mirror of
https://github.com/KhronosGroup/SPIRV-Cross.git
synced 2024-11-15 16:30:08 +00:00
acae607703
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. |
||
|---|---|---|
| .. | ||
| 16bit-constants.frag | ||
| array-lut-no-loop-variable.frag | ||
| basic.frag | ||
| complex-expression-in-access-chain.frag | ||
| composite-extract-forced-temporary.frag | ||
| constant-array.frag | ||
| constant-composites.frag | ||
| false-loop-init.frag | ||
| flush_params.frag | ||
| for-loop-init.frag | ||
| frexp-modf.frag | ||
| front-facing.frag | ||
| gather-dref.frag | ||
| ground.frag | ||
| helper-invocation.frag | ||
| hoisted-temporary-use-continue-block-as-value.frag | ||
| image-load-store-uint-coord.asm.frag | ||
| loop-dominator-and-switch-default.frag | ||
| lut-promotion.frag | ||
| mix.frag | ||
| partial-write-preserve.frag | ||
| pls.frag | ||
| sample-parameter.frag | ||
| sampler-ms.frag | ||
| sampler-proj.frag | ||
| sampler.frag | ||
| switch-unsigned-case.frag | ||
| swizzle.frag | ||
| texel-fetch-offset.frag | ||
| ubo_layout.frag | ||
| unary-enclose.frag | ||