Extract program

Author: georgwiese

cli-openvm/src/main.rs

@@ -4,7 +4,7 @@ use metrics_util::{debugging::DebuggingRecorder, layers::Layer};
 use openvm_sdk::StdIn;
 use openvm_stark_sdk::bench::serialize_metric_snapshot;
 use powdr_autoprecompiles::pgo::{pgo_config, PgoType};
-use powdr_openvm::{default_powdr_openvm_config, CompiledProgram, GuestOptions};
+use powdr_openvm::{compile_openvm, default_powdr_openvm_config, CompiledProgram, GuestOptions};
 
 #[cfg(feature = "metrics")]
 use openvm_stark_sdk::metrics_tracing::TimingMetricsLayer;
@@ -140,14 +140,13 @@ fn run_command(command: Commands) {
             if let Some(apc_candidates_dir) = apc_candidates_dir {
                 powdr_config = powdr_config.with_apc_candidates_dir(apc_candidates_dir);
             }
-            let execution_profile = powdr_openvm::execution_profile_from_guest(
-                &guest,
-                guest_opts.clone(),
-                stdin_from(input),
-            );
+            let guest_program = compile_openvm(&guest, guest_opts.clone()).unwrap();
+            let execution_profile =
+                powdr_openvm::execution_profile_from_guest(&guest_program, stdin_from(input));
+
             let pgo_config = pgo_config(pgo, max_columns, execution_profile);
             let program =
-                powdr_openvm::compile_guest(&guest, guest_opts, powdr_config, pgo_config).unwrap();
+                powdr_openvm::compile_exe(guest_program, powdr_config, pgo_config).unwrap();
             write_program_to_file(program, &format!("{guest}_compiled.cbor")).unwrap();
         }
 
@@ -165,16 +164,13 @@ fn run_command(command: Commands) {
             if let Some(apc_candidates_dir) = apc_candidates_dir {
                 powdr_config = powdr_config.with_apc_candidates_dir(apc_candidates_dir);
             }
-            let execution_profile = powdr_openvm::execution_profile_from_guest(
-                &guest,
-                guest_opts.clone(),
-                stdin_from(input),
-            );
+            let guest_program = compile_openvm(&guest, guest_opts.clone()).unwrap();
+            let execution_profile =
+                powdr_openvm::execution_profile_from_guest(&guest_program, stdin_from(input));
             let pgo_config = pgo_config(pgo, max_columns, execution_profile);
             let compile_and_exec = || {
                 let program =
-                    powdr_openvm::compile_guest(&guest, guest_opts, powdr_config, pgo_config)
-                        .unwrap();
+                    powdr_openvm::compile_exe(guest_program, powdr_config, pgo_config).unwrap();
                 powdr_openvm::execute(program, stdin_from(input)).unwrap();
             };
             if let Some(metrics_path) = metrics {
@@ -203,16 +199,14 @@ fn run_command(command: Commands) {
             if let Some(apc_candidates_dir) = apc_candidates_dir {
                 powdr_config = powdr_config.with_apc_candidates_dir(apc_candidates_dir);
             }
-            let execution_profile = powdr_openvm::execution_profile_from_guest(
-                &guest,
-                guest_opts.clone(),
-                stdin_from(input),
-            );
+            let guest_program = compile_openvm(&guest, guest_opts).unwrap();
+
+            let execution_profile =
+                powdr_openvm::execution_profile_from_guest(&guest_program, stdin_from(input));
             let pgo_config = pgo_config(pgo, max_columns, execution_profile);
             let compile_and_prove = || {
                 let program =
-                    powdr_openvm::compile_guest(&guest, guest_opts, powdr_config, pgo_config)
-                        .unwrap();
+                    powdr_openvm::compile_exe(guest_program, powdr_config, pgo_config).unwrap();
                 powdr_openvm::prove(&program, mock, recursion, stdin_from(input), None).unwrap()
             };
             if let Some(metrics_path) = metrics {

openvm/src/customize_exe.rs

@@ -1,4 +1,4 @@
-use std::collections::{BTreeSet, HashMap};
+use std::collections::HashMap;
 
 use std::fmt::Display;
 use std::hash::Hash;
@@ -10,7 +10,6 @@ use crate::bus_map::OpenVmBusType;
 use crate::extraction_utils::{get_air_metrics, AirWidthsDiff, OriginalAirs, OriginalVmConfig};
 use crate::instruction_formatter::openvm_instruction_formatter;
 use crate::memory_bus_interaction::OpenVmMemoryBusInteraction;
-use crate::opcode::branch_opcodes_bigint_set;
 use crate::powdr_extension::chip::PowdrAir;
 use crate::utils::UnsupportedOpenVmReferenceError;
 use crate::OriginalCompiledProgram;
@@ -27,7 +26,7 @@ use openvm_stark_sdk::p3_baby_bear::BabyBear;
 use powdr_autoprecompiles::adapter::{
     Adapter, AdapterApc, AdapterApcWithStats, AdapterVmConfig, ApcWithStats, PgoAdapter,
 };
-use powdr_autoprecompiles::blocks::{collect_basic_blocks, BasicBlock, Instruction, Program};
+use powdr_autoprecompiles::blocks::{BasicBlock, Instruction, Program};
 use powdr_autoprecompiles::evaluation::{evaluate_apc, EvaluationResult};
 use powdr_autoprecompiles::expression::try_convert;
 use powdr_autoprecompiles::pgo::{ApcCandidateJsonExport, Candidate, KnapsackItem};
@@ -78,7 +77,7 @@ impl<'a> Adapter for BabyBearOpenVmApcAdapter<'a> {
 
 /// A newtype wrapper around `OpenVmProgram` to implement the `Program` trait.
 /// This is necessary because we cannot implement a foreign trait for a foreign type.
-pub struct Prog<'a, F>(&'a OpenVmProgram<F>);
+pub struct Prog<'a, F>(pub &'a OpenVmProgram<F>);
 
 impl<'a, F> From<&'a OpenVmProgram<F>> for Prog<'a, F> {
     fn from(program: &'a OpenVmProgram<F>) -> Self {
@@ -140,30 +139,20 @@ impl<'a, F: PrimeField32> Program<Instr<F>> for Prog<'a, F> {
 }
 
 pub fn customize<'a, P: PgoAdapter<Adapter = BabyBearOpenVmApcAdapter<'a>>>(
-    OriginalCompiledProgram {
-        exe,
-        vm_config,
-        elf,
-    }: OriginalCompiledProgram,
+    original_program: OriginalCompiledProgram,
     config: PowdrConfig,
     pgo: P,
 ) -> CompiledProgram {
-    let labels = elf.text_labels();
-    let debug_info = elf.debug_info();
-    let original_config = OriginalVmConfig::new(vm_config.clone());
+    let original_config = OriginalVmConfig::new(original_program.vm_config.clone());
     let airs = original_config.airs(config.degree_bound.identities).expect("Failed to convert the AIR of an OpenVM instruction, even after filtering by the blacklist!");
     let bus_map = original_config.bus_map();
 
-    let jumpdest_set = add_extra_targets(
-        &exe.program,
-        labels.clone(),
-        exe.program.pc_base,
-        DEFAULT_PC_STEP,
-    );
-
-    let program = Prog(&exe.program);
-
-    let range_tuple_checker_sizes = vm_config.sdk.rv32m.unwrap().range_tuple_checker_sizes;
+    let range_tuple_checker_sizes = original_program
+        .vm_config
+        .sdk
+        .rv32m
+        .unwrap()
+        .range_tuple_checker_sizes;
     let vm_config = VmConfig {
         instruction_handler: &airs,
         bus_interaction_handler: OpenVmBusInteractionHandler::new(
@@ -173,13 +162,9 @@ pub fn customize<'a, P: PgoAdapter<Adapter = BabyBearOpenVmApcAdapter<'a>>>(
         bus_map: bus_map.clone(),
     };
 
-    // Convert the jump destinations to u64 for compatibility with the `collect_basic_blocks` function.
-    let jumpdest_set = jumpdest_set
-        .iter()
-        .map(|&x| x as u64)
-        .collect::<BTreeSet<_>>();
-
-    let blocks = collect_basic_blocks::<BabyBearOpenVmApcAdapter>(&program, &jumpdest_set, &airs);
+    let blocks = original_program.collect_basic_blocks(config.degree_bound.identities);
+    let exe = original_program.exe;
+    let debug_info = original_program.elf.debug_info();
     tracing::info!(
         "Got {} basic blocks from `collect_basic_blocks`",
         blocks.len()
@@ -264,36 +249,6 @@ pub fn customize<'a, P: PgoAdapter<Adapter = BabyBearOpenVmApcAdapter<'a>>>(
     }
 }
 
-/// Besides the base RISCV-V branching instructions, the bigint extension adds two more branching
-/// instruction classes over BranchEqual and BranchLessThan.
-/// Those instructions have the form <INSTR rs0 rs1 target_offset ...>, where target_offset is the
-/// relative jump we're interested in.
-/// This means that for a given program address A containing the instruction above,
-/// we add A + target_offset as a target as well.
-fn add_extra_targets<F: PrimeField32>(
-    program: &OpenVmProgram<F>,
-    mut labels: BTreeSet<u32>,
-    base_pc: u32,
-    pc_step: u32,
-) -> BTreeSet<u32> {
-    let branch_opcodes_bigint = branch_opcodes_bigint_set();
-    let new_labels = program
-        .instructions_and_debug_infos
-        .iter()
-        .enumerate()
-        .filter_map(|(i, instr)| {
-            let instr = instr.as_ref().unwrap().0.clone();
-            let adjusted_pc = base_pc + (i as u32) * pc_step;
-            let op = instr.opcode;
-            branch_opcodes_bigint
-                .contains(&op)
-                .then_some(adjusted_pc + instr.c.as_canonical_u32())
-        });
-    labels.extend(new_labels);
-
-    labels
-}
-
 pub fn openvm_bus_interaction_to_powdr<F: PrimeField32>(
     interaction: &SymbolicInteraction<F>,
     columns: &[Arc<String>],