Add more inbounds and inlining

[charleskawczynski]

This PR applies more @inbounds and inlining to

• some of the spectral element operators (written for the cpu)
• some of the dss kernels

[charleskawczynski]

bors try

[bors]

try

Build succeeded:

• test-os (windows-latest)
• test-os (macos-latest)
• lib-climacore-vtk
• lib-climacore-makie
• lib-climacore-plots
• lib-climacore-tempestremap
• buildkite/climacore-ci
• format
• docbuild
• test-os (ubuntu-20.04)

[charleskawczynski]

Interestingly, this doesn’t seem to have much of an impact. Any differences could easily be variation.

[charleskawczynski]

The CPU vector dss is the slowest of the spectral kernels, and using IntrospectionTools along with julia --check-bounds=no:

using Revise; using ClimaCore
include(joinpath(pkgdir(ClimaCore), "test", "Operators", "spectralelement", "benchmark_utils.jl"))
include(joinpath(pkgdir(ClimaCore), "test", "Operators", "spectralelement", "benchmark_kernels.jl"))
args = setup_kernel_args(["--device", "CPU"]);
kernel_vector_dss!(args)
using IntrospectionTools
@code_summary kernel_vector_dss!(args)

shows that we're currently generating the same native code as if we used julia --check-bounds=no, so these changes are not needed.

[charleskawczynski]

Actually, the allocation job shows that this fixes some allocation sites. So perhaps we should reconsider.

[charleskawczynski]

Superseded by #1338