Intrinsic test tool to compare neon intrinsics with C (rust-lang#1170)

Author: JamieCunliffe

.gitignore

@@ -4,3 +4,5 @@ target
 tags
 crates/stdarch-gen/aarch64.rs
 crates/stdarch-gen/arm.rs
+c_programs/*
+rust_programs/*

Cargo.toml

@@ -4,6 +4,7 @@ members = [
   "crates/core_arch",
   "crates/std_detect",
   "crates/stdarch-gen",
+  "crates/intrinsic-test",
   "examples/"
 ]
 exclude = [

ci/docker/aarch64-unknown-linux-gnu/Dockerfile

@@ -1,13 +1,17 @@
 FROM ubuntu:20.04
 RUN apt-get update && apt-get install -y --no-install-recommends \
   gcc \
+  g++ \
   ca-certificates \
   libc6-dev \
   gcc-aarch64-linux-gnu \
+  g++-aarch64-linux-gnu \
   libc6-dev-arm64-cross \
   qemu-user \
   make \
-  file
+  file \
+  clang-12 \
+  lld
 
 ENV CARGO_TARGET_AARCH64_UNKNOWN_LINUX_GNU_LINKER=aarch64-linux-gnu-gcc \
     CARGO_TARGET_AARCH64_UNKNOWN_LINUX_GNU_RUNNER="qemu-aarch64 -L /usr/aarch64-linux-gnu" \

ci/run-docker.sh

@@ -9,7 +9,7 @@ run() {
     target=$(echo "${1}" | sed 's/-emulated//')
     echo "Building docker container for TARGET=${1}"
     docker build -t stdarch -f "ci/docker/${1}/Dockerfile" ci/
-    mkdir -p target
+    mkdir -p target c_programs rust_programs
     echo "Running docker"
     # shellcheck disable=SC2016
     docker run \
@@ -29,6 +29,8 @@ run() {
       --volume "$(rustc --print sysroot)":/rust:ro \
       --volume "$(pwd)":/checkout:ro \
       --volume "$(pwd)"/target:/checkout/target \
+      --volume "$(pwd)"/c_programs:/checkout/c_programs \
+      --volume "$(pwd)"/rust_programs:/checkout/rust_programs \
       --init \
       --workdir /checkout \
       --privileged \

ci/run.sh

@@ -65,6 +65,8 @@ cargo_test() {
 CORE_ARCH="--manifest-path=crates/core_arch/Cargo.toml"
 STD_DETECT="--manifest-path=crates/std_detect/Cargo.toml"
 STDARCH_EXAMPLES="--manifest-path=examples/Cargo.toml"
+INTRINSIC_TEST="--manifest-path=crates/intrinsic-test/Cargo.toml"
+
 cargo_test "${CORE_ARCH} --release"
 
 if [ "$NOSTD" != "1" ]; then
@@ -111,6 +113,11 @@ case ${TARGET} in
 
 esac
 
+if [ "${TARGET}" = "aarch64-unknown-linux-gnu" ]; then
+    export CPPFLAGS="-fuse-ld=lld -I/usr/aarch64-linux-gnu/include/ -I/usr/aarch64-linux-gnu/include/c++/9/aarch64-linux-gnu/"
+    cargo run ${INTRINSIC_TEST} --release --bin intrinsic-test -- crates/intrinsic-test/neon-intrinsics.csv --runner "${CARGO_TARGET_AARCH64_UNKNOWN_LINUX_GNU_RUNNER}" --cppcompiler "clang++-12"
+fi
+
 if [ "$NORUN" != "1" ] && [ "$NOSTD" != 1 ]; then
     # Test examples
     (

crates/intrinsic-test/Cargo.toml

@@ -0,0 +1,16 @@
+[package]
+name = "intrinsic-test"
+version = "0.1.0"
+authors = ["Jamie Cunliffe <[email protected]>"]
+edition = "2018"
+
+[dependencies]
+lazy_static = "1.4.0"
+serde = { version = "1", features = ["derive"] }
+csv = "1.1"
+clap = "2.33.3"
+regex = "1.4.2"
+log = "0.4.11"
+pretty_env_logger = "0.4.0"
+rayon = "1.5.0"
+diff = "0.1.12"

crates/intrinsic-test/README.md

@@ -0,0 +1,24 @@
+Generate and run programs using equivalent C and Rust intrinsics, checking that
+each produces the same result from random inputs.
+
+# Usage
+```
+USAGE:
+    intrinsic-test [OPTIONS] <INPUT>
+
+FLAGS:
+    -h, --help       Prints help information
+    -V, --version    Prints version information
+
+OPTIONS:
+        --cppcompiler <CPPCOMPILER>    The C++ compiler to use for compiling the c++ code [default: clang++]
+        --runner <RUNNER>              Run the C programs under emulation with this command
+        --toolchain <TOOLCHAIN>        The rust toolchain to use for building the rust code
+
+ARGS:
+    <INPUT>    The input file containing the intrinsics
+```
+
+The intrinsic.csv is the arm neon tracking google sheet (https://docs.google.com/spreadsheets/d/1MqW1g8c7tlhdRWQixgdWvR4uJHNZzCYAf4V0oHjZkwA/edit#gid=0)
+that contains the intrinsic list. The done percentage column should be renamed to "enabled".
+

crates/intrinsic-test/neon-intrinsics.csv

@@ -0,0 +1,137 @@
+use serde::{Deserialize, Deserializer};
+
+use crate::types::IntrinsicType;
+use crate::Language;
+
+/// An argument for the intrinsic.
+#[derive(Debug, PartialEq, Clone)]
+pub struct Argument {
+    /// The argument's index in the intrinsic function call.
+    pub pos: usize,
+    /// The argument name.
+    pub name: String,
+    /// The type of the argument.
+    pub ty: IntrinsicType,
+}
+
+impl Argument {
+    /// Creates an argument from a Rust style signature i.e. `name: type`
+    fn from_rust(pos: usize, arg: &str) -> Result<Self, String> {
+        let mut parts = arg.split(':');
+        let name = parts.next().unwrap().trim().to_string();
+        let ty = IntrinsicType::from_rust(parts.next().unwrap().trim())?;
+
+        Ok(Self { pos, name, ty })
+    }
+
+    fn to_c_type(&self) -> String {
+        self.ty.c_type()
+    }
+
+    fn is_simd(&self) -> bool {
+        self.ty.is_simd()
+    }
+
+    pub fn is_ptr(&self) -> bool {
+        self.ty.is_ptr()
+    }
+}
+
+#[derive(Debug, PartialEq, Clone)]
+pub struct ArgumentList {
+    pub args: Vec<Argument>,
+}
+
+impl ArgumentList {
+    /// Creates an argument list from a Rust function signature, the data for
+    /// this function should only be the arguments.
+    /// e.g. for `fn test(a: u32, b: u32) -> u32` data should just be `a: u32, b: u32`
+    fn from_rust_arguments(data: &str) -> Result<Self, String> {
+        let args = data
+            .split(',')
+            .enumerate()
+            .map(|(idx, arg)| Argument::from_rust(idx, arg))
+            .collect::<Result<_, _>>()?;
+
+        Ok(Self { args })
+    }
+
+    /// Converts the argument list into the call paramters for a C function call.
+    /// e.g. this would generate something like `a, &b, c`
+    pub fn as_call_param_c(&self) -> String {
+        self.args
+            .iter()
+            .map(|arg| match arg.ty {
+                IntrinsicType::Ptr { .. } => {
+                    format!("&{}", arg.name)
+                }
+                IntrinsicType::Type { .. } => arg.name.clone(),
+            })
+            .collect::<Vec<String>>()
+            .join(", ")
+    }
+
+    /// Converts the argument list into the call paramters for a Rust function.
+    /// e.g. this would generate something like `a, b, c`
+    pub fn as_call_param_rust(&self) -> String {
+        self.args
+            .iter()
+            .map(|arg| arg.name.clone())
+            .collect::<Vec<String>>()
+            .join(", ")
+    }
+
+    /// Creates a line that initializes this argument for C code.
+    /// e.g. `int32x2_t a = { 0x1, 0x2 };`
+    pub fn init_random_values_c(&self, pass: usize) -> String {
+        self.iter()
+            .map(|arg| {
+                format!(
+                    "{ty} {name} = {{ {values} }};",
+                    ty = arg.to_c_type(),
+                    name = arg.name,
+                    values = arg.ty.populate_random(pass, &Language::C)
+                )
+            })
+            .collect::<Vec<_>>()
+            .join("\n    ")
+    }
+
+    /// Creates a line that initializes this argument for Rust code.
+    /// e.g. `let a = transmute([0x1, 0x2]);`
+    pub fn init_random_values_rust(&self, pass: usize) -> String {
+        self.iter()
+            .map(|arg| {
+                if arg.is_simd() {
+                    format!(
+                        "let {name} = ::std::mem::transmute([{values}]);",
+                        name = arg.name,
+                        values = arg.ty.populate_random(pass, &Language::Rust),
+                    )
+                } else {
+                    format!(
+                        "let {name} = {value};",
+                        name = arg.name,
+                        value = arg.ty.populate_random(pass, &Language::Rust)
+                    )
+                }
+            })
+            .collect::<Vec<_>>()
+            .join("\n        ")
+    }
+
+    pub fn iter(&self) -> std::slice::Iter<'_, Argument> {
+        self.args.iter()
+    }
+}
+
+impl<'de> Deserialize<'de> for ArgumentList {
+    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
+    where
+        D: Deserializer<'de>,
+    {
+        use serde::de::Error;
+        let s = String::deserialize(deserializer)?;
+        Self::from_rust_arguments(&s).map_err(Error::custom)
+    }
+}

crates/intrinsic-test/src/argument.rs

@@ -0,0 +1,112 @@
+use crate::types::{IntrinsicType, TypeKind};
+
+use super::argument::ArgumentList;
+use serde::de::Unexpected;
+use serde::{de, Deserialize, Deserializer};
+
+/// An intrinsic
+#[derive(Deserialize, Debug, PartialEq, Clone)]
+pub struct Intrinsic {
+    /// If the intrinsic should be tested.
+    #[serde(deserialize_with = "bool_from_string")]
+    pub enabled: bool,
+
+    /// The function name of this intrinsic.
+    pub name: String,
+
+    /// Any arguments for this intrinsinc.
+    #[serde(rename = "args")]
+    pub arguments: ArgumentList,
+
+    /// The return type of this intrinsic.
+    #[serde(rename = "return")]
+    pub results: IntrinsicType,
+}
+
+impl Intrinsic {
+    /// Generates a std::cout for the intrinsics results that will match the
+    /// rust debug output format for the return type.
+    pub fn print_result_c(&self, index: usize) -> String {
+        let lanes = if self.results.num_lanes() > 1 {
+            (0..self.results.num_lanes())
+                .map(|idx| -> std::string::String {
+                    format!(
+                        "{cast}{lane_fn}(__return_value, {lane})",
+                        cast = self.results.c_promotion(),
+                        lane_fn = self.results.get_lane_function(),
+                        lane = idx
+                    )
+                })
+                .collect::<Vec<_>>()
+                .join(r#" << ", " << "#)
+        } else {
+            format!(
+                "{promote}cast<{cast}>(__return_value)",
+                cast = match self.results.kind() {
+                    TypeKind::Float if self.results.inner_size() == 32 => "float".to_string(),
+                    TypeKind::Float if self.results.inner_size() == 64 => "double".to_string(),
+                    TypeKind::Int => format!("int{}_t", self.results.inner_size()),
+                    TypeKind::UInt => format!("uint{}_t", self.results.inner_size()),
+                    TypeKind::Poly => format!("poly{}_t", self.results.inner_size()),
+                    ty => todo!("print_result_c - Unknown type: {:#?}", ty),
+                },
+                promote = self.results.c_promotion(),
+            )
+        };
+
+        format!(
+            r#"std::cout << "Result {idx}: {ty}" << std::fixed << std::setprecision(150) <<  {lanes} << "{close}" << std::endl;"#,
+            ty = if self.results.is_simd() {
+                format!("{}(", self.results.c_type())
+            } else {
+                String::from("")
+            },
+            close = if self.results.is_simd() { ")" } else { "" },
+            lanes = lanes,
+            idx = index,
+        )
+    }
+
+    pub fn generate_pass_rust(&self, index: usize) -> String {
+        format!(
+            r#"
+    unsafe {{
+        {initialized_args}
+        let res = {intrinsic_call}({args});
+        println!("Result {idx}: {{:.150?}}", res);
+    }}"#,
+            initialized_args = self.arguments.init_random_values_rust(index),
+            intrinsic_call = self.name,
+            args = self.arguments.as_call_param_rust(),
+            idx = index,
+        )
+    }
+
+    pub fn generate_pass_c(&self, index: usize) -> String {
+        format!(
+            r#"  {{
+    {initialized_args}
+    auto __return_value = {intrinsic_call}({args});
+    {print_result}
+  }}"#,
+            initialized_args = self.arguments.init_random_values_c(index),
+            intrinsic_call = self.name,
+            args = self.arguments.as_call_param_c(),
+            print_result = self.print_result_c(index)
+        )
+    }
+}
+
+fn bool_from_string<'de, D>(deserializer: D) -> Result<bool, D::Error>
+where
+    D: Deserializer<'de>,
+{
+    match String::deserialize(deserializer)?.to_uppercase().as_ref() {
+        "TRUE" => Ok(true),
+        "FALSE" => Ok(false),
+        other => Err(de::Error::invalid_value(
+            Unexpected::Str(other),
+            &"TRUE or FALSE",
+        )),
+    }
+}