Update to v075r06 release.

byuu says:

Removed the floating-point volume adjustments from the wave channel and
the left/right speaker mixers. Also, against my better judgment I'm
backing out of left/right computation when they are both turned off.
This basically makes non-stereo games run faster, but will make stereo
games appear to run slower. I don't like it when end-users experience
mystery slowdowns.

Anyway, it appears that the audio calculation is really fucking
demanding. Knocks FPS from 800 down to 300. I thought it might be libco,
so I took it out and it only went up to 305fps o.O

There is also some sort of problem with bsnes/Super Game Boy audio. The
latency is really great when you first start, but it seems to drift
apart over time until it is well over 500ms, and then it either pops or
fades back to very low, sub-50ms latency again. The way I handle mixing
is that the coprocessor audio samples go into a resampler to the native
SNES rate, and fed to an output buffer. SNES audio samples go there
untouched. When there is a sample in each, I add them together and
average the result (I still don't understand why we divide by two since
these are signed integers), and output it immediately. It's just-in-time
sampling, so as long as DSP v Coprocessor do not drift very far, it
should have very low latency. And I make the CPU sync DSP and
Coprocessor once per scanline, which is something like 15 samples or so.

bsnes/Makefile

@@ -2,7 +2,7 @@ include nall/Makefile
 snes := snes
 gameboy := gameboy
 profile := compatibility
-ui := ui-gameboy
+ui := ui
 
 # compiler
 c       := $(compiler) -std=gnu99

bsnes/gameboy/apu/apu.cpp

@@ -21,23 +21,24 @@ void APU::main() {
       scheduler.exit(Scheduler::ExitReason::SynchronizeEvent);
     }
 
-    if((counter & 8191) == 0) {  //512hz
-      if(sequencer == 0 || sequencer == 2 || sequencer == 4 || sequencer == 6) {  //256hz
+    if(sequencer_base == 0) {  //512hz
+      if(sequencer_step == 0 || sequencer_step == 2 || sequencer_step == 4 || sequencer_step == 6) {  //256hz
         square1.clock_length();
         square2.clock_length();
         wave.clock_length();
         noise.clock_length();
       }
-      if(sequencer == 2 || sequencer == 6) {  //128hz
+      if(sequencer_step == 2 || sequencer_step == 6) {  //128hz
         square1.clock_sweep();
       }
-      if(sequencer == 7) {  //64hz
+      if(sequencer_step == 7) {  //64hz
         square1.clock_envelope();
         square2.clock_envelope();
         noise.clock_envelope();
       }
-      sequencer = (sequencer + 1) & 7;
+      sequencer_step++;
     }
+    sequencer_base++;
 
     square1.run();
     square2.run();
@@ -46,9 +47,6 @@ void APU::main() {
     master.run();
 
     system.interface->audio_sample(master.center, master.left, master.right);
-
-    if(++counter == 4194304) counter = 0;
-
     if(++clock >= 0) co_switch(scheduler.active_thread = cpu.thread);
   }
 }
@@ -58,8 +56,8 @@ void APU::power() {
   for(unsigned n = 0xff10; n <= 0xff3f; n++) bus.mmio[n] = this;
 
   foreach(n, mmio_data) n = 0x00;
-  counter = 0;
-  sequencer = 0;
+  sequencer_base = 0;
+  sequencer_step = 0;
 
   square1.power();
   square2.power();

bsnes/gameboy/apu/apu.hpp

@@ -6,8 +6,8 @@ struct APU : Processor, MMIO {
   #include "master/master.hpp"
 
   uint8 mmio_data[48];
-  unsigned counter;
-  unsigned sequencer;
+  uint13 sequencer_base;
+  uint3 sequencer_step;
 
   Square1 square1;
   Square2 square2;

bsnes/gameboy/apu/master/master.cpp

@@ -1,15 +1,26 @@
 #ifdef APU_CPP
 
 void APU::Master::run() {
-  signed sample = 0, channels = 4;
+  if(enable == false) {
+    center = 0;
+    left   = 0;
+    right  = 0;
+    return;
+  }
+
+  signed sample = 0, channels;
   sample += apu.square1.output;
   sample += apu.square2.output;
   sample +=    apu.wave.output;
   sample +=   apu.noise.output;
-  sample /= channels;
-  if(sample < -32768) sample = -32768;
-  if(sample > +32767) sample = +32767;
-  center = sample;
+  sample >>= 2;
+  center = sclamp<16>(sample);
+
+  if(left_enable == false && right_enable == false) {
+    left  = center;
+    right = center;
+    return;
+  }
 
   sample = 0;
   channels = 0;
@@ -18,17 +29,16 @@ void APU::Master::run() {
   if(channel3_left_enable) { sample +=    apu.wave.output; channels++; }
   if(channel4_left_enable) { sample +=   apu.noise.output; channels++; }
   if(channels) sample /= channels;
-  left = sample;
+  left = sclamp<16>(sample);
 
   switch(left_volume) {
-    case 0: left *= 0.125; break;
-    case 1: left *= 0.250; break;
-    case 2: left *= 0.375; break;
-    case 3: left *= 0.500; break;
-    case 4: left *= 0.625; break;
-    case 5: left *= 0.750; break;
-    case 6: left *= 0.875; break;
-    case 7: left *= 1.000; break;
+    case 0: left >>= 3;                       break;  // 12.5%
+    case 1: left >>= 2;                       break;  // 25.0%
+    case 2: left = (left >> 2) + (left >> 3); break;  // 37.5%
+    case 3: left >>= 1;                       break;  // 50.0%
+    case 4: left = (left >> 1) + (left >> 3); break;  // 62.5%
+    case 5: left -= (left >> 2);              break;  // 75.0%
+    case 6: left -= (left >> 3);              break;  // 87.5%
   }
   if(left_enable == false) left = 0;
 
@@ -39,30 +49,18 @@ void APU::Master::run() {
   if(channel3_right_enable) { sample +=    apu.wave.output; channels++; }
   if(channel4_right_enable) { sample +=   apu.noise.output; channels++; }
   if(channels) sample /= channels;
-  right = sample;
+  right = sclamp<16>(sample);
 
   switch(right_volume) {
-    case 0: right *= 0.125; break;
-    case 1: right *= 0.250; break;
-    case 2: right *= 0.375; break;
-    case 3: right *= 0.500; break;
-    case 4: right *= 0.625; break;
-    case 5: right *= 0.750; break;
-    case 6: right *= 0.875; break;
-    case 7: right *= 1.000; break;
+    case 0: right >>= 3;                         break;  // 12.5%
+    case 1: right >>= 2;                         break;  // 25.0%
+    case 2: right = (right >> 2) + (right >> 3); break;  // 37.5%
+    case 3: right >>= 1;                         break;  // 50.0%
+    case 4: right = (right >> 1) + (right >> 3); break;  // 62.5%
+    case 5: right -= (right >> 2);               break;  // 75.0%
+    case 6: right -= (right >> 3);               break;  // 87.5%
   }
   if(right_enable == false) right = 0;
-
-  if(left_enable == false && right_enable == false) {
-    left  = center;
-    right = center;
-  }
-
-  if(enable == false) {
-    center = 0;
-    left   = 0;
-    right  = 0;
-  }
 }
 
 void APU::Master::write(unsigned r, uint8 data) {

bsnes/gameboy/apu/serialization.cpp

@@ -2,8 +2,8 @@
 
 void APU::serialize(serializer &s) {
   s.array(mmio_data);
-  s.integer(counter);
-  s.integer(sequencer);
+  s.integer(sequencer_base);
+  s.integer(sequencer_step);
 
   square1.serialize(s);
   square2.serialize(s);

bsnes/gameboy/apu/wave/wave.cpp

@@ -12,12 +12,7 @@ void APU::Wave::run() {
   if(enable == false) sample = 0;
 
   output = (sample * 4369) - 32768;
-  switch(volume) {
-    case 0: output *= 0.00; break;
-    case 1: output *= 1.00; break;
-    case 2: output *= 0.50; break;
-    case 3: output *= 0.25; break;
-  }
+  output >>= volume;
 }
 
 void APU::Wave::clock_length() {
@@ -38,7 +33,12 @@ void APU::Wave::write(unsigned r, uint8 data) {
   }
 
   if(r == 2) {
-    volume = (data >> 5) & 3;
+    switch((data >> 5) & 3) {
+      case 0: volume = 16; break;  //  0%
+      case 1: volume =  0; break;  //100%
+      case 2: volume =  1; break;  // 50%
+      case 3: volume =  2; break;  // 25%
+    }
   }
 
   if(r == 3) {

bsnes/gameboy/cpu/cpu.cpp

@@ -94,7 +94,7 @@ void CPU::interrupt_exec(uint16 pc) {
 }
 
 void CPU::power() {
-  create(Main, 4 * 1024 * 1024);
+  create(Main, 4194304);
 
   for(unsigned n = 0xc000; n <= 0xdfff; n++) bus.mmio[n] = this;  //WRAM
   for(unsigned n = 0xe000; n <= 0xfdff; n++) bus.mmio[n] = this;  //WRAM (mirror)

bsnes/gameboy/cpu/timing/timing.cpp

@@ -1,13 +1,9 @@
 //4194304hz (4 * 1024 * 1024)
+
 //70224 clocks/frame
 //  456 clocks/scanline
 //  154 scanlines/frame
 
-//4194304 /   4096 = 1024
-//4194304 / 262144 =   16
-//4194304 /  65536 =   64
-//4394304 /  16384 =  256
-
 #ifdef CPU_CPP
 
 #include "opcode.cpp"
@@ -17,8 +13,8 @@ void CPU::add_clocks(unsigned clocks) {
   scheduler.exit(Scheduler::ExitReason::StepEvent);
 
   status.clock += clocks;
-  if(status.clock >= 4 * 1024 * 1024) {
-    status.clock -= 4 * 1024 * 1024;
+  if(status.clock >= 4194304) {
+    status.clock -= 4194304;
     cartridge.mbc3.second();
   }
 

bsnes/gameboy/gameboy.hpp

@@ -5,7 +5,7 @@
 namespace GameBoy {
   namespace Info {
     static const char Name[] = "bgameboy";
-    static const char Version[] = "000.16";
+    static const char Version[] = "000.17";
     static unsigned SerializerVersion = 1;
   }
 }
@@ -33,12 +33,38 @@ namespace GameBoy {
   typedef uint32_t uint32;
   typedef uint64_t uint64;
 
-  typedef int_t< 4> int4;
-  typedef int_t<15> int15;
-
+  typedef uint_t< 1> uint1;
+  typedef uint_t< 2> uint2;
+  typedef uint_t< 3> uint3;
   typedef uint_t< 4> uint4;
+  typedef uint_t< 5> uint5;
+  typedef uint_t< 6> uint6;
+  typedef uint_t< 7> uint7;
+
+  typedef uint_t< 9> uint9;
+  typedef uint_t<10> uint10;
+  typedef uint_t<11> uint11;
+  typedef uint_t<12> uint12;
+  typedef uint_t<13> uint13;
+  typedef uint_t<14> uint14;
   typedef uint_t<15> uint15;
 
+  typedef uint_t<17> uint17;
+  typedef uint_t<18> uint18;
+  typedef uint_t<19> uint19;
+  typedef uint_t<20> uint20;
+  typedef uint_t<21> uint21;
+  typedef uint_t<22> uint22;
+  typedef uint_t<23> uint23;
+  typedef uint_t<24> uint24;
+  typedef uint_t<25> uint25;
+  typedef uint_t<26> uint26;
+  typedef uint_t<27> uint27;
+  typedef uint_t<28> uint28;
+  typedef uint_t<29> uint29;
+  typedef uint_t<30> uint30;
+  typedef uint_t<31> uint31;
+
   template<uint16 lo, uint16 hi>
   alwaysinline bool within(uint16 addr) {
     static const uint16 mask = ~(hi ^ lo);

bsnes/gameboy/lcd/lcd.cpp

@@ -190,7 +190,7 @@ void LCD::render_obj() {
 }
 
 void LCD::power() {
-  create(Main, 4 * 1024 * 1024);
+  create(Main, 4194304);
 
   for(unsigned n = 0x8000; n <= 0x9fff; n++) bus.mmio[n] = this;  //VRAM
   for(unsigned n = 0xff40; n <= 0xff4b; n++) bus.mmio[n] = this;  //MMIO

bsnes/snes/audio/audio.cpp

@@ -24,8 +24,8 @@ void Audio::sample(int16 left, int16 right) {
     system.interface->audio_sample(left, right);
   } else {
     dsp_buffer[dsp_wroffset] = ((uint16)left << 0) + ((uint16)right << 16);
-    dsp_wroffset = (dsp_wroffset + 1) & 32767;
-    dsp_length = (dsp_length + 1) & 32767;
+    dsp_wroffset = (dsp_wroffset + 1) & buffer_mask;
+    dsp_length = (dsp_length + 1) & buffer_mask;
     flush();
   }
 }
@@ -50,8 +50,8 @@ void Audio::coprocessor_sample(int16 left, int16 right) {
   r_frac = r_step - first;
 
   cop_buffer[cop_wroffset] = (output_left << 0) + (output_right << 16);
-  cop_wroffset = (cop_wroffset + 1) & 32767;
-  cop_length = (cop_length + 1) & 32767;
+  cop_wroffset = (cop_wroffset + 1) & buffer_mask;
+  cop_length = (cop_length + 1) & buffer_mask;
   flush();
 }
 
@@ -63,8 +63,8 @@ void Audio::flush() {
     uint32 dsp_sample = dsp_buffer[dsp_rdoffset];
     uint32 cop_sample = cop_buffer[cop_rdoffset];
 
-    dsp_rdoffset = (dsp_rdoffset + 1) & 32767;
-    cop_rdoffset = (cop_rdoffset + 1) & 32767;
+    dsp_rdoffset = (dsp_rdoffset + 1) & buffer_mask;
+    cop_rdoffset = (cop_rdoffset + 1) & buffer_mask;
 
     dsp_length--;
     cop_length--;

bsnes/snes/audio/audio.hpp

@@ -8,7 +8,8 @@ public:
 
 private:
   bool coprocessor;
-  uint32 dsp_buffer[32768], cop_buffer[32768];
+  enum : unsigned { buffer_size = 32768, buffer_mask = buffer_size - 1 };
+  uint32 dsp_buffer[buffer_size], cop_buffer[buffer_size];
   unsigned dsp_rdoffset, cop_rdoffset;
   unsigned dsp_wroffset, cop_wroffset;
   unsigned dsp_length, cop_length;

bsnes/snes/snes.hpp

@@ -1,7 +1,7 @@
 namespace SNES {
   namespace Info {
     static const char Name[] = "bsnes";
-    static const char Version[] = "075.05";
+    static const char Version[] = "075.06";
     static const unsigned SerializerVersion = 18;
   }
 }

bsnes/ui-gameboy/main.cpp

@@ -38,7 +38,7 @@ void Application::main(int argc, char **argv) {
   audio.driver("ALSA");
   #endif
   audio.set(Audio::Handle, (uintptr_t)mainWindow.viewport.handle());
-  audio.set(Audio::Synchronize, true);
+  audio.set(Audio::Synchronize, false);
   audio.set(Audio::Volume, 100U);
   audio.set(Audio::Latency, 80U);
   audio.set(Audio::Frequency, 44100U);