We currently generate our own I/O signatures inside the DXIL parser, but
use the element counts from the DXBC container signatures to allocate
the input_params/output_params/patch_constant_params arrays. That
happens to work for well-behaved inputs, but it's asking for trouble.
We already check for error instructions when parsing swizzles, but if allocation
fails at codegen time we would like to avoid asserting when subsequently
constructing a swizzle.
Similar to d1c2ae3f0e, this is a bit too strict
and may prevent e.g. simultaneous use of float and float1 at codegen time.
However, in this case the inciting factor is that in the case of allocation
failure at codegen time, we would like to allow one or more arguments to have
error type.
The primary motivation here is to avoid needing to worry about instructions
potentially pointing to the preallocated error instruction in the case of
allocation failure.
This doesn't cover all passes, but none of the other passes make assumptions
about instruction sources.
This is because lower_nonconstant_array_loads() can potentially turn
nonconstant loads into constant loads, allowing copy-prop to turn these
loads into previous instructions, which might help other passes as well.
This patch lowers the number of required temps for the following ps_2_0
shader from 19 to 16:
int i;
float3x3 mats[4];
float4 main() : sv_target
{
return mul(mats[i], float3(1, 2, 3)).xyzz;
}
This can save a significant amount of temp registers because it allows to
avoid referencing the temp (and having to store it) when not needed.
For instance, this patch lowers the number of required temps for the
following ps_2_0 shader from 24 to 19:
int i;
float3x3 mats[4];
float4 main() : sv_target
{
return mul(mats[i], float3(1, 2, 3)).xyzz;
}
Also, it is needed for SM1 vertex shader relative addressing since
non-constant loads are required to be directly on the uniform ('c'
registers) instead of the temp, and non-constant loads cannot be
transformed by copy propagation.
Since 4a94bfc2f6 we segregate
different D3D12 descriptor types in different Vulkan descriptor sets.
This change was introduced to reduce descriptor wasting when
allocating a new descriptor pool; that can be very useful when
using virtual heaps, which have to often cycle through many
descriptors, but it is expected to have limited impact for Vulkan
heaps, given that in that case most descriptors are allocated through
the descriptor heap rather than through the command allocator.
Instead, it has a rather detrimental effect with Vulkan heaps,
because it tends to use many more Vulkan descriptor sets than
necessary, often with just a handful of descriptors each. This
causes a regression on some Vulkan implementations that support
too few descriptor sets.
With this change we revert to a situation similar to before,
stuffing all the descriptors that do not live in a root
descriptor table in as few descriptor sets as possible (at most
one or two, depending on whether push descriptors are used).
