Skip to content

feat: optimize out unnecessary truncation instructions #3717

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

Merged
TomAFrench merged 4 commits into from
Dec 9, 2023
Merged

feat: optimize out unnecessary truncation instructions #3717

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
588f58e
feat: remove unnecessary truncations into larger types
TomAFrench Dec 6, 2023
80b67fe
chore: use specialised functions over `insert_instruction`
TomAFrench Dec 7, 2023
5e8e567
Merge branch 'master' into tf/remove-unnecessary-truncation
kevaundray Dec 9, 2023
5d8c293
Merge branch 'master' into tf/remove-unnecessary-truncation
kevaundray Dec 9, 2023
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
11 changes: 3 additions & 8 deletions compiler/noirc_evaluator/src/ssa/function_builder/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -314,16 +314,11 @@ impl FunctionBuilder {
(FieldElement::max_num_bits(), self.insert_binary(predicate, BinaryOp::Mul, pow))
};

let instruction = Instruction::Binary(Binary { lhs, rhs: pow, operator: BinaryOp::Mul });
if max_bit <= bit_size {
self.insert_instruction(instruction, None).first()
self.insert_binary(lhs, BinaryOp::Mul, pow)
} else {
let result = self.insert_instruction(instruction, None).first();
self.insert_instruction(
Instruction::Truncate { value: result, bit_size, max_bit_size: max_bit },
None,
)
.first()
let result = self.insert_binary(lhs, BinaryOp::Mul, pow);
self.insert_truncate(result, bit_size, max_bit)
}
}

Expand Down
14 changes: 13 additions & 1 deletion compiler/noirc_evaluator/src/ssa/ir/instruction.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -467,11 +467,23 @@ impl Instruction {
}
None
}
Instruction::Truncate { value, bit_size, .. } => {
Instruction::Truncate { value, bit_size, max_bit_size } => {
if let Some((numeric_constant, typ)) = dfg.get_numeric_constant_with_type(*value) {
let integer_modulus = 2_u128.pow(*bit_size);
let truncated = numeric_constant.to_u128() % integer_modulus;
SimplifiedTo(dfg.make_constant(truncated.into(), typ))
} else if let Value::Instruction { instruction, .. } = &dfg[dfg.resolve(*value)] {
if let Instruction::Truncate { bit_size: src_bit_size, .. } = &dfg[*instruction]
{
// If we're truncating the value to fit into the same or larger bit size then this is a noop.
if src_bit_size <= bit_size && src_bit_size <= max_bit_size {
SimplifiedTo(*value)
} else {
None
}
} else {
None
}
} else {
None
}
Expand Down
49 changes: 12 additions & 37 deletions compiler/noirc_evaluator/src/ssa/ssa_gen/context.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -275,16 +275,11 @@ impl<'a> FunctionContext<'a> {
let bit_width =
self.builder.numeric_constant(FieldElement::from(2_i128.pow(bit_size)), Type::field());
let sign_not = self.builder.insert_binary(one, BinaryOp::Sub, sign);
let as_field =
self.builder.insert_instruction(Instruction::Cast(input, Type::field()), None).first();
let sign_field =
self.builder.insert_instruction(Instruction::Cast(sign, Type::field()), None).first();
let as_field = self.builder.insert_cast(input, Type::field());
let sign_field = self.builder.insert_cast(sign, Type::field());
let positive_predicate = self.builder.insert_binary(sign_field, BinaryOp::Mul, as_field);
let two_complement = self.builder.insert_binary(bit_width, BinaryOp::Sub, as_field);
let sign_not_field = self
.builder
.insert_instruction(Instruction::Cast(sign_not, Type::field()), None)
.first();
let sign_not_field = self.builder.insert_cast(sign_not, Type::field());
let negative_predicate =
self.builder.insert_binary(sign_not_field, BinaryOp::Mul, two_complement);
self.builder.insert_binary(positive_predicate, BinaryOp::Add, negative_predicate)
Expand Down Expand Up @@ -315,17 +310,8 @@ impl<'a> FunctionContext<'a> {
match operator {
BinaryOpKind::Add | BinaryOpKind::Subtract => {
// Result is computed modulo the bit size
let mut result = self
.builder
.insert_instruction(
Instruction::Truncate {
value: result,
bit_size,
max_bit_size: bit_size + 1,
},
None,
)
.first();
let mut result =
self.builder.insert_truncate(result, bit_size, bit_size + 1);
result = self.builder.insert_cast(result, Type::unsigned(bit_size));

self.check_signed_overflow(result, lhs, rhs, operator, bit_size, location);
Expand All @@ -335,17 +321,7 @@ impl<'a> FunctionContext<'a> {
// Result is computed modulo the bit size
let mut result =
self.builder.insert_cast(result, Type::unsigned(2 * bit_size));
result = self
.builder
.insert_instruction(
Instruction::Truncate {
value: result,
bit_size,
max_bit_size: 2 * bit_size,
},
None,
)
.first();
result = self.builder.insert_truncate(result, bit_size, 2 * bit_size);

self.check_signed_overflow(result, lhs, rhs, operator, bit_size, location);
self.builder.insert_cast(result, result_type)
Expand Down Expand Up @@ -476,17 +452,16 @@ impl<'a> FunctionContext<'a> {

// Then we check the signed product fits in a signed integer of bit_size-bits
let not_same = self.builder.insert_binary(one, BinaryOp::Sub, same_sign);
let not_same_sign_field = self
.builder
.insert_instruction(Instruction::Cast(not_same, Type::unsigned(bit_size)), None)
.first();
let not_same_sign_field =
self.builder.insert_cast(not_same, Type::unsigned(bit_size));
let positive_maximum_with_offset =
self.builder.insert_binary(half_width, BinaryOp::Add, not_same_sign_field);
let product_overflow_check =
self.builder.insert_binary(product, BinaryOp::Lt, positive_maximum_with_offset);
self.builder.set_location(location).insert_instruction(
Instruction::Constrain(product_overflow_check, one, Some(message)),
None,
self.builder.set_location(location).insert_constrain(
product_overflow_check,
one,
Some(message),
);
}
BinaryOpKind::ShiftLeft => unreachable!("shift is not supported for signed integer"),
Expand Down