Handbook
Glossary
##single>double-float
Basic block instructions
Prev:
##shuffle-vector-imm
Next:
##store-memory-imm
Vocabulary
compiler
.
cfg
.
instructions
Class description
Converts a single precision value (32-bit usually) stored in a SIMD register to a double precision one (64-bit usually).
Definition
IN:
compiler.cfg.instructions
TUPLE:
##single>double-float
<
foldable-insn
dst src insn#
;
Methods
USING:
accessors
compiler.cfg.gvn.avail
compiler.cfg.instructions
kernel
;
M:
##single>double-float
>avail-insn-defs
[
]
change-dst
drop
;
USING:
accessors
compiler.cfg.gvn.avail
compiler.cfg.instructions
kernel
;
M:
##single>double-float
>avail-insn-uses
[
dup
>avail-vreg
or*
]
change-src
drop
;
USING:
accessors
arrays
compiler.cfg.gvn.expressions
compiler.cfg.gvn.graph
compiler.cfg.instructions
kernel
sequences.private
;
M:
##single>double-float
>expr
\
##single>double-float
swap
[
src>>
vreg>vn
]
keep
drop
2
f
<array>
1
swap
[
set-array-nth
]
keep
0
swap
[
set-array-nth
]
keep
;
USING:
accessors
arrays
compiler.cfg.instructions
compiler.cfg.value-numbering.expressions
compiler.cfg.value-numbering.graph
kernel
sequences.private
;
M:
##single>double-float
>expr
\
##single>double-float
swap
[
src>>
vreg>vn
]
keep
drop
2
f
<array>
1
swap
[
set-array-nth
]
keep
0
swap
[
set-array-nth
]
keep
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.linear-scan.assignment
kernel
;
M:
##single>double-float
assign-insn-defs
[
vreg>reg
]
change-dst
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.linear-scan.assignment
kernel
;
M:
##single>double-float
assign-insn-uses
[
vreg>reg
]
change-src
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.representations.rewrite
kernel
;
M:
##single>double-float
convert-insn-defs
[
converted-value
]
change-dst
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.representations.rewrite
kernel
;
M:
##single>double-float
convert-insn-uses
[
converted-value
]
change-src
drop
;
USING:
compiler.cfg.instructions
compiler.cfg.representations.preferred
cpu.architecture
kernel
;
M:
##single>double-float
defs-vreg-reps
drop
{
double-rep
}
;
USING:
accessors
arrays
compiler.cfg.def-use
compiler.cfg.instructions
;
M:
##single>double-float
defs-vregs
dst>>
1array
;
USING:
accessors
compiler.cfg.instructions
compiler.codegen
cpu.architecture
kernel
;
M:
##single>double-float
generate-insn
[
dst>>
]
keep
[
src>>
]
keep
drop
%single>double-float
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.renaming
kernel
;
M:
##single>double-float
rename-insn-defs
[
rename-value
]
change-dst
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.renaming
kernel
;
M:
##single>double-float
rename-insn-uses
[
rename-value
]
change-src
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.ssa.construction.private
kernel
;
M:
##single>double-float
ssa-rename-insn-defs
[
gen-name
]
change-dst
drop
;
USING:
accessors
compiler.cfg.instructions
compiler.cfg.ssa.construction.private
kernel
;
M:
##single>double-float
ssa-rename-insn-uses
[
top-name
]
change-src
drop
;
USING:
compiler.cfg.instructions
compiler.cfg.representations.preferred
cpu.architecture
kernel
;
M:
##single>double-float
uses-vreg-reps
drop
{
float-rep
}
;
USING:
accessors
arrays
compiler.cfg.def-use
compiler.cfg.instructions
;
M:
##single>double-float
uses-vregs
src>>
1array
;