Hypertile Logo

Hypertile

A unified, work-stealing executor collocating free-threaded Python coroutines and Send Rust futures in one right-sized thread pool.

Rust Python Type Checked uv License

1. Executive Summary & Problem Space

Modern high-performance applications combining Python and Rust (such as FastAPI web services, AI/ML inference servers, and distributed ETL pipelines) conventionally run two completely independent async runtime stacks:

  1. Python's Async Stack: Single-threaded asyncio event loop driving coroutines, combined with a separate ThreadPoolExecutor (allocating 8–32 OS threads) for offloading blocking work.
  2. Rust's Async Stack: A native multi-threaded runtime (Tokio-style) with its own thread pool sized to available CPU cores.

The Hidden Bottlenecks of Dual Pools:

  • 💥 Hardware Oversubscription: Two independent pools sized against hardware cores create $2 \times \text{CPUs}$ active OS threads, resulting in relentless thread parking, context switching, and cache-line invalidation.
  • 🐢 Double-Hop Boundary Latency: Crossing between Python and Rust requires two scheduling hops: $$\text{Native Wake} \longrightarrow \text{Native Tick} \longrightarrow \texttt{call_soon_threadsafe} \longrightarrow \texttt{eventfd} \longrightarrow \text{Asyncio Tick} \longrightarrow \text{Python Task}$$
  • 🔒 Zero Capacity Sharing: Idle interpreter threads cannot assist with stranded native CPU work, and idle native threads cannot step Python coroutines.
  • 🛑 GIL Overhead vs. Free-Threaded Promise: Standard CPython serializes bytecode execution under the Global Interpreter Lock (GIL). However, with PEP 779 free-threaded Python (3.13t / 3.14t+), Python bytecode can execute truly in parallel across multiple OS threads—if and only if the executor is designed to step coroutines natively without lock contention.

2. Architecture & Core Innovations

Hypertile provides one unified multi-threaded runtime whose workers are individually either native-only (pure Rust) or bilingual (interpreter-attached):

                     +-----------------------------------------------+
                     |            HYPERTILE GLOBAL POOL              |
                     |   shared injector (MPSC) + per-worker local   |
                     |   Chase-Lev deques (crossbeam-deque)          |
                     +-----------------------------------------------+
                           |                   |                  |
             +-------------+---------+  +------+------+   +------+---------+
             |  Bilingual worker    |  |  Native     |   |  Bilingual     |
             |  (interpreter-attach |  |  worker     |   |  worker        |
             |   on 3.13t/3.14t+)   |  |  (Rust-only)|   |  (… N workers) |
             +----------------------+  +-------------+   +----------------+
                  |          |                |
        +---------+          |                +------------------+
        | Step Python        |                |  Poll Rust       |
        | coroutines, in     |                |  futures (any    |
        | batches, via       |                |  Send future,    |
        | coro.send()        |                |  never two       |
        | (interpreter       |                |  threads at once)|
        |  mode)             |                |                  |
        +--------------------+                +------------------+

Architectural Highlights:

  1. Single-Hop Continuation Handoff: When a native task finishes on worker $W$, $W$ pushes the awaiting continuation directly onto its own local Chase-Lev deque. $W$ resumes it immediately without returning to an event loop or waking another thread (~1.01 µs per hop).
  2. Cache-Line False-Sharing Elimination (CachePadded): Hot atomics (idle_count) and worker locks (idle_stack, stealers_cache) are isolated via crossbeam_utils::CachePadded, eliminating MESI/MOESI cache-line bouncing across 64-byte (x86/ARM) and 128-byte (Apple Silicon M-series/POWER) lines.
  3. Sub-Nanosecond PRNG (fast_rand): Replaced thread-local cryptographic RNG calls with an ultra-fast non-cryptographic SplitMix64 PRNG, executing victim selection in 2–3 single-cycle ALU instructions.
  4. Inter-Core Batch Work-Stealing (steal_batch_and_pop): When stealing from the global injector or peer workers, an idle thread atomically claims half of the victim's queue in a single transaction, executing remaining tasks locally out of L1 cache with zero contention.
  5. Hardware-Adaptive Vectorized Batching: Vectorized APIs (batch_native_pipeline and gather_to_thread) cross the Python $\leftrightarrow$ Rust FFI boundary once per slice and dynamically chunk workloads across physical cores, yielding >237,000 to >697,000 operations/sec.
  6. Dynamic Worker Registration & Capacity Sharing: External server threads (e.g. FastAPI / Uvicorn workers) dynamically join Hypertile's work-stealing pool as auxiliary workers while idle and leave cleanly in 2.76 µs / cycle.
  7. Panic Containment: Native panics are caught via catch_unwind and surfaced as hypertile.PanicInTask without poisoning mutexes or killing worker threads.

3. GIL vs. Free-Threaded Support (Honest Matrix)

Capability Free-Threaded (3.13t / 3.14t+, PEP 779) Standard GIL Builds (3.11–3.14)
Parallel Python Bytecode Execution Yes (bilingual workers step simultaneously) No (bytecode requires GIL)
Native Task Work-Stealing Yes (all workers) Yes (all workers)
Single-Hop Continuation Handoff Yes (on any bilingual worker) Yes (on event loop thread)
Vectorized Micro-Batching Yes (>237,000 req/s) Yes (>226,000 req/s)
Dynamic Worker Registration Yes (2.76 µs / cycle) Yes (2.76 µs / cycle)
Operating Mode Full Native Work-Stealing Cooperative Mode

4. API Reference & Developer Ergonomics

Asynchronous Offloading: hypertile.to_thread

Ultra-low-latency drop-in replacement for asyncio.to_thread(). Dispatches callables directly to Hypertile's bilingual workers without allocating OS threads or intermediate asyncio futures:

import hypertile

# Dispatch arbitrary Python callable
result = await hypertile.to_thread(crypto_hash, payload, rounds=50)

Vectorized Parallel Dispatch: hypertile.gather_to_thread

Vectorized parallel execution across an iterable of inputs. Crosses the FFI boundary once for the entire batch:

# Process a collection of items in parallel across bilingual workers
results = await hypertile.gather_to_thread(process_record, [r1, r2, r3, r4])

Function Decorator: @hypertile.task

Converts synchronous functions into awaitable Hypertile tasks:

@hypertile.task
def verify_token(raw_jwt: str) -> dict:
    return jwt.decode(raw_jwt, key, algorithms=["RS256"])

# Inside an async endpoint:
claims = await verify_token(header_auth)

Native Pipelines: hypertile.spawn_native_pipeline

Route CPU-intensive numerical and cryptographic transforms directly onto the native Rust work-stealing queue with single-hop continuation:

native_task = hypertile.spawn_native_pipeline(raw_bytes, rounds=100)
digest = await native_task

High-Throughput Batch Pipeline: hypertile.batch_native_pipeline

Submits a batch of payloads directly into the native work-stealing engine in a single FFI crossing:

# Dispatches thousands of payloads with sub-5µs amortized latency
digests = await hypertile.batch_native_pipeline(payload_chunks, rounds=100)

Dynamic Worker Registration

Join the work-stealing pool from long-running server threads (e.g., FastAPI lifespan):

with hypertile.register_worker(kind="bilingual") as worker:
    # Assist the executor while waiting for requests
    worker.run_until_idle()

Cooperative Cancellation: CancellationToken

Propagate cooperative cancellation across bilingual and native workers:

token = hypertile.CancellationToken()
token.cancel()
assert token.is_cancelled()

5. Benchmarks & Empirical Performance

All benchmarks were measured on Windows AMD64 (8 physical cores / 16 threads) comparing identical cryptographic/numerical workloads:

A. Free-Threaded (No-GIL, Python 3.13t) Benchmark:

.venv-313t/Scripts/python showcase/free_threaded_showcase.py
Metric Standard ThreadPool (Baseline) Hypertile Direct (Scalar) Hypertile Vector (Batch) Speedup vs Baseline
Throughput (req/s) 10,747 req/s 19,340 req/s 237,270 req/s 1.80x (scalar) / 22.1x (vector)
Wall Time (10,000 reqs) 0.930 s 0.517 s 0.042 s -44.4% (scalar) / -95.5% (vector)
Median Latency (p50) 17.54 ms 8.21 ms 4.21 µs / item -53.2% latency reduction
Tail Latency (p95) 24.08 ms 11.23 ms 4.21 µs / item -53.4% tail latency reduction
Tail Latency (p99) 74.17 ms 56.39 ms 4.21 µs / item -24.0% tail latency reduction

B. Standard GIL (Python 3.11) Honest Benchmark:

.venv/Scripts/python showcase/showcase_benchmark.py
Metric Standard asyncio (Baseline) Hypertile L2 Hook Hypertile Vector Batch Speedup vs Baseline
Cross-Thread Offload Latency 115.41 µs 113.22 µs N/A 1.02x faster offload
Pipeline Throughput 9,653 req/s 8,295 req/s 226,924 req/s 23.5x throughput multiplier
Median Latency (p50) 21.68 ms 25.62 ms 4.41 µs / item Sub-5µs per item
Dynamic Worker Cycle N/A 2.76 µs / cycle N/A Sub-3µs thread registration

6. Development with uv & Multi-Environment Setup

Hypertile strictly recommends Astral uv for fast, reproducible virtual environment management:

1. Free-Threaded Environment (Python 3.13t)

# Install free-threaded Python 3.13t
uv python install 3.13t

# Create virtual environment
uv venv --python 3.13t .venv-313t

# Install dependencies with uv
uv pip install maturin pytest fastapi httpx --python .venv-313t/Scripts/python.exe

# Build and install editable release wheel
$env:VIRTUAL_ENV = "d:\HyperTile\.venv-313t"
& .venv-313t\Scripts\maturin.exe develop --release --uv

2. Standard GIL Environment (Python 3.11)

# Create virtual environment
uv venv --python 3.11 .venv

# Install dependencies with uv
uv pip install maturin pytest fastapi httpx --python .venv/Scripts/python.exe

# Build and install editable release wheel
$env:VIRTUAL_ENV = "d:\HyperTile\.venv"
& .venv\Scripts\maturin.exe develop --release --uv

7. Production Examples

Complete, executable production examples are provided in examples/:

  1. FastAPI Microservice (examples/fastapi_service.py): Lifespan worker registration, @hypertile.task offloading, and native pipeline endpoints.

    python examples/fastapi_service.py

  2. Batch Data Pipeline (examples/data_pipeline.py): Multi-stage ETL pipeline, cooperative cancellation tokens, dynamic worker scaling, and vectorized native batches.

    python examples/data_pipeline.py

8. Multi-OS & Hardware Compatibility Matrix

Platform Architecture Tier Verification Status
Windows x86_64, aarch64 Tier 1 Verified with MSVC toolchain, WaitOnAddress, keyed events, PyO3 .pyd.
Linux x86_64, aarch64 Tier 1 POSIX threads, standard futexes, multi-OS GitHub Actions CI workflow enabled.
macOS x86_64, aarch64 (Apple Silicon) Tier 1 Pthread primitives, Mach monotonic timing, 128B Apple Silicon cache alignment.

9. License

Dual-licensed under either of:

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Newer release available (0.1.1)

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for Flaxmbot from gravatar.com Flaxmbot

Credits

Author: Hypertile Contributors

License

MIT OR Apache-2.0

Requires

Python >=3.11

Classifiers

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

hypertile-0.1.0.tar.gz (46.1 kB view details)

Uploaded Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

hypertile-0.1.0-cp314-cp314-win_amd64.whl (259.3 kB view details)

Uploaded CPython 3.14Windows x86-64

hypertile-0.1.0-cp314-cp314-macosx_11_0_arm64.whl (310.6 kB view details)

Uploaded CPython 3.14macOS 11.0+ ARM64

hypertile-0.1.0-cp314-cp314-macosx_10_12_x86_64.whl (321.3 kB view details)

Uploaded CPython 3.14macOS 10.12+ x86-64

hypertile-0.1.0-cp313-cp313-win_amd64.whl (259.3 kB view details)

Uploaded CPython 3.13Windows x86-64

hypertile-0.1.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (345.9 kB view details)

Uploaded CPython 3.13manylinux: glibc 2.17+ x86-64

hypertile-0.1.0-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (337.0 kB view details)

Uploaded CPython 3.13manylinux: glibc 2.17+ ARM64

hypertile-0.1.0-cp313-cp313-macosx_11_0_arm64.whl (310.6 kB view details)

Uploaded CPython 3.13macOS 11.0+ ARM64

hypertile-0.1.0-cp313-cp313-macosx_10_12_x86_64.whl (321.3 kB view details)

Uploaded CPython 3.13macOS 10.12+ x86-64

hypertile-0.1.0-cp312-cp312-win_amd64.whl (258.9 kB view details)

Uploaded CPython 3.12Windows x86-64

hypertile-0.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (345.8 kB view details)

Uploaded CPython 3.12manylinux: glibc 2.17+ x86-64

hypertile-0.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (336.9 kB view details)

Uploaded CPython 3.12manylinux: glibc 2.17+ ARM64

hypertile-0.1.0-cp312-cp312-macosx_11_0_arm64.whl (310.4 kB view details)

Uploaded CPython 3.12macOS 11.0+ ARM64

hypertile-0.1.0-cp312-cp312-macosx_10_12_x86_64.whl (321.0 kB view details)

Uploaded CPython 3.12macOS 10.12+ x86-64

hypertile-0.1.0-cp311-cp311-win_amd64.whl (257.7 kB view details)

Uploaded CPython 3.11Windows x86-64

hypertile-0.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (347.2 kB view details)

Uploaded CPython 3.11manylinux: glibc 2.17+ x86-64

hypertile-0.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (337.8 kB view details)

Uploaded CPython 3.11manylinux: glibc 2.17+ ARM64

hypertile-0.1.0-cp311-cp311-macosx_11_0_arm64.whl (312.2 kB view details)

Uploaded CPython 3.11macOS 11.0+ ARM64

hypertile-0.1.0-cp311-cp311-macosx_10_12_x86_64.whl (322.1 kB view details)

Uploaded CPython 3.11macOS 10.12+ x86-64

File details

Details for the file hypertile-0.1.0.tar.gz.

File metadata

  • Download URL: hypertile-0.1.0.tar.gz
  • Upload date:
  • Size: 46.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for hypertile-0.1.0.tar.gz
Algorithm Hash digest
SHA256 fc31aaf2cfa1df7772a400d3fd518db43c6688191d1778d531e9595ffc9d241f
MD5 57f7db46899b8dfc05eda877d7befae1
BLAKE2b-256 129ae42eb4973b3838937c5253e53050d5015d470b00a9b8025cccf593e89845

See more details on using hashes here.

File details

Details for the file hypertile-0.1.0-cp314-cp314-win_amd64.whl.

File metadata

  • Download URL: hypertile-0.1.0-cp314-cp314-win_amd64.whl
  • Upload date:
  • Size: 259.3 kB
  • Tags: CPython 3.14, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for hypertile-0.1.0-cp314-cp314-win_amd64.whl
Algorithm Hash digest
SHA256 78bd6cc1ffe4018dac1720201395b7747f7caf26a327419bb88324a3e54fd964
MD5 02059c3b68281f253259411bba15ada8
BLAKE2b-256 3816916eccca3ef458610a4ecf3d1e35239b1bf0ad45a6f85cdbb09c9ede5802

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp314-cp314-win_amd64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp314-cp314-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp314-cp314-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 2ad33292e0a7567230a4929384ab8d9c94560a9a8171da0361f0eedda7471eee
MD5 27c0b3bd0388d37563f4ae8e3cf46401
BLAKE2b-256 62bd603e44913db808a79cddbf297822e075b02c67a302a5154ba1a79bbcb72a

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp314-cp314-macosx_11_0_arm64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp314-cp314-macosx_10_12_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp314-cp314-macosx_10_12_x86_64.whl
Algorithm Hash digest
SHA256 7616f8640d1c7653cfdbcad77b6039ab05d0a1597f8a8d54ca8625fa2ce05752
MD5 d88b171479774262c7864dce8d26dd12
BLAKE2b-256 9fde67e46d677dee9ae1112284617f21d23ed660e8b1f95951465b2782a71157

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp314-cp314-macosx_10_12_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp313-cp313-win_amd64.whl.

File metadata

  • Download URL: hypertile-0.1.0-cp313-cp313-win_amd64.whl
  • Upload date:
  • Size: 259.3 kB
  • Tags: CPython 3.13, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for hypertile-0.1.0-cp313-cp313-win_amd64.whl
Algorithm Hash digest
SHA256 2d18567eda4e231ad5f4dd571be323350021478d1242f21538c70d0ff99282dd
MD5 5824059492ef1339838879032dbb1eab
BLAKE2b-256 b055a2c960b5a09cde15956e6ce47f2d2056089b6cd4e371a8c1be2b6cf18ae5

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp313-cp313-win_amd64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 9f0b29d3cebf8dc65b7e8d6930c18ec80ddcbb2f4190ea6818c24255ad654b8d
MD5 d87b1da1e1b5064db3e2ba4a0063b383
BLAKE2b-256 c9bb2e1cec43b981a28a880a36cbc61be9f34aee2a3fb21ce736e3005979ef46

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 1e46c503bc66213fc548675c3605319e578c397a2d67cde2239a20e6f6fff70d
MD5 23c475bc845a844339c7d45536944fa6
BLAKE2b-256 a5ecdc91714ed95a83d6a3612aac0736d2d5d6e90eff929e08be30fcee4aa6e5

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp313-cp313-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp313-cp313-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 016998209b1a0bb972769e6d32522a965f858a23fae156185784ff1ce57f7814
MD5 bde1bb8f8663c2c71601a61fd211e063
BLAKE2b-256 4887e1fe3afb71fa048dfea6b7fa09c7a07bfbe334b358100d35478af8a1a040

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp313-cp313-macosx_11_0_arm64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp313-cp313-macosx_10_12_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp313-cp313-macosx_10_12_x86_64.whl
Algorithm Hash digest
SHA256 f482a258cda66d2b3c4603b1032ae5b1491a00362b42d60a8bf0a0f1dd49c432
MD5 d1aaf6ae1a7aa321d2a546f7f20b1381
BLAKE2b-256 a15b8736887229d6aa54bff6928e7d0dd1a739b67bdd0e7801633010981a0f87

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp313-cp313-macosx_10_12_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp312-cp312-win_amd64.whl.

File metadata

  • Download URL: hypertile-0.1.0-cp312-cp312-win_amd64.whl
  • Upload date:
  • Size: 258.9 kB
  • Tags: CPython 3.12, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for hypertile-0.1.0-cp312-cp312-win_amd64.whl
Algorithm Hash digest
SHA256 b1550ac4f84f9e8db9c419ccfb5c49475a7c19d17e754542076b44ab072ca002
MD5 fb6a73a2bad665a324af32d3fe3ad9df
BLAKE2b-256 3a9dc6d02276c8261352debc675a4915d7a57fc4f2169656bbbbdbe2a0f66375

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp312-cp312-win_amd64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 e93d23c9b7cdf10e08949f87ea0bc50f4ad3944b5710c0b31186da7f87caf9f5
MD5 d9a4015973e475f6ebdf64b9ec1b6df4
BLAKE2b-256 797be13c8b8f639f754c64596bc6918500a47bae98902348f963dada250990c8

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 c79b1c46d5a056ebbc9cfac199904a7f6942ef4b90b960b003e53b6bc268fe87
MD5 fd287b3322c9952282b6c866652491e9
BLAKE2b-256 866683a9e58fab80b027418611bb0595657a9d758b119192e8aecd7a6f241ca0

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp312-cp312-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp312-cp312-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 24469b8284e87baedc83fbf29d85b81e79b5a91ef116b76af70e33ec8f3b7699
MD5 021a2413f98482fc0e556e6bc7d5b08e
BLAKE2b-256 14e206b1195cd17ddb30a14cfb14e97a664f53ca730fdfa484254a49e83aca26

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp312-cp312-macosx_11_0_arm64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp312-cp312-macosx_10_12_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp312-cp312-macosx_10_12_x86_64.whl
Algorithm Hash digest
SHA256 875144c2d34644b67c0978ba5f5e5e796c5ab41e41139ce598e5e39de70570e3
MD5 7d1e8c4b15091d32bf53803b360abad1
BLAKE2b-256 a2cebec8e338c385c86cb6a1b2d9be2a5b0d828c7580c92f689524a938b1b7d7

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp312-cp312-macosx_10_12_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp311-cp311-win_amd64.whl.

File metadata

  • Download URL: hypertile-0.1.0-cp311-cp311-win_amd64.whl
  • Upload date:
  • Size: 257.7 kB
  • Tags: CPython 3.11, Windows x86-64
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for hypertile-0.1.0-cp311-cp311-win_amd64.whl
Algorithm Hash digest
SHA256 17bb633134e45247e131374e82f4a8b98b83d71ec6a13114a4b5f0e517fceaf2
MD5 0717d38824a3840c88ec6a0c138f3506
BLAKE2b-256 fe4638510721763e6f4f2407137298360b8aff1c8b5d085c1f5325f1600e6554

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp311-cp311-win_amd64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
Algorithm Hash digest
SHA256 7d4be046cee7c4b2f22ac4c2f0c8781d2f5c27395c969037e252170856f8aa53
MD5 d579fcf052e113bfb7c9550c43be2262
BLAKE2b-256 247469acaad3576483271dff234e7a0d430b5d1c8071521d349a5472bd9a671a

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
Algorithm Hash digest
SHA256 a4224d970c1aff213f005e0a86a08ef41b7ecdc2279a15a83337b896d3e6295c
MD5 f50e36e8a9707d25104c067f1f64eb8d
BLAKE2b-256 78d6eb949fa197227ad0f2bc677be5a656ef6f170bcbf0cc11b032f140c084d1

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp311-cp311-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp311-cp311-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 0e0fd5d511e116a20b8dd3d5aa3badddd34f5154fcf78d95a88613794cc50470
MD5 606731a62920a157438523916291c1f6
BLAKE2b-256 852538d727956de8aaf4d5cbe93d00da7d2fb3d7702cba2a45383c7f4cbd1874

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp311-cp311-macosx_11_0_arm64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file hypertile-0.1.0-cp311-cp311-macosx_10_12_x86_64.whl.

File metadata

File hashes

Hashes for hypertile-0.1.0-cp311-cp311-macosx_10_12_x86_64.whl
Algorithm Hash digest
SHA256 074198674fbf7a92c6a5e40652b048926b45579cbb62f8e686b0768591189981
MD5 7d14a30c79b5598a4ccbd961e11bee8f
BLAKE2b-256 6d4a5b74298b10b29e38ac7ae5483420d65dd145cd6fb116f28b10489601c78c

See more details on using hashes here.

Provenance

The following attestation bundles were made for hypertile-0.1.0-cp311-cp311-macosx_10_12_x86_64.whl:

Publisher: release.yml on kuntal-devrat/hypertile

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Release history Release notifications | RSS feed

0.1.1

19 files

This release

0.1.0 This release

19 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page