use runmat_accelerate_api::GpuTensorHandle;
use runmat_builtins::{CharArray, ComplexTensor, Tensor, Value};
use runmat_macros::runtime_builtin;
use crate::builtins::common::spec::{
BroadcastSemantics, BuiltinFusionSpec, BuiltinGpuSpec, ConstantStrategy, FusionError,
FusionExprContext, FusionKernelTemplate, GpuOpKind, ProviderHook, ReductionNaN,
ResidencyPolicy, ScalarType, ShapeRequirements,
};
use crate::builtins::common::{gpu_helpers, map_control_flow_with_builtin, tensor};
use crate::builtins::math::type_resolvers::numeric_unary_type;
use crate::{build_runtime_error, BuiltinResult, RuntimeError};
#[runmat_macros::register_gpu_spec(builtin_path = "crate::builtins::math::elementwise::sign")]
pub const GPU_SPEC: BuiltinGpuSpec = BuiltinGpuSpec {
name: "sign",
op_kind: GpuOpKind::Elementwise,
supported_precisions: &[ScalarType::F32, ScalarType::F64],
broadcast: BroadcastSemantics::Matlab,
provider_hooks: &[ProviderHook::Unary { name: "unary_sign" }],
constant_strategy: ConstantStrategy::InlineLiteral,
residency: ResidencyPolicy::NewHandle,
nan_mode: ReductionNaN::Include,
two_pass_threshold: None,
workgroup_size: None,
accepts_nan_mode: false,
notes:
"Providers may execute sign on-device via unary_sign; the runtime gathers to the host when the hook is unavailable or complex normalisation is required.",
};
#[runmat_macros::register_fusion_spec(builtin_path = "crate::builtins::math::elementwise::sign")]
pub const FUSION_SPEC: BuiltinFusionSpec = BuiltinFusionSpec {
name: "sign",
shape: ShapeRequirements::BroadcastCompatible,
constant_strategy: ConstantStrategy::InlineLiteral,
elementwise: Some(FusionKernelTemplate {
scalar_precisions: &[ScalarType::F32, ScalarType::F64],
wgsl_body: |ctx: &FusionExprContext| {
let input = ctx
.inputs
.first()
.ok_or(FusionError::MissingInput(0))?;
Ok(format!("sign({input})"))
},
}),
reduction: None,
emits_nan: true,
notes: "Fusion kernels emit WGSL `sign` ops; providers can override via fused pipelines when advantageous.",
};
const BUILTIN_NAME: &str = "sign";
fn builtin_error(message: impl Into<String>) -> RuntimeError {
build_runtime_error(message)
.with_builtin(BUILTIN_NAME)
.build()
}
#[runtime_builtin(
name = "sign",
category = "math/elementwise",
summary = "Sign of scalars, vectors, matrices, or N-D tensors with real or complex values.",
keywords = "sign,signum,elementwise,complex,gpu",
accel = "unary",
type_resolver(numeric_unary_type),
builtin_path = "crate::builtins::math::elementwise::sign"
)]
async fn sign_builtin(value: Value) -> BuiltinResult<Value> {
match value {
Value::GpuTensor(handle) => sign_gpu(handle).await,
Value::Complex(re, im) => {
let (re_out, im_out) = sign_complex(re, im);
Ok(Value::Complex(re_out, im_out))
}
Value::ComplexTensor(ct) => sign_complex_tensor(ct),
Value::CharArray(ca) => sign_char_array(ca),
Value::String(_) | Value::StringArray(_) => Err(builtin_error(
"sign: expected numeric, logical, or character input",
)),
other => sign_real(other),
}
}
async fn sign_gpu(handle: GpuTensorHandle) -> BuiltinResult<Value> {
if let Some(provider) = runmat_accelerate_api::provider_for_handle(&handle) {
if let Ok(out) = provider.unary_sign(&handle).await {
return Ok(gpu_helpers::resident_gpu_value(out));
}
}
let tensor = gpu_helpers::gather_tensor_async(&handle)
.await
.map_err(|flow| map_control_flow_with_builtin(flow, BUILTIN_NAME))?;
Ok(tensor::tensor_into_value(sign_tensor(tensor)?))
}
fn sign_real(value: Value) -> BuiltinResult<Value> {
let tensor = tensor::value_into_tensor_for("sign", value)
.map_err(|e| builtin_error(format!("sign: {e}")))?;
Ok(tensor::tensor_into_value(sign_tensor(tensor)?))
}
fn sign_tensor(tensor: Tensor) -> BuiltinResult<Tensor> {
let data = tensor.data.iter().map(|&x| sign_real_scalar(x)).collect();
Tensor::new(data, tensor.shape.clone()).map_err(|e| builtin_error(format!("sign: {e}")))
}
fn sign_char_array(ca: CharArray) -> BuiltinResult<Value> {
let data = ca
.data
.iter()
.map(|&ch| sign_real_scalar(ch as u32 as f64))
.collect::<Vec<_>>();
let tensor = Tensor::new(data, vec![ca.rows, ca.cols])
.map_err(|e| builtin_error(format!("sign: {e}")))?;
Ok(Value::Tensor(tensor))
}
fn sign_complex_tensor(ct: ComplexTensor) -> BuiltinResult<Value> {
let mapped = ct
.data
.iter()
.map(|&(re, im)| sign_complex(re, im))
.collect::<Vec<_>>();
let tensor = ComplexTensor::new(mapped, ct.shape.clone())
.map_err(|e| builtin_error(format!("sign: {e}")))?;
Ok(Value::ComplexTensor(tensor))
}
#[inline]
fn sign_real_scalar(x: f64) -> f64 {
if x > 0.0 {
1.0
} else if x < 0.0 {
-1.0
} else if x == 0.0 {
0.0
} else {
x
}
}
fn sign_complex(re: f64, im: f64) -> (f64, f64) {
if re == 0.0 && im == 0.0 {
return (0.0, 0.0);
}
if re.is_nan() || im.is_nan() {
return (f64::NAN, f64::NAN);
}
let re_inf = re.is_infinite();
let im_inf = im.is_infinite();
if re_inf || im_inf {
let real = if re_inf { re.signum() } else { 0.0 };
let imag = if im_inf { im.signum() } else { 0.0 };
let norm = (real * real + imag * imag).sqrt();
if norm == 0.0 {
return (real, imag);
}
return (real / norm, imag / norm);
}
let scale = re.abs().max(im.abs());
if scale == 0.0 {
return (0.0, 0.0);
}
let nr = re / scale;
let ni = im / scale;
let magnitude = (nr * nr + ni * ni).sqrt();
if magnitude == 0.0 {
(0.0, 0.0)
} else {
(nr / magnitude, ni / magnitude)
}
}
#[cfg(test)]
pub(crate) mod tests {
use super::*;
use crate::builtins::common::test_support;
use futures::executor::block_on;
use runmat_builtins::{IntValue, LogicalArray, ResolveContext, Type};
fn sign_builtin(value: Value) -> BuiltinResult<Value> {
block_on(super::sign_builtin(value))
}
#[test]
fn sign_type_preserves_tensor_shape() {
let out = numeric_unary_type(
&[Type::Tensor {
shape: Some(vec![Some(2), Some(3)]),
}],
&ResolveContext::new(Vec::new()),
);
assert_eq!(
out,
Type::Tensor {
shape: Some(vec![Some(2), Some(3)])
}
);
}
#[test]
fn sign_type_scalar_tensor_returns_num() {
let out = numeric_unary_type(
&[Type::Tensor {
shape: Some(vec![Some(1), Some(1)]),
}],
&ResolveContext::new(Vec::new()),
);
assert_eq!(out, Type::Num);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_scalar_positive_negative_zero() {
assert_eq!(sign_builtin(Value::Num(3.5)).unwrap(), Value::Num(1.0));
assert_eq!(sign_builtin(Value::Num(-2.0)).unwrap(), Value::Num(-1.0));
assert_eq!(sign_builtin(Value::Num(0.0)).unwrap(), Value::Num(0.0));
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_scalar_nan_propagates() {
let result = sign_builtin(Value::Num(f64::NAN)).unwrap();
match result {
Value::Num(v) => assert!(v.is_nan()),
other => panic!("expected scalar NaN, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_tensor_mixed_values() {
let tensor = Tensor::new(vec![-2.0, -0.0, 0.0, 5.0], vec![2, 2]).unwrap();
let result = sign_builtin(Value::Tensor(tensor)).unwrap();
match result {
Value::Tensor(out) => {
assert_eq!(out.shape, vec![2, 2]);
assert_eq!(out.data, vec![-1.0, 0.0, 0.0, 1.0]);
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_complex_scalar_normalises() {
let result = sign_builtin(Value::Complex(3.0, 4.0)).unwrap();
match result {
Value::Complex(re, im) => {
assert!((re - 0.6).abs() < 1e-12);
assert!((im - 0.8).abs() < 1e-12);
}
other => panic!("expected complex value, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_complex_tensor_handles_zero() {
let tensor = ComplexTensor::new(vec![(0.0, 0.0), (1.0, -1.0)], vec![2, 1]).unwrap();
let result = sign_builtin(Value::ComplexTensor(tensor)).unwrap();
match result {
Value::ComplexTensor(out) => {
assert_eq!(out.shape, vec![2, 1]);
assert_eq!(out.data[0], (0.0, 0.0));
let (re, im) = out.data[1];
assert!((re - 0.7071067811865475).abs() < 1e-12);
assert!((im + 0.7071067811865475).abs() < 1e-12);
}
other => panic!("expected complex tensor, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_character_array() {
let ca = CharArray::new("RunMat".chars().collect(), 1, 6).unwrap();
let result = sign_builtin(Value::CharArray(ca)).unwrap();
match result {
Value::Tensor(out) => {
assert_eq!(out.shape, vec![1, 6]);
assert!(out.data.iter().all(|&v| (v - 1.0).abs() < 1e-12));
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_logical_array() {
let logical = LogicalArray::new(vec![0, 1, 0, 1], vec![2, 2]).unwrap();
let result = sign_builtin(Value::LogicalArray(logical)).unwrap();
match result {
Value::Tensor(out) => {
assert_eq!(out.shape, vec![2, 2]);
assert_eq!(out.data, vec![0.0, 1.0, 0.0, 1.0]);
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_int_values() {
let value = Value::Int(IntValue::I32(-7));
let result = sign_builtin(value).unwrap();
match result {
Value::Num(v) => assert_eq!(v, -1.0),
other => panic!("expected scalar result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_bool_values() {
let t = sign_builtin(Value::Bool(true)).unwrap();
let f = sign_builtin(Value::Bool(false)).unwrap();
assert_eq!(t, Value::Num(1.0));
assert_eq!(f, Value::Num(0.0));
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_infinite_values() {
let tensor = Tensor::new(
vec![f64::INFINITY, f64::NEG_INFINITY, 0.0, f64::NAN],
vec![2, 2],
)
.unwrap();
let result = sign_builtin(Value::Tensor(tensor)).unwrap();
match result {
Value::Tensor(out) => {
assert_eq!(out.data[0], 1.0);
assert_eq!(out.data[1], -1.0);
assert_eq!(out.data[2], 0.0);
assert!(out.data[3].is_nan());
}
other => panic!("expected tensor result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_string_input_errors() {
let err = sign_builtin(Value::String("runmat".to_string())).unwrap_err();
assert!(
err.message()
.contains("expected numeric, logical, or character input"),
"unexpected error message: {err}"
);
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_complex_with_nan() {
let result = sign_builtin(Value::Complex(f64::NAN, 1.0)).unwrap();
match result {
Value::Complex(re, im) => {
assert!(re.is_nan());
assert!(im.is_nan());
}
other => panic!("expected complex result, got {other:?}"),
}
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_gpu_provider_roundtrip() {
test_support::with_test_provider(|provider| {
let tensor = Tensor::new(vec![-3.0, -0.5, 0.0, 2.5], vec![2, 2]).unwrap();
let view = runmat_accelerate_api::HostTensorView {
data: &tensor.data,
shape: &tensor.shape,
};
let handle = provider.upload(&view).expect("upload");
let result = sign_builtin(Value::GpuTensor(handle)).expect("sign");
let gathered = test_support::gather(result).expect("gather");
assert_eq!(gathered.shape, vec![2, 2]);
assert_eq!(gathered.data, vec![-1.0, -1.0, 0.0, 1.0]);
});
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
fn sign_gpu_fallback_for_complex() {
test_support::with_test_provider(|provider| {
let tensor = Tensor::new(vec![3.0, -4.0], vec![2, 1]).unwrap();
let view = runmat_accelerate_api::HostTensorView {
data: &tensor.data,
shape: &tensor.shape,
};
let handle = provider.upload(&view).unwrap();
let result = sign_builtin(Value::GpuTensor(handle)).expect("sign");
let gathered = test_support::gather(result).expect("gather");
assert_eq!(gathered.data, vec![1.0, -1.0]);
});
}
#[cfg_attr(target_arch = "wasm32", wasm_bindgen_test::wasm_bindgen_test)]
#[test]
#[cfg(feature = "wgpu")]
fn sign_wgpu_matches_cpu() {
let _ = runmat_accelerate::backend::wgpu::provider::register_wgpu_provider(
runmat_accelerate::backend::wgpu::provider::WgpuProviderOptions::default(),
);
let tensor = Tensor::new(vec![-3.0, 0.0, 4.0, f64::NAN], vec![2, 2]).unwrap();
let cpu = sign_real(Value::Tensor(tensor.clone())).unwrap();
let view = runmat_accelerate_api::HostTensorView {
data: &tensor.data,
shape: &tensor.shape,
};
let handle = runmat_accelerate_api::provider()
.unwrap()
.upload(&view)
.unwrap();
let gpu = block_on(sign_gpu(handle)).unwrap();
let gathered = test_support::gather(gpu).expect("gather");
match (cpu, gathered) {
(Value::Tensor(ct), gt) => {
assert_eq!(gt.shape, ct.shape);
for (a, b) in gt.data.iter().zip(ct.data.iter()) {
if a.is_nan() && b.is_nan() {
continue;
}
assert_eq!(a, b);
}
}
_ => panic!("unexpected shapes"),
}
}
}