Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix inbounds codegen for CPU #431

Merged
merged 1 commit into from
Nov 6, 2023
Merged

Fix inbounds codegen for CPU #431

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 17 additions & 12 deletions src/macros.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,13 +14,6 @@ function __kernel(expr, generate_cpu=true, force_inbounds=false)
def = splitdef(expr)
name = def[:name]
args = def[:args]
if force_inbounds
body_qt = quote
@inbounds $(def[:body])
end
def[:body] = body_qt
end

find_return(expr) && error("Return statement not permitted in a kernel function $name")

constargs = Array{Bool}(undef, length(args))
Expand All @@ -45,13 +38,13 @@ function __kernel(expr, generate_cpu=true, force_inbounds=false)
if generate_cpu
def_cpu = deepcopy(def)
def_cpu[:name] = cpu_name
transform_cpu!(def_cpu, constargs)
transform_cpu!(def_cpu, constargs, force_inbounds)
cpu_function = combinedef(def_cpu)
end

def_gpu = deepcopy(def)
def_gpu[:name] = gpu_name = Symbol(:gpu_, name)
transform_gpu!(def_gpu, constargs)
transform_gpu!(def_gpu, constargs, force_inbounds)
gpu_function = combinedef(def_gpu)

# create constructor functions
Expand Down Expand Up @@ -83,17 +76,23 @@ end

# The easy case, transform the function for GPU execution
# - mark constant arguments by applying `constify`.
function transform_gpu!(def, constargs)
function transform_gpu!(def, constargs, force_inbounds)
let_constargs = Expr[]
for (i, arg) in enumerate(def[:args])
if constargs[i]
push!(let_constargs, :($arg = $constify($arg)))
end
end
pushfirst!(def[:args], :__ctx__)
body = def[:body]
if force_inbounds
body = quote
@inbounds $(body)
end
end
body = quote
if $__validindex(__ctx__)
$(def[:body])
$(body)
end
return nothing
end
Expand All @@ -110,7 +109,7 @@ end
# - handle indicies
# - hoist workgroup definitions
# - hoist uniform variables
function transform_cpu!(def, constargs)
function transform_cpu!(def, constargs, force_inbounds)
let_constargs = Expr[]
for (i, arg) in enumerate(def[:args])
if constargs[i]
Expand All @@ -121,7 +120,13 @@ function transform_cpu!(def, constargs)
new_stmts = Expr[]
body = MacroTools.flatten(def[:body])
push!(new_stmts, Expr(:aliasscope))
if force_inbounds
push!(new_stmts, Expr(:inbounds, true))
end
append!(new_stmts, split(body.args))
if force_inbounds
push!(new_stmts, Expr(:inbounds, :pop))
end
push!(new_stmts, Expr(:popaliasscope))
push!(new_stmts, :(return nothing))
def[:body] = Expr(:let,
Expand Down