SPIRV-Cross/reference/shaders-msl/frag/complex-expression-in-access-chain.frag
Hans-Kristian Arntzen acae607703 Register implied expression reads in OpLoad/OpAccessChain.
This is required to avoid relying on complex sub-expression elimination
in compilers, and generates cleaner code.

The problem case is if a complex expression is used in an access chain,
like:

Composite comp = buffer[texture(...)];
vec4 a = comp.a + comp.b + comp.c;

Before, we did not have common subexpression tracking for
OpLoad/OpAccessChain, so we easily ended up with code like:

vec4 a = buffer[texture(...)].a + buffer[texture(...)].b + buffer[texture(...)].c;

A good compiler will optimize this, but we should not rely on it, and
forcing texture(...) to a temporary also looks better.

The solution is to add a vector "implied_expression_reads", which works
similarly to expression_dependencies. We also need an extra mechanism in
to_expression which lets us skip expression read checking and do it
later. E.g. for expr -> access chain -> load, we should only trigger
a read of expr when using the loaded expression.
2019-01-04 14:56:12 +01:00

33 lines
703 B
GLSL

#include <metal_stdlib>
#include <simd/simd.h>
using namespace metal;
struct UBO
{
float4 results[1024];
};
struct main0_out
{
float4 FragColor [[color(0)]];
};
struct main0_in
{
int vIn [[user(locn0)]];
int vIn2 [[user(locn1)]];
};
fragment main0_out main0(main0_in in [[stage_in]], device UBO& _34 [[buffer(0)]], texture2d<int> Buf [[texture(1)]], sampler BufSmplr [[sampler(1)]], float4 gl_FragCoord [[position]])
{
main0_out out = {};
int4 coords = Buf.read(uint2(int2(gl_FragCoord.xy)), 0);
float4 foo = _34.results[coords.x % 16];
int c = in.vIn * in.vIn;
int d = in.vIn2 * in.vIn2;
out.FragColor = (foo + foo) + _34.results[c + d];
return out;
}