feat: Add estimate subcommand (#93)

Closes #78 

Uses the risc0 emulation tools to step through the ELF and return the
user cycles, total cycles and segment count.

.github/docker/Dockerfile.build

@@ -57,4 +57,4 @@ RUN /go/bin/yamlfmt -lint .github/workflows/*.yaml .github/workflows/*.yml .gith
 
 RUN cargo check
 RUN cargo +nightly fmt --all -- --check
-RUN cargo test   
+RUN cargo test

.github/workflows/integration-tests.yml

@@ -0,0 +1,34 @@
+name: Integration Tests
+on:
+  pull_request:
+    branches:
+      - '**'
+
+concurrency:
+  group: "integration-tests"
+  cancel-in-progress: true
+
+permissions: read-all
+
+jobs:
+  integration-tests:
+    name: Setup Toolchain and Test
+    runs-on: ubuntu-latest-m
+    permissions:
+      id-token: "write"
+      contents: "read"
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+      - name: Install Nix With Bonsol Binary Cache
+        uses: DeterminateSystems/nix-installer-action@main
+        with:
+          extra-conf: |
+            extra-substituters = https://bonsol.cachix.org
+            extra-trusted-public-keys = bonsol.cachix.org-1:yz7vi1rCPW1BpqoszdJvf08HZxQ/5gPTPxft4NnT74A=
+      - name: Setup Toolchain, Build and Test
+        run: |
+          nix develop --command bash -c "
+            cargo build &&
+            cargo test --features integration -- --nocapture
+          "

CHANGELOG.md

@@ -9,6 +9,9 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 ### Changed
 * `bonsol` cli option requirements and error messages updated for added clarity
 
+### Added
+* `bonsol estimate` for estimating execution cost of bonsol programs.
+
 ### Fixed
 * **Breaking**: `execute_v1` interface instruction now uses the new `InputRef` to improve CU usage.
 * Adds a callback struct to use the input_hash and committed_outputs from the callback program ergonomically.

cli/Cargo.toml

@@ -11,13 +11,16 @@ path = "src/main.rs"
 [features]
 mac = ["risc0-zkvm/metal"]
 linux = ["risc0-zkvm/cuda"]
+integration = []
 
 [dependencies]
 anyhow = "1.0.86"
 atty = "0.2.14"
 bincode = "1.3.3"
+bonsol-interface.workspace = true
 bonsol-prover = { path = "../prover" }
 bonsol-sdk = { path = "../sdk" }
+bytemuck = "1.15.0"
 hex = "0.4.3"
 byte-unit = "4.0.19"
 bytes = "1.4.0"
@@ -34,7 +37,9 @@ reqwest = { version = "0.11.26", features = [
   "native-tls-vendored",
 ] }
 risc0-binfmt = { workspace = true }
-risc0-zkvm = { workspace = true, features = ["prove"] }
+risc0-zkvm = { workspace = true, default-features = false, features = ["prove", "std"] }
+risc0-zkvm-platform = { git = "https://github.com/anagrambuild/risc0", branch = "v1.0.1-bonsai-fix" }
+risc0-circuit-rv32im = { git = "https://github.com/anagrambuild/risc0", branch = "v1.0.1-bonsai-fix" }
 serde = { version = "1.0.197", features = ["derive"] }
 serde_json = "1.0.104"
 sha2 = "0.10.6"
@@ -46,4 +51,6 @@ tera = "1.17.1"
 thiserror = "1.0.65"
 tokio = { version = "1.38.0", features = ["full"] }
 
-bonsol-interface.workspace = true
+[dev-dependencies]
+assert_cmd = "2.0.16"
+predicates = "3.1.2"

cli/README.md

@@ -58,3 +58,20 @@ todo
 
 ### Prove
 todo
+
+### Estimate
+
+You can estimate the number of cycles and segments using risc0 emulation to step through an ELF by passing the `estimate` command the path to a manifest.json and an inputs file (if required).
+
+```
+bonsol -k ./keypair.json -u http://localhost:8899 estimate \
+    --manifest-path program/manifest.json \
+    --input-file program/inputs.json \
+    --max-cycles 16777216 # this is the default
+
+# Example Output:
+#
+# User cycles: 3380
+# Total cycles: 65536
+# Segments: 1
+```

cli/src/command.rs

@@ -227,6 +227,7 @@ pub enum Command {
         )]
         auto_confirm: bool,
     },
+    #[command(about = "Build a ZK program")]
     Build {
         #[arg(
             help = "The path to a ZK program folder containing a Cargo.toml",
@@ -235,6 +236,25 @@ pub enum Command {
         )]
         zk_program_path: String,
     },
+    #[command(about = "Estimate the execution cost of a ZK RISC0 program")]
+    Estimate {
+        #[arg(
+            help = "The path to the program's manifest file (manifest.json)",
+            short = 'm',
+            long
+        )]
+        manifest_path: String,
+
+        #[arg(help = "The path to the program input file", short = 'i', long)]
+        input_file: Option<String>,
+
+        #[arg(
+            help = "Set the maximum number of cycles [default: 16777216u64]",
+            short = 'c',
+            long
+        )]
+        max_cycles: Option<u64>,
+    },
     Execute {
         #[arg(short = 'f', long)]
         execution_request_file: Option<String>,
@@ -296,6 +316,11 @@ pub enum ParsedCommand {
     Build {
         zk_program_path: String,
     },
+    Estimate {
+        manifest_path: String,
+        input_file: Option<String>,
+        max_cycles: Option<u64>,
+    },
     Execute {
         execution_request_file: Option<String>,
 
@@ -351,6 +376,15 @@ impl TryFrom<Command> for ParsedCommand {
                 ),
             }),
             Command::Build { zk_program_path } => Ok(ParsedCommand::Build { zk_program_path }),
+            Command::Estimate {
+                manifest_path,
+                input_file,
+                max_cycles,
+            } => Ok(ParsedCommand::Estimate {
+                manifest_path,
+                input_file,
+                max_cycles,
+            }),
             Command::Execute {
                 execution_request_file,
                 program_id,

cli/src/estimate.rs

@@ -0,0 +1,74 @@
+//! Bare bones upper bound estimator that uses the rv32im
+//! emulation utils for fast lookups in the opcode list
+//! to extract the cycle count from an elf.
+
+use anyhow::Result;
+use risc0_binfmt::{MemoryImage, Program};
+use risc0_zkvm::{ExecutorEnv, ExecutorImpl, Session, GUEST_MAX_MEM};
+use risc0_zkvm_platform::PAGE_SIZE;
+
+pub fn estimate<E: MkImage>(elf: E, env: ExecutorEnv) -> Result<()> {
+    let session = get_session(elf, env)?;
+    println!(
+        "User cycles: {}\nTotal cycles: {}\nSegments: {}",
+        session.user_cycles,
+        session.total_cycles,
+        session.segments.len()
+    );
+
+    Ok(())
+}
+
+/// Get the total number of cycles by stepping through the ELF using emulation
+/// tools from the risc0_circuit_rv32im module.
+pub fn get_session<E: MkImage>(elf: E, env: ExecutorEnv) -> Result<Session> {
+    Ok(ExecutorImpl::new(env, elf.mk_image()?)?.run()?)
+}
+
+/// Helper trait for loading an image from an elf.
+pub trait MkImage {
+    fn mk_image(self) -> Result<MemoryImage>;
+}
+impl<'a> MkImage for &'a [u8] {
+    fn mk_image(self) -> Result<MemoryImage> {
+        let program = Program::load_elf(self, GUEST_MAX_MEM as u32)?;
+        MemoryImage::new(&program, PAGE_SIZE as u32)
+    }
+}
+
+#[cfg(test)]
+mod estimate_tests {
+    use anyhow::Result;
+    use risc0_binfmt::MemoryImage;
+    use risc0_circuit_rv32im::prove::emu::{
+        exec::DEFAULT_SEGMENT_LIMIT_PO2,
+        testutil::{basic as basic_test_program, DEFAULT_SESSION_LIMIT},
+    };
+    use risc0_zkvm::{ExecutorEnv, PAGE_SIZE};
+
+    use super::MkImage;
+    use crate::estimate;
+
+    impl MkImage for MemoryImage {
+        fn mk_image(self) -> Result<MemoryImage> {
+            Ok(self)
+        }
+    }
+
+    #[test]
+    fn estimate_basic() {
+        let program = basic_test_program();
+        let mut env = &mut ExecutorEnv::builder();
+        env = env
+            .segment_limit_po2(DEFAULT_SEGMENT_LIMIT_PO2 as u32)
+            .session_limit(DEFAULT_SESSION_LIMIT);
+        let image = MemoryImage::new(&program, PAGE_SIZE as u32)
+            .expect("failed to create image from basic program");
+        let res = estimate::get_session(image, env.build().unwrap());
+
+        assert_eq!(
+            res.ok().and_then(|session| Some(session.total_cycles)),
+            Some(16384)
+        );
+    }
+}