fixing shader execution

example/main.zig

@@ -4,8 +4,8 @@ const spv = @import("spv");
 
 const shader_source = @embedFile("shader.spv");
 
-const screen_width = 200;
-const screen_height = 200;
+const screen_width = 480;
+const screen_height = 240;
 
 pub fn main() !void {
     {
@@ -50,9 +50,7 @@ pub fn main() !void {
         }
 
         var thread_pool: std.Thread.Pool = undefined;
-        try thread_pool.init(.{
-            .allocator = allocator,
-        });
+        try thread_pool.init(.{ .allocator = allocator });
 
         var timer = try std.time.Timer.start();
 
@@ -73,15 +71,15 @@ pub fn main() !void {
 
                 const pixel_map: [*]u32 = @as([*]u32, @ptrCast(@alignCast((surface.getPixels() orelse return).ptr)));
 
+                const delta: f32 = @as(f32, @floatFromInt(timer.read())) / std.time.ns_per_s;
+
                 var frame_timer = try std.time.Timer.start();
                 defer {
                     const ns = frame_timer.lap();
                     const ms = @as(f32, @floatFromInt(ns)) / std.time.ns_per_s;
-                    std.log.info("Took {d:.3}s - {d:.3}fps to render", .{ ms, 1.0 / ms });
+                    std.log.info("Took {d:.3}s - {d:.3}fps to render {d:.2}", .{ ms, 1.0 / ms, delta });
                 }
 
-                const delta: f32 = @as(f32, @floatFromInt(timer.read())) / std.time.ns_per_s;
-
                 var wait_group: std.Thread.WaitGroup = .{};
                 for (0..screen_height) |y| {
                     const runner = &runner_cache.items[y];
@@ -131,10 +129,10 @@ const Runner = struct {
             try rt.readOutput(f32, output[0..], self.color);
 
             const rgba = self.surface.mapRgba(
-                @truncate(@as(u32, @intFromFloat(output[0] * 255.0))),
-                @truncate(@as(u32, @intFromFloat(output[1] * 255.0))),
-                @truncate(@as(u32, @intFromFloat(output[2] * 255.0))),
-                @truncate(@as(u32, @intFromFloat(output[3] * 255.0))),
+                @intCast(@max(@min(@as(i32, @intFromFloat(output[0] * 255.0)), 255), 0)),
+                @intCast(@max(@min(@as(i32, @intFromFloat(output[1] * 255.0)), 255), 0)),
+                @intCast(@max(@min(@as(i32, @intFromFloat(output[2] * 255.0)), 255), 0)),
+                @intCast(@max(@min(@as(i32, @intFromFloat(output[3] * 255.0)), 255), 0)),
             );
 
             pixel_map[(y * self.surface.getWidth()) + x] = rgba.value;

example/mangohud.conf

@@ -0,0 +1,6 @@
+gpu_stats=0
+font_size=16
+resolution
+hud_compact
+background_alpha=0
+width=140

example/shader.nzsl

@@ -16,16 +16,16 @@ struct FragOut
 [entry(frag)]
 fn main(input: FragIn) -> FragOut
 {
-    const I: i32 = 32;
+    const I: i32 = 128;
     const A: f32 = 7.5;
-    const MA: f32 = 20.0;
+    const MA: f32 = 100.0;
     const MI: f32 = 0.001;
 
     let uv0 = input.pos / input.res * 2.0 - vec2[f32](1.0, 1.0);
     let uv  = vec2[f32](uv0.x * (input.res.x / input.res.y), uv0.y);
 
-    let col = vec3[f32](0.0, 0.0, 0.0);
-    let ro  = vec3[f32](0.0, 0.0, -2.0);
+    let col = vec3[f32](0.0, 0.0, 0.0);  
+    let ro  = vec3[f32](0.0, 0.0, -2.0); 
     let rd  = vec3[f32](uv.x, uv.y, 1.0);
     let dt  = 0.0;
     let ds  = 0.0;

example/shader.spv.txt

@@ -55,11 +55,11 @@ Schema: 0
  %29 = OpConstant %3 f32(0.2)
  %30 = OpConstant %3 f32(4)
  %31 = OpTypePointer StorageClass(Function) %6
- %32 = OpConstant %6 i32(32)
+ %32 = OpConstant %6 i32(128)
  %33 = OpTypeBool
  %34 = OpConstant %3 f32(0.001)
  %35 = OpConstant %3 f32(0.35)
- %36 = OpConstant %3 f32(20)
+ %36 = OpConstant %3 f32(100)
  %37 = OpConstant %3 f32(0.15)
  %38 = OpConstant %3 f32(0.05)
  %39 = OpConstant %3 f32(1.15)