
--- a/src/GL/gl_internal.h
+++ b/src/GL/gl_internal.h
@@ -808,6 +808,7 @@ inline void* gl_get_attrib_pointer(gl_obj_attributes_t *attribs, gl_array_type_t
         return attribs->mtx_index;
     default:
         assert(0);
+        return NULL;
     }
 }
 
diff --git a/src/debug.c b/src/debug.c
index 539bf32b70..9f0475cd02 100644
--- a/src/debug.c
+++ b/src/debug.c
@@ -2,7 +2,7 @@
  * @file debug.c
  * @brief Debugging Support
  */
-
+#ifndef NDEBUG
 #include <string.h>
 #include <fcntl.h>
 #include <assert.h>
@@ -651,3 +651,12 @@ void debug_backtrace(void)
 {
 	__debug_backtrace(stderr, false);
 }
+#else
+
+#include <stdlib.h>
+
+void debug_assert_func(...) {
+	abort();
+}
+
+#endif
diff --git a/src/display.c b/src/display.c
index b2fec553fb..a1fa64ef49 100644
--- a/src/display.c
+++ b/src/display.c
@@ -19,7 +19,8 @@
 
 /** @brief Maximum number of video backbuffers */
 #define NUM_BUFFERS         32
-
+/** @brief Number of past frames used to evaluate FPS */
+#define FPS_WINDOW          32
 
 static surface_t *surfaces;
 /** @brief Currently active bit depth */
@@ -40,6 +41,12 @@ static int now_showing = -1;
 static uint32_t drawing_mask = 0;
 /** @brief Bitmask of surfaces that are ready to be shown */
 static volatile uint32_t ready_mask = 0;
+/** @brief Window of absolute times at which previous frames were shown */
+static uint32_t frame_times[FPS_WINDOW];
+/** @brief Current index into the frame times window */
+static int frame_times_index = 0;
+/** @brief Current duration of the frame window (time elapsed for FPS_WINDOW frames) */
+static uint32_t frame_times_duration;
 
 /** @brief Get the next buffer index (with wraparound) */
 static inline int buffer_next(int idx) {
@@ -86,15 +93,12 @@ void display_init( resolution_t res, bitdepth_t bit, uint32_t num_buffers, gamma
     /* Minimum is two buffers. */
     __buffers = MAX(1, MIN(NUM_BUFFERS, num_buffers));
 
-
-    if( res.interlaced != INTERLACE_OFF )
-    {
-        /* Serrate on to stop vertical jitter */
-        control |= VI_CTRL_SERRATE;
-    }
+    bool serrate = res.interlaced != INTERLACE_OFF;
+    /* Serrate on to stop vertical jitter */
+    if(serrate) control |= VI_CTRL_SERRATE;
 
     /* Copy over extra initializations */
-    vi_write_config(&vi_config_presets[res.interlaced][tv_type]);
+    vi_write_config(&vi_config_presets[serrate][tv_type]);
 
     /* Figure out control register based on input given */
     switch( bit )
@@ -349,6 +353,16 @@ void display_show( surface_t* surf )
     drawing_mask &= ~(1 << i);
     ready_mask |= 1 << i;
 
+    /* Record the time at which this frame was (asked to be) shown */
+    uint32_t old_ticks = frame_times[frame_times_index];
+    uint32_t now = get_ticks();
+    if (old_ticks)
+        frame_times_duration = TICKS_DISTANCE(old_ticks, now);
+    frame_times[frame_times_index] = now;
+    frame_times_index++;
+    if (frame_times_index == FPS_WINDOW)
+        frame_times_index = 0;
+
     enable_interrupts();
 }
 
@@ -373,22 +387,28 @@ void display_show_force( display_context_t disp )
     enable_interrupts();
 }
 
-uint32_t display_get_width()
+uint32_t display_get_width(void)
 {
     return __width;
 }
 
-uint32_t display_get_height()
+uint32_t display_get_height(void)
 {
     return __height;
 }
 
-uint32_t display_get_bitdepth()
+uint32_t display_get_bitdepth(void)
 {
     return __bitdepth;
 }
 
-uint32_t display_get_num_buffers()
+uint32_t display_get_num_buffers(void)
 {
     return __buffers;
 }
+
+float display_get_fps(void)
+{
+    if (!frame_times_duration) return 0;
+    return (float)(FPS_WINDOW * TICKS_PER_SECOND) / frame_times_duration;
+}
diff --git a/src/exception.c b/src/exception.c
index 12591f1302..3d7f6c7260 100644
--- a/src/exception.c
+++ b/src/exception.c
@@ -202,6 +202,7 @@ void __exception_dump_fpr(exception_t* ex, void (*cb)(void *arg, const char *reg
 	}
 }
 
+#ifndef NDEBUG
 static void debug_exception(exception_t* ex) {
 	debugf("\n\n******* CPU EXCEPTION *******\n");
 	__exception_dump_header(stderr, ex);
@@ -225,6 +226,7 @@ static void debug_exception(exception_t* ex) {
 		debugf("\n");
 	}
 }
+#endif
 
 /**
  * @brief Default exception handler.
@@ -234,6 +236,7 @@ static void debug_exception(exception_t* ex) {
  * of all GPR/FPR registers. It then calls abort() to abort execution.
  */
 void exception_default_handler(exception_t* ex) {
+	#ifndef NDEBUG
 	static bool backtrace_exception = false;
 
 	// Write immediately as much data as we can to the debug spew. This is the
@@ -250,6 +253,7 @@ void exception_default_handler(exception_t* ex) {
 
 	// Run the inspector
 	__inspector_exception(ex);
+	#endif
 
 	abort();
 }
diff --git a/src/fmath.c b/src/fmath.c
index ba515cab2b..054287f6fa 100644
--- a/src/fmath.c
+++ b/src/fmath.c
@@ -49,6 +49,12 @@ float fm_cosf(float x) {
     return fm_sinf(half_pi_hi - x);
 }
 
+void fm_sincosf(float x, float *sin, float *cos) {
+    float y = fm_sinf_approx(x, 0);
+    *sin = y;
+    *cos = sqrtf(1.0f - y * y);
+}
+
 float fm_atan2f(float y, float x) {
     // Approximation of atan2f using a polynomial minmax approximation in [0,1]
     // calculated via the Remez algorithm (https://math.stackexchange.com/a/1105038).
diff --git a/src/inspector.c b/src/inspector.c
index d5bc2948bd..069af8dd61 100644
--- a/src/inspector.c
+++ b/src/inspector.c
@@ -1,3 +1,4 @@
+#ifndef NDEBUG
 #include "graphics.h"
 #include "display.h"
 #include "debug.h"
@@ -583,3 +584,4 @@ void __inspector_init(void) {
     }
     register_syscall_handler(handler, 0x00001, 0x00002);
 }
+#endif /* NDBUEG */
diff --git a/src/rsp.c b/src/rsp.c
index df1a63066e..98c423343f 100644
--- a/src/rsp.c
+++ b/src/rsp.c
@@ -41,7 +41,9 @@ static void __SP_DMA_wait(void)
 
 static void rsp_interrupt(void)
 {
+    #ifndef NDEBUG
     __rsp_check_assert(__FILE__, __LINE__, __func__);
+    #endif
 }
 
 void rsp_init(void)
@@ -170,6 +172,7 @@ void rsp_run(void)
     rsp_wait();
 }
 
+#ifndef NDEBUG
 /// @cond
 // Check if the RSP has hit an internal assert, and call rsp_crash if so.
 // This function is invoked by #RSP_WAIT_LOOP while waiting for the RSP
@@ -196,7 +199,9 @@ void __rsp_check_assert(const char *file, int line, const char *func)
     }
 }
 /// @endcond
+#endif /* NDEBUG */
 
+#ifndef NDEBUG
 /// @cond
 // RSP crash handler implementation
 __attribute__((noreturn, format(printf, 4, 5)))
@@ -447,5 +452,6 @@ void __rsp_crash(const char *file, int line, const char *func, const char *msg,
     abort();
 }
 /// @endcond
+#endif /* NDEBUG */
 
 extern inline void rsp_run_async(void);
diff --git a/src/rspq/rspq.c b/src/rspq/rspq.c
index 4bd74077f6..c3ceb11cb4 100644
--- a/src/rspq/rspq.c
+++ b/src/rspq/rspq.c
@@ -754,6 +754,7 @@ static uint32_t rspq_overlay_get_command_count(rspq_overlay_header_t *header)
     }
 
     assertf(0, "Overlays can only define up to %d commands!", RSPQ_MAX_OVERLAY_COMMAND_COUNT);
+    return 0;
 }
 
 static uint32_t rspq_find_new_overlay_index(void)
diff --git a/tests/testrom.c b/tests/testrom.c
index a2788ebeac..c72aca38dd 100644
--- a/tests/testrom.c
+++ b/tests/testrom.c
@@ -58,7 +58,7 @@ typedef void (*TestFunc)(TestContext *ctx);
 
 // Fair and fast random generation (using xorshift32, with explicit seed)
 static uint32_t rand_state = 1;
-static uint32_t rand(void) {
+static uint32_t myrand(void) {
 	uint32_t x = rand_state;
 	x ^= x << 13;
 	x ^= x >> 7;
@@ -72,8 +72,8 @@ static uint32_t rand(void) {
 // RANDN(n): generate a random number from 0 to n-1
 #define RANDN(n) ({ \
 	__builtin_constant_p((n)) ? \
-		(rand()%(n)) : \
-		(uint32_t)(((uint64_t)rand() * (n)) >> 32); \
+		(myrand()%(n)) : \
+		(uint32_t)(((uint64_t)myrand() * (n)) >> 32); \
 })
 
 // ASSERT(cond, msg): fail the test if the condition is false (with log message)