Fix shaderf16 support on vulkan/nvidia

[cryvosh]

Connections
N/A

Description
On my Laptop 4090, vulkaninfo reports:

VkPhysicalDeviceVulkan11Features:
---------------------------------
storageBuffer16BitAccess = true
uniformAndStorageBuffer16BitAccess = true
storagePushConstant16 = true
storageInputOutput16 = false

However the wgpu vulkan backend requires all four to be true for the SHADER_F16 feature to be supported.
This doesn't seem necessary, and as far as I can tell, storage_input_output16 isn't even read anywhere in wgpu.

This is also how dawn handles f16 support checking here.

Testing
My personal project using f16s (developed in wasm which doesn't have this issue) was tested and works fine on Windows 11, Vulkan Backend.

Squash or Rebase?
Squash

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy --tests. If applicable, add:
  • --target wasm32-unknown-unknown
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[sagudev]

storage_input_output16 is somehow implicitly used here:

wgpu/naga/src/back/spv/writer.rs

Line 1207 in 50eb207

.insert(spirv::Capability::StorageInputOutput16);

As for Dawn https://issues.chromium.org/issues/42251212:

Tint can now polyfill f16 shader IO by using f32 types, so we can
enable the F16 feature on Vulkan without this capability.

I think naga also does pollyfill somewhere.

[cryvosh]

Any idea where would be the best place to add the polyfill?