gl-block

gl-block

Using

Primitive types should have the Block instance. With that in place, you can build structures and get Storable instances derived generically, according to intented usage.

import GHC.Generics (Generic)
import Graphics.Gl.Block (Block, Packed(..), Std140(..), Std430(..))

-- | Attribute streams can be packed tightly together
data VertexAttrs = VertexAttrs
  { color :: Vec4
  , texCoords :: Vec2
  }
  deriving stock (Eq, Ord, Show, Generic) -- The regular stuff and Generic
  deriving anyclass Block -- The layout class, with Generic defaults
  deriving Storable via (Packed VertexAttrs) -- Free goodies!

-- | Uniform data require jumping through padding and alignment flaming hoop.
-- You can use derive-storable-plugin or hs2c instead, but there are gotchas.
data SceneUniform = SceneUniform
  { projection :: Mat4
  , viewPosition :: Vec3 -- Here comes the jazz
  , viewDirection :: Vec3
  }
  deriving stock (Eq, Ord, Show, Generic)
  deriving anyclass Block
  deriving Storable via (Std140 VertexAttrs) -- With comfy padding

-- | Shader buffer objects are less vacuous, but the rules are specific to the domain.
data Material = Material
  { baseColor :: Vec4
  , metallicRoughness :: Vec2
  , emission :: Vec4
  }
  deriving stock (Eq, Ord, Show, Generic)
  deriving anyclass Block
  deriving Storable via (Std430 VertexAttrs) -- Less alignment, less calculations

• Shader Buffer Memory Layout Info

Benchmarks

The benchmark consists of filling a Storable vector.

• Packed layout is on par with manual instances.
• Std140 is slower, but not catastrophically so.
• Std430 seems to regain some performance due to being a tad simpler.

There’s only one “manual” case standing for all the layouts since it would only be different in pointer offsets. And no way in hell I’m going to calculate them by hand!

  struct
    10
      manual: OK (0.36s)
        19.0 ns ± 1.9 ns, 325 B  allocated,   0 B  copied, 117 MB peak memory
      packed: OK (0.68s)
        18.5 ns ± 782 ps, 327 B  allocated,   0 B  copied, 117 MB peak memory, 0.98x
      std140: OK (0.58s)
        64.9 ns ± 4.0 ns, 827 B  allocated,   0 B  copied, 117 MB peak memory, 3.42x
      std430: OK (0.28s)
        62.1 ns ± 3.3 ns, 658 B  allocated,   0 B  copied, 117 MB peak memory, 3.28x
    1000
      manual: OK (0.41s)
        334  ns ±  30 ns,  23 KB allocated,   0 B  copied, 135 MB peak memory
      packed: OK (0.40s)
        320  ns ±  23 ns,  23 KB allocated,   0 B  copied, 135 MB peak memory, 0.96x
      std140: OK (0.44s)
        733  ns ±  64 ns,  47 KB allocated,   0 B  copied, 137 MB peak memory, 2.19x
      std430: OK (0.55s)
        460  ns ±  30 ns,  32 KB allocated,   0 B  copied, 137 MB peak memory, 1.38x
    1000000
      manual: OK (0.42s)
        344  μs ±  27 μs,  23 MB allocated, 167 B  copied, 141 MB peak memory
      packed: OK (0.42s)
        350  μs ±  14 μs,  23 MB allocated, 164 B  copied, 141 MB peak memory, 1.02x
      std140: OK (0.92s)
        817  μs ±  36 μs,  46 MB allocated, 241 B  copied, 163 MB peak memory, 2.38x
      std430: OK (1.20s)
        523  μs ±  20 μs,  30 MB allocated, 157 B  copied, 164 MB peak memory, 1.52x

Caveat: nested structures have degraded performance.