AutoGineer
Generated banner for Warp Compile Time Optimizer
Generated logo for Warp Compile Time Optimizer

Warp Compile Time Optimizer

NVIDIAOfficialCreated by Warp TeamDistributed by NVIDIA

  • Skill
  • Workflow
  • Software development
  • Apache-2.0
  • Medium risk
  • Has scripts
Download ZIP

From the author

Use when compile time or startup time is the problem in code that uses Warp: a request to improve, optimize, or cut compile times; an app that is slow to start or stalls at the first wp.launch; seconds of compiling before real work begins; JIT modules recompiling on every run or every CI job. Only applies when the code being optimized uses Warp kernels. Not for steady-state kernel runtime, memory, correctness, building Warp itself from source, or nvcc/C++ build times.
— the description by Warp Team

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill provides a structured approach to identifying and resolving excessive compile times or startup delays in Warp-based applications. It focuses on isolating compilation bottlenecks, detecting redundant module builds, and optimizing module configuration to ensure efficient kernel generation. The process relies on a dedicated probe script to measure cold-start performance and identify specific mechanisms causing compilation churn, such as identity changes, redundant module options, or inefficient module boundaries.

What it does

  • Measures cold and warm compilation times using an isolated probe script.
  • Identifies redundant module builds caused by identity churn or configuration changes.
  • Provides guidance on aligning module lifecycles and options to prevent unnecessary recompilation.
  • Helps distinguish between structural compilation issues and runtime performance bottlenecks.
  • Offers a protocol for verifying optimizations through comparative analysis of launch topology and timing.
  • Supplies a framework for reporting measured improvements and trade-offs.

When to use it

  • When an application experiences significant delays before the first kernel launch.
  • When Warp modules are observed to recompile repeatedly across runs or CI jobs.
  • When optimizing startup latency for production environments.
  • When investigating why specific kernels or modules trigger unexpected build activity.

Advantages

  • Provides objective, data-driven insights into compilation costs.
  • Minimizes noise by isolating cache directories during measurement.
  • Offers clear, mechanism-based diagnostics to target specific build inefficiencies.
  • Ensures that optimizations preserve existing workload behavior and numerical accuracy.

Usage examples

“My application stalls for several seconds at the first wp.launch. How can I find out why?”
The agent will guide you to use the warp_compile_probe.py script to measure your command, identify which modules are recompiling, and check for identity churn or configuration issues.
“I suspect my Warp kernels are recompiling every time I run my tests. How do I fix this?”
The agent will help you analyze the probe output to determine if the cache is being reused correctly or if module options are being set after the module has already loaded.

Install

  1. Download the skill as a ZIP:

    https://mp-staging.autogineer.com/agent-skills/nvidia/warp-compile-time-optimizer/download

  2. Unzip it into the cross-client skills folder of your project:

    mkdir -p .agents/skills && unzip warp-compile-time-optimizer.zip -d .agents/skills

Documentation by Warp Team

Warp cold-start compile time

Start with a runnable command

The probe runs the target in a subprocess, writes only to temporary directories, and needs no network, external tool servers, or Warp checkout.

ScriptPurposeArguments
scripts/warp_compile_probe.pyMeasure isolated cold/warm compilation and launches.measure [OPTIONS] -- COMMAND...; use --help.

Use run_script("scripts/warp_compile_probe.py", args=[...]) when supported; otherwise use the Python command below. The target must run to completion.

Compilation model

Warp compiles modules, not individual kernels. A module's identity is:

(live kernel & function set) x (module options) x (CUDA block_dim) x (generic instances)

Each identity requires code generation and native compilation for the full module.

Cold-start cost is roughly:

number of distinct module identities you touch  x  size of each module

Reduce it in two ways:

  1. Stop identity churn. A module that changes after loading compiles again.

  2. Stop source duplication. A module used at three block dimensions compiles every kernel three times.

Deleting one kernel from a module that still builds saves only part of one build. Removing an unnecessary module identity saves the full build.

When neither applies, overlap independent CUDA builds (CS-13). This changes when work happens, not how much is compiled, so judge it on elapsed time.

When a module's options are fixed

Options have two deadlines:

  1. Module creation (Python import). The module copies enable_backward, max_unroll, lineinfo, deterministic, deterministic_max_records, and compile_time_trace from warp.config. Setting a global later is silently ignored by that module. default_grid_stride is the exception.

  2. First load. Before compilation, change the existing module with wp.set_module_options() or wp.get_module(name).options. Changing an option after load creates a new identity and rebuilds the module (CS-3).

Missed deadlineSymptomCost
wp.config.* set after importhash unchanged, option silently absentthe entire benefit, invisibly
module options set after loada second hash, module builds twiceone extra build, visible in the trace

After changing an option, confirm the hash moved for every target module. An unchanged hash means the option never arrived.

Preserve behavior

Change how Warp compiles the code, not the workload.

Do not delete or merge kernels to claim a gain. Apparently redundant stages may preserve ownership, aliasing, retained outputs, numerical boundaries, or API behavior. Fix duplication at the module level.

Preserve every launch and its order, dimensions, dtypes, devices, block dimensions, gradients, numerical modes, dynamic/plugin behavior, and public API signatures. Keep kernel names when moving definitions to module scope because logs, cache artifacts, and external tools expose them.

Instructions

1. Find the real cost, and confirm it is compilation

Ask what command the user actually waits on, then measure it cold:

python scripts/warp_compile_probe.py measure --samples 3 \
    --json baseline.json -- <the user's command>

The probe gives each sample private WARP_CACHE_PATH, WARP_CACHE_ROOT, and CUDA_CACHE_PATH directories, enables module timers, and records launches. It creates those directories under the system temporary directory and removes them itself, so isolating a sample never requires writing a cache into the project or deleting anything to re-measure. Isolate a hand-rolled sample the same way. Never clear a live cache with wp.clear_kernel_cache() or wp.clear_lto_cache(); clearing is not isolated and can disrupt other processes.

Read the probe output before source. If compilation is a small part of wall time, report the real bottleneck and stop. For libraries and tests, use the smallest command that compiles the workload's modules.

Modules that each compiled once, with no repeated hashes, block-dimension variants, or LTO, have no structural churn. This rules out redundant builds, not oversized builds; still check cache reuse (CS-2), backward codegen (CS-10), unrolling (CS-11), the precompiled header (CS-12), and overlap when several CUDA modules remain (CS-13).

Every sample also re-runs the command against the cache it just populated. If warm module work is not near zero, diagnose cache reuse (CS-2) before changing module structure.

2. Ask about runtime tradeoffs when needed

Apply ordering fixes, lifecycle grouping, and option hoists without asking. Ask before changing fast_math, max_unroll, or a MathDx/tile implementation:

Some of these knobs cut compile time but can make the compiled kernels slower or change numerics. Are you optimizing a fast edit-run loop (where slower kernels are usually fine), or production startup (where they usually are not)?

If the user is unavailable:

  • Leave numerics-changing options and implementation swaps alone.

  • Before removing a capability such as backward codegen, test whether it is used; report what is removed, the measured benefit, and how to revert.

  • Set global wp.config.* options at application entry points, not in library code.

Record declined options and their measured benefits in the step 6 ledger.

Match the scope of the change to the scope of the evidence

A profile supports a change to the measured application, not every consumer of a shared library. Repository searches also miss out-of-tree and future callers. For example, a forward-only application does not justify disabling gradients inside a solver library that another application differentiates through.

Scope the option to the measured process, before importing the library:

import warp as wp

wp.config.enable_backward = False   # must precede the library import

import the_library

This also reaches every module the application loads. CS-10 covers the silent import-order trap. If only a library change works, send its maintainers the measurement and let them decide the contract.

3. Diagnose from the measurement, not from reading the source

The probe prints every compiled module identity with its name, hash, device, and block dimension, then names which modules built more than once. Match what you see:

What the probe showsWhat it meansWhere to look
One module name, several hashesIdentity churn: its kernel set, options, or generic instances changed after it first loadedCS-1, CS-3, CS-6
One module name, several block_dim valuesThe whole module is recompiled per block dimension (CUDA)CS-5
Many one-kernel modules in one featureFixed per-module cost repeatedCS-4
A hash-named module per kernelmodule="unique" used on stable kernelsCS-9
Big gap between module time and native compile time, plus .lto artifactsMathDx/LTO setupCS-7
(compiled) on a run that should have been warmCache is not being reusedCS-2
Modules load, then "Failed to find module"Concurrent CPU JIT first-use raceCS-8
Large generated source, no rebuild problemUnroll budgetCS-11
Adjoint code in a module nothing differentiatesBackward codegenCS-10
Compiles slow across the board, or a few small modules on CUDA below toolkit 13The precompiled header is turned off, or is not paying for itselfCS-12
Several independent modules, each built once, overlap_factor near 1.0Builds are running one at a time; parallel loading is off by defaultCS-13
An option you set changed nothing, and that module's hash is unchangedIt was assigned after the module was created, so it never arrived"When a module's options are fixed"
No row above firesNothing is being built redundantly; the cost is the size of the builds themselvesStep 6

references/mechanisms.md has one section per mechanism: how to confirm it, the fix, its limits, and its failure mode. Read only the sections selected by the measurement.

4. Choose module boundaries deliberately

Group kernels in one module only when they share:

  • lifecycle: they are defined, loaded, and invalidated together;

  • option set: they need the same fast_math, enable_backward, max_unroll, and MathDx settings;

  • stable block dimension on CUDA.

Kernels with the same lifecycle but different stable block dimensions should not share a module because each would compile twice. Separate kernels with independent lifecycles too.

Kernels whose block dimension varies at runtime (chosen from input size, say) have no stable mapping, so keep them in their own module rather than dragging a whole shared module into an extra variant.

Prefer the least invasive change that removes a build. Ordering fixes and option hoists are cheaper and safer than re-architecting module layout; regroup only when fixed per-module cost or block-dimension duplication dominates.

wp.set_module_options() targets its calling Python module, not kernels with an explicit module="pkg.name". Either use a real Python module or update the named module before it loads:

wp.set_module_options({"enable_backward": False})  # at module scope

wp.get_module("pkg.name").options.update({"enable_backward": False})

Do not pass wp.get_module() to wp.set_module_options(module=...), or use @wp.kernel(module_options={...}) without module="unique". Per-kernel enable_backward=False has a tile-module exception covered by CS-10. Confirm the module hash after every option change.

5. Verify
python scripts/warp_compile_probe.py measure --samples 3 \
    --json candidate.json -- <the same command>
python scripts/warp_compile_probe.py compare baseline.json candidate.json

compare rejects changed launch topology and treats a result inside max(1% of baseline, 2 x baseline MAD) as inconclusive.

For BUILDS OVERLAPPED, judge scheduling changes on compile elapsed rather than summed module timers. The required warm pass supplies that clock. See references/measurement.md.

Then check what the probe cannot see:

  • Diff numeric output and run the project's tests or entry points.

  • After changing enable_backward or boundaries, verify a gradient path.

  • After changing fast_math, max_unroll, MathDx, or an implementation, benchmark steady-state runtime.

  • Exercise uncovered dynamic kernels, dtypes, profiles, and modes.

6. Report results

Read "Reporting results" in references/measurement.md. Report:

Describe every optimization in plain language: name the behavior, the evidence, and the effect. For example, write "moved module options before the first load to avoid a redundant rebuild," not "applied CS-3." Treat CS-* labels as internal navigation aids, not user-facing explanations.

OptionMeasuredWhy not takenTo take it
enable_backward=False on pkg.solver−38% colda live tape traverses these kernelsset at the entry point, then re-check adjoints
max_unroll=4−2%, inside noisechanges generated code for no measured gain—

  • before/after medians and sample counts;

  • the reduction and residual cost, sized against the original complaint;

  • mechanisms fixed, ruled out, measured and declined, or not reached;

  • tradeoffs and behavior not verified;

  • the ledger above for declined or incompletely verified options;

  • the measured value of relaxing any constraint that blocked a fix.

State only what the evidence supports. "No structural churn" does not mean "optimal" or "irreducible." Measure declined levers when practical; label any estimate untested. Before reporting no available fix, check CS-13. For a possible module split, first measure a one-kernel module with the same options to establish the repeated fixed cost.

Troubleshooting

Run the target command directly before debugging the probe. See references/measurement.md for cache/noise issues and references/mechanisms.md for mechanism-specific failures.

Limitations

Two rules override any gain:

  • Isolate both Warp and CUDA caches for every cold sample.

  • Keep max_workers <= 1 when a load can target CPU, including device=None and mixed device lists. Concurrent CPU first loads can lose kernels; retries do not make them safe. CUDA-only loading is unaffected.

Measurements are environment-specific: cold times move with CPU, GPU, driver, toolchain, and Warp version. Mechanisms transfer; numbers do not. Read the known unknowns in references/mechanisms.md before making broad claims.

Reference files

  • references/mechanisms.md: the thirteen compile-time mechanisms, each with its confirming signal, fix, applicability limits, and failure mode. Read the sections your measurement points to.

  • references/measurement.md: measurement protocol, what each metric does and does not mean, reporting guidance, log examples, and manual measurement.

Files

LICENSE

Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.


                                 Apache License
                           Version 2.0, January 2004
                        http://www.apache.org/licenses/

   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

   1. Definitions.

      "License" shall mean the terms and conditions for use, reproduction,
      and distribution as defined by Sections 1 through 9 of this document.

      "Licensor" shall mean the copyright owner or entity authorized by
      the copyright owner that is granting the License.

      "Legal Entity" shall mean the union of the acting entity and all
      other entities that control, are controlled by, or are under common
      control with that entity. For the purposes of this definition,
      "control" means (i) the power, direct or indirect, to cause the
      direction or management of such entity, whether by contract or
      otherwise, or (ii) ownership of fifty percent (50%) or more of the
      outstanding shares, or (iii) beneficial ownership of such entity.

      "You" (or "Your") shall mean an individual or Legal Entity
      exercising permissions granted by this License.

      "Source" form shall mean the preferred form for making modifications,
      including but not limited to software source code, documentation
      source, and configuration files.

      "Object" form shall mean any form resulting from mechanical
      transformation or translation of a Source form, including but
      not limited to compiled object code, generated documentation,
      and conversions to other media types.

      "Work" shall mean the work of authorship, whether in Source or
      Object form, made available under the License, as indicated by a
      copyright notice that is included in or attached to the work
      (an example is provided in the Appendix below).

      "Derivative Works" shall mean any work, whether in Source or Object
      form, that is based on (or derived from) the Work and for which the
      editorial revisions, annotations, elaborations, or other modifications
      represent, as a whole, an original work of authorship. For the purposes
      of this License, Derivative Works shall not include works that remain
      separable from, or merely link (or bind by name) to the interfaces of,
      the Work and Derivative Works thereof.

      "Contribution" shall mean any work of authorship, including
      the original version of the Work and any modifications or additions
      to that Work or Derivative Works thereof, that is intentionally
      submitted to Licensor for inclusion in the Work by the copyright owner
      or by an individual or Legal Entity authorized to submit on behalf of
      the copyright owner. For the purposes of this definition, "submitted"
      means any form of electronic, verbal, or written communication sent
      to the Licensor or its representatives, including but not limited to
      communication on electronic mailing lists, source code control systems,
      and issue tracking systems that are managed by, or on behalf of, the
      Licensor for the purpose of discussing and improving the Work, but
      excluding communication that is conspicuously marked or otherwise
      designated in writing by the copyright owner as "Not a Contribution."

      "Contributor" shall mean Licensor and any individual or Legal Entity
      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,
      publicly display, publicly perform, sublicense, and distribute the
      Work and such Derivative Works in Source or Object form.

   3. Grant of Patent License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      (except as stated in this section) patent license to make, have made,
      use, offer to sell, sell, import, and otherwise transfer the Work,
      where such license applies only to those patent claims licensable
      by such Contributor that are necessarily infringed by their
      Contribution(s) alone or by combination of their Contribution(s)
      with the Work to which such Contribution(s) was submitted. If You
      institute patent litigation against any entity (including a
      cross-claim or counterclaim in a lawsuit) alleging that the Work
      or a Contribution incorporated within the Work constitutes direct
      or contributory patent infringement, then any patent licenses
      granted to You under this License for that Work shall terminate
      as of the date such litigation is filed.

   4. Redistribution. You may reproduce and distribute copies of the
      Work or Derivative Works thereof in any medium, with or without
      modifications, and in Source or Object form, provided that You
      meet the following conditions:

      (a) You must give any other recipients of the Work or
          Derivative Works a copy of this License; and

      (b) You must cause any modified files to carry prominent notices
          stating that You changed the files; and

      (c) You must retain, in the Source form of any Derivative Works
          that You distribute, all copyright, patent, trademark, and
          attribution notices from the Source form of the Work,
          excluding those notices that do not pertain to any part of
          the Derivative Works; and

      (d) If the Work includes a "NOTICE" text file as part of its
          distribution, then any Derivative Works that You distribute must
          include a readable copy of the attribution notices contained
          within such NOTICE file, excluding those notices that do not
          pertain to any part of the Derivative Works, in at least one
          of the following places: within a NOTICE text file distributed
          as part of the Derivative Works; within the Source form or
          documentation, if provided along with the Derivative Works; or,
          within a display generated by the Derivative Works, if and
          wherever such third-party notices normally appear. The contents
          of the NOTICE file are for informational purposes only and
          do not modify the License. You may add Your own attribution
          notices within Derivative Works that You distribute, alongside
          or as an addendum to the NOTICE text from the Work, provided
          that such additional attribution notices cannot be construed
          as modifying the License.

      You may add Your own copyright statement to Your modifications and
      may provide additional or different license terms and conditions
      for use, reproduction, or distribution of Your modifications, or
      for any such Derivative Works as a whole, provided Your use,
      reproduction, and distribution of the Work otherwise complies with
      the conditions stated in this License.

   5. Submission of Contributions. Unless You explicitly state otherwise,
      any Contribution intentionally submitted for inclusion in the Work
      by You to the Licensor shall be under the terms and conditions of
      this License, without any additional terms or conditions.
      Notwithstanding the above, nothing herein shall supersede or modify
      the terms of any separate license agreement you may have executed
      with Licensor regarding such Contributions.

   6. Trademarks. This License does not grant permission to use the trade
      names, trademarks, service marks, or product names of the Licensor,
      except as required for reasonable and customary use in describing the
      origin of the Work and reproducing the content of the NOTICE file.

   7. Disclaimer of Warranty. Unless required by applicable law or
      agreed to in writing, Licensor provides the Work (and each
      Contributor provides its Contributions) on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
      implied, including, without limitation, any warranties or conditions
      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
      PARTICULAR PURPOSE. You are solely responsible for determining the
      appropriateness of using or redistributing the Work and assume any
      risks associated with Your exercise of permissions under this License.

   8. Limitation of Liability. In no event and under no legal theory,
      whether in tort (including negligence), contract, or otherwise,
      unless required by applicable law (such as deliberate and grossly
      negligent acts) or agreed to in writing, shall any Contributor be
      liable to You for damages, including any direct, indirect, special,
      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

   9. Accepting Warranty or Additional Liability. While redistributing
      the Work or Derivative Works thereof, You may choose to offer,
      and charge a fee for, acceptance of support, warranty, indemnity,
      or other liability obligations and/or rights consistent with this
      License. However, in accepting such obligations, You may act only
      on Your own behalf and on Your sole responsibility, not on behalf
      of any other Contributor, and only if You agree to indemnify,
      defend, and hold each Contributor harmless for any liability
      incurred by, or claims asserted against, such Contributor by reason
      of your accepting any such warranty or additional liability.

   END OF TERMS AND CONDITIONS

   APPENDIX: How to apply the Apache License to your work.

      To apply the Apache License to your work, attach the following
      boilerplate notice, with the fields enclosed by brackets "[]"
      replaced with your own identifying information. (Don't include
      the brackets!)  The text should be enclosed in the appropriate
      comment syntax for the file format. We also recommend that a
      file or class name and description of purpose be included on the
      same "printed page" as the copyright notice for easier
      identification within third-party archives.

   Copyright [yyyy] [name of copyright owner]

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

       http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Declares allowed tools, Needs Python

The author says: Requires Python 3.10+ and an installed warp-lang package. A CUDA device is needed to diagnose CUDA-specific mechanisms.

Security

What our scanner noticed· Medium risk

  • Large base64-encoded blobs
  • Scripts that your agent may run

This skill includes scripts that your agent may run. Read them before you use it.

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by
Warp Team
Publisher
NVIDIA (Official)
Distributed by
NVIDIA
Source
NVIDIA/skills · skills/warp-compile-time-optimizer · commit fce3cca
Licence
Apache License 2.0 (redistribution allowed)
  • LICENSE (added by AutoGineer from LICENSE-APACHE)
  • ATTRIBUTION.md (written by AutoGineer)
Synced

AutoGineer is not affiliated with or endorsed by Warp Team or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download warp-compile-time-optimizer.zip