perf(prover): per-table scheduler with VRAM admission for multi_prove

.gitignore

@@ -9,3 +9,7 @@ executor/program_artifacts/
 # Shared cargo target directory for ELF builds
 executor/shared_target/
 
+
+# Python bytecode
+__pycache__/
+*.pyc

bin/cli/Cargo.toml

@@ -20,3 +20,5 @@ env_logger = "0.11"
 jemalloc-stats = ["dep:tikv-jemalloc-ctl"]
 disk-spill = ["prover/disk-spill"]
 instruments = ["prover/instruments", "stark/instruments"]
+# GPU profiling build (Nsight): CUDA prover + instruments spans + NVTX ranges.
+nvtx = ["prover/nvtx", "instruments"]

crypto/crypto/src/merkle_tree/merkle.rs

@@ -172,6 +172,19 @@ where
     /// but no nodes. Used when paths are gathered from a device resident copy
     /// (GPU) instead of this host tree, so the host nodes are never built.
     /// [`get_proof_by_pos`](Self::get_proof_by_pos) must NOT be called on it.
+    /// True when this tree carries only its root (the nodes live elsewhere,
+    /// e.g. device-resident): openings must not walk this tree.
+    pub fn is_root_only(&self) -> bool {
+        #[cfg(feature = "disk-spill")]
+        {
+            self.nodes.is_empty() && self.mmap_backing.is_none()
+        }
+        #[cfg(not(feature = "disk-spill"))]
+        {
+            self.nodes.is_empty()
+        }
+    }
+
     pub fn from_root(root: B::Node) -> Self {
         MerkleTree {
             root,

crypto/math-cuda/Cargo.toml

@@ -30,11 +30,17 @@ cudarc = { version = "0.19", default-features = false, features = [
 ] }
 math = { path = "../math" }
 rayon = "1.7"
+# NVTX range emission for Nsight timelines (dlopen'd at runtime, see src/nvtx.rs).
+libloading = { version = "0.8", optional = true }
 
 [features]
 # Test-only fault injection in FriCommitState. Production builds leave this
 # off so the fault check is fully elided at compile time.
 test-faults = []
+# NVTX bindings for Nsight Systems profiling (ranges are emitted by the
+# stark/prover layers). Zero-cost when disabled; when enabled but
+# libnvToolsExt is absent at runtime, every call is a cheap no-op.
+nvtx = ["dep:libloading"]
 
 [dev-dependencies]
 crypto = { path = "../crypto" }

crypto/math-cuda/build.rs

@@ -82,6 +82,7 @@ fn compile_kernel(src: &str, out_name: &str, have_nvcc: bool) {
     println!("cargo:rerun-if-env-changed=CUDA_HOME");
     println!("cargo:rerun-if-env-changed=CUDA_PATH");
     println!("cargo:rerun-if-env-changed=CUDARC_NVCC_ARCH");
+    println!("cargo:rerun-if-env-changed=LAMBDA_VM_NVCC_LINEINFO");
 
     // When nvcc is missing from PATH, emit an empty cubin stub so the crate
     // still compiles. include_bytes! in src/device.rs needs the file to exist
@@ -115,8 +116,15 @@ fn compile_kernel(src: &str, out_name: &str, have_nvcc: bool) {
         .map(|a| to_real_arch(&a))
         .unwrap_or_else(|_| detect_arch());
 
-    let status = Command::new(nvcc_path())
-        .args(["--cubin", "-O3", "-std=c++17", "-arch", &arch, "-o"])
+    let mut cmd = Command::new(nvcc_path());
+    cmd.args(["--cubin", "-O3", "-std=c++17", "-arch", &arch]);
+    // SASS→source line mapping for Nsight Compute. Unlike -G this does not
+    // change codegen, but keep it opt-in so production cubins stay byte-stable.
+    if env::var("LAMBDA_VM_NVCC_LINEINFO").is_ok_and(|v| v != "0" && !v.is_empty()) {
+        cmd.arg("-lineinfo");
+    }
+    let status = cmd
+        .arg("-o")
         .arg(&out_path)
         .arg(&src_path)
         .status()