perf(evm): depth-indexed InterpreterExecContext pool for nested calls#482
Merged
zoowii merged 1 commit intoApr 29, 2026
Merged
Conversation
starwarfan
reviewed
Apr 28, 2026
Contributor
There was a problem hiding this comment.
Pull request overview
This PR significantly changes the EVM execution hot path to reduce interpreter overhead and introduce a profile-guided, background JIT compilation flow, while also adding a depth-indexed InterpreterExecContext pool to avoid per-nested-call allocations.
Changes:
- Reuse
InterpreterExecContextper call depth to eliminate repeated large frame-stack allocations in deeply nested calls. - Add profile-guided JIT triggering with a sliding-window profiler and a background compilation thread pool, plus new config/CLI options.
- Inline several “pure”/hot EVM opcodes in the interpreter dispatch loop to reduce handler overhead.
Reviewed changes
Copilot reviewed 12 out of 12 changed files in this pull request and generated 7 comments.
Show a summary per file
| File | Description |
|---|---|
| src/vm/dt_evmc_vm.cpp | Adds depth-indexed exec-context reuse and implements profile-guided JIT + background compilation plumbing in execute() |
| src/tests/spec_unit_tests.cpp | Adds --enable-profile-guided-jit CLI flag for the spec test runner |
| src/tests/solidity_contract_tests.cpp | Adds --enable-profile-guided-jit flag but removes --enable-multipass-lazy |
| src/runtime/runtime.cpp | Switches between interpreter/JIT based on whether JIT code is actually available |
| src/runtime/evm_module.h | Makes JIT code pointer atomic and adds a std::future for background compilation |
| src/runtime/evm_module.cpp | Waits for any in-flight background JIT compilation in EVMModule destructor; skips eager compile when PGJ is enabled |
| src/runtime/config.h | Adds EnableProfileGuidedJIT and NumJITCompileThreads runtime config fields |
| src/evm/interpreter.cpp | Inlines several opcode implementations inside the dispatch loop (calldata/txcontext/memory/misc/transient storage) |
| src/compiler/evm_frontend/evm_mir_compiler.cpp | Minor comment tweak in MIR builder init |
| src/compiler/evm_compiler.cpp | Wraps eager EVM JIT compile in try/catch and changes how JIT code pointer is published |
| src/cli/dtvm.cpp | Adds --enable-profile-guided-jit to the CLI |
| src/action/compiler.cpp | Removes the lazy-compilation warning branch and always runs eager EVM JIT compile in multipass mode |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
⚡ Performance Regression Check Results✅ Performance Check Passed (interpreter)Performance Benchmark Results (threshold: 25%)
Summary: 194 benchmarks, 0 regressions ✅ Performance Check Passed (multipass)Performance Benchmark Results (threshold: 25%)
Summary: 194 benchmarks, 0 regressions |
ys8888john
force-pushed
the
feat/perf-fold-stack-check-into-chunk
branch
from
47a7f32 to
99f4061
Compare
Contributor
There was a problem hiding this comment.
Pull request overview
This PR targets interpreter-mode EVM performance by reducing per-call allocations for nested calls and inlining several hot opcodes directly into the computed-goto dispatch loop.
Changes:
- Add a depth-indexed
InterpreterExecContextpool to reuse execution contexts across nested EVMC calls. - Inline multiple “pure read” opcodes and several memory/misc/transient-storage opcodes in
BaseInterpreter::interpret()to avoid handler overhead and reduce hot-path work.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| src/vm/dt_evmc_vm.cpp | Adds a per-depth InterpreterExecContext reuse pool for interpreter fast-path execution. |
| src/evm/interpreter.cpp | Inlines multiple opcode implementations inside the computed-goto interpreter loop for performance. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+1182
to
+1194
| // ---- Memory ops (inlined: MLOAD/MSTORE/MSTORE8) ---- | ||
| // Each opcode performs: | ||
| // 1. stack underflow check | ||
| // 2. memory expansion + gas charge (mirror of | ||
| // checkMemoryExpandAndChargeGas in opcode_handlers.cpp) | ||
| // 3. memcpy load/store | ||
| // The gas formula MUST match the canonical | ||
| // calculateMemoryExpansionCost EXACTLY: | ||
| // MemoryCost(W) = (W*W)/512 + 3*W (computed in __int128) | ||
| // delta = MemoryCost(NewWords) - MemoryCost(CurrentWords) | ||
| // NewWords/CurrentWords = ceil(size/32) where size is byte length. | ||
| // Any deviation (e.g. inlining the subtraction before the divide) | ||
| // can desynchronize gas accounting and break consensus. |
Contributor
Author
There was a problem hiding this comment.
Agree the duplication is a maintenance concern. I'll extract a static inline helper for the expansion+charge sequence in a follow-up.
Contributor
|
CI failed. @ys8888john |
ys8888john
force-pushed
the
feat/perf-fold-stack-check-into-chunk
branch
2 times, most recently
from
512ee4e to
66d72f1
Compare
zoowii
reviewed
Apr 28, 2026
| if (OffsetVal > intx::uint256(InputSize)) { | ||
| Frame->Stack[sp - 1] = intx::uint256(0); | ||
| } else { | ||
| // OffsetVal <= InputSize fits safely in size_t. |
Contributor
There was a problem hiding this comment.
only inline small functions (keep it easy to read by human).
the code in else branch is too large for inline
Contributor
Author
There was a problem hiding this comment.
This inline policy is inspired by evmone baseline, but for readability larger ops are still dispatched through *Handler::doExecute(), e.g. RETURN, RETURNDATACOPY, etc.
ys8888john
force-pushed
the
feat/perf-fold-stack-check-into-chunk
branch
from
66d72f1 to
cad7286
Compare
ys8888john
force-pushed
the
feat/perf-fold-stack-check-into-chunk
branch
from
cad7286 to
1a62057
Compare
abmcar
added a commit
to abmcar/DTVM
that referenced
this pull request
May 12, 2026
After rebasing onto current upstream/main (which now includes DTVMStack#458 / DTVMStack#460 / DTVMStack#482 / DTVMStack#483 perf work) and running a 10-rep evmone-bench on the 27 paper benches, the cumulative PR delta has collapsed to noise (raw geomean +1.15%, +0.46% after correcting a single-iteration outlier on main/blake2b_shifts/8415nulls via a focused 20-rep re-measurement). 0 benches above the +/-25% CI gate. The A-vs-PR-base -2.73% from this commit's own optimization is unchanged; the framing shift is that the absolute runtime delta of the whole PR vs unmodified main has been absorbed by the intervening upstream perf optimizations. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
6 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
1. Does this PR affect any open issues?(Y/N) and add issue references (e.g. "fix #123", "re #123".):
2. What is the scope of this PR (e.g. component or file name):
3. Provide a description of the PR(e.g. more details, effects, motivations or doc link):
4. Are there any breaking changes?(Y/N) and describe the breaking changes(e.g. more details, motivations or doc link):
5. Are there test cases for these changes?(Y/N) select and add more details, references or doc links:
6. Release note