https://twitter.com/CuPy_Team/status/1313569720310865927
We're presenting at NVIDIA #GTC20!
Come and learn the basics of CuPy and highlights of the latest v8 release:
https://www.nvidia.com/en-us/gtc/session-catalog/?search=A21531#/session/1597988469506001EUWx
https://twitter.com/CuPy_Team/status/1321717254866677760
We have released CuPy v9.0.0a1, this is the first step towards the v9 release!
CUDA 11.1 support has been added so now you can use CuPy with your GeForce RTX 30 series or Quadro RTX. Stay safe and check the release note to download CUDA 11.1 builds at:
https://t.co/al3GTzPPRl
https://twitter.com/CuPy_Team/status/1321717791100035073
CuPy v8.1.0, the latest stable release with CUDA 11.1 support, multiple bug corrections and enhancements, is also released today. Check the release note at:
https://t.co/2WSzwniATe
https://twitter.com/CuPy_Team/status/1331861803035365377
Some fixes and bug corrections in the new CuPy v8.2.0. Stay safe and check the release note at: https://github.com/cupy/cupy/releases/tag/v8.2.0
https://twitter.com/CuPy_Team/status/1331862091121131520
CuPy v9.0.0a2 has been released with multiple improvements! We have added the initial support for our Python CUDA JIT that we will be enhancing towards the major release. Stay safe and check the release note at: https://t.co/MLWgdpIa04
https://twitter.com/CuPy_Team/status/1334240930501521411
Do you have questions or need help with CuPy? Or just want to have a casual chat with the developers? Join us on our Gitter channel: https://t.co/Zs67iTtUB6
https://twitter.com/CuPy_Team/status/1342343695081689088
Happy Holidays! We are releasing CuPy v9.0.0b1 where we add several new SciPy sparse routines and improve the overall stability of the library. Also we are releasing v8.3.0 with several corrections to the v8 series. Check the release notes at: https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1354664559202582531
Just released CuPy v9.0.0b2 with initial support for the new @numpy_team random API. Also, improved @AMD #ROCm support and other features that you can check in the release note. v8.4.0 with several corrections to the v8 series is also released today.
https://t.co/Fu0ranE13s
https://twitter.com/adamlikesai/status/1359150475212447746
CuPy retweeted:
1/ Ever wondered how to profile @CuPy_Team code with @nvidia #NSIGHTSystems? Use the prof time_range markers to den… https://twitter.com/i/web/status/1359150475212447746
https://twitter.com/CuPy_Team/status/1365212719121457153
Released CuPy v9.0.0b3/8.5.0. Improved CuPy JIT support in
cupy.vectorize and support the latest NumPy/SciPy releases. As always thanks to @NVIDIA and @RAPIDSai for their continued support! Check the list of changes in the release notes and stay safe!
https://twitter.com/CuPy_Team/status/1365213415954731010
Released CuPy v9.0.0b3/8.5.0. Improved CuPy JIT support in
cupy.vectorize and support the latest NumPy/SciPy releases. As always thanks to @NVIDIA and @RAPIDSai for their continued support! Check the list of changes in the release notes and stay safe!https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1365213741873172483
We are releasing a wheel with @AMD ROCm/HIP support, get your hands on it by installing the
cupy-rocm-4-0 package in PyPI!https://t.co/nEGR5wZgUa
https://twitter.com/CuPy_Team/status/1365219592604295170
We have found a PyPI package named "cupy-cuda112" which was NOT uploaded by the CuPy project. Please do not install it. We are contacting @PyPI to address this issue. https://t.co/6MpMXR5i02
https://twitter.com/CuPy_Team/status/1374977760867753989
Released CuPy v8.6.0. This is the last release for CuPy v8 where we pave the road for the upcoming v9 release next month! We have fixed several issues to make your CuPy experience smoother while reving up our v9 engine! Check the release note at:
https://t.co/MJFcPT50QP
https://twitter.com/CuPy_Team/status/1374978269217349641
Released CuPy v9.0.0rc1, the feature-complete release of series 9!
Write CUDA kernels in Python, support for GUFuncs thanks to @dask_dev, and other features thanks to @NVIDIA @RAPIDSai support and many other contributors! Check the release note at:
https://t.co/cL1CDVOkqS
[Twitter]
https://twitter.com/CuPy_Team/status/1377120980783390724
We're excited to be participating in Google Summer of Code @gsoc again this year!
:bulb: Project Ideas: https://t.co/ZMnznJLj9h
:memo: Guide: https://t.co/1R7cCzHRsT & https://t.co/KCqr2FBS9G
:speech_balloon: Gitter Chat (feel free to ask questions or discuss with mentors): https://t.co/Zs67iTtUB6
https://twitter.com/gmarkall/status/1379355919851646976
CuPy retweeted:
@numba_jit Not demoed in the notebook, more important @numba_jit 0.53 CUDA features:
- CUDA 11.2 support
- CUDA Ar… https://twitter.com/i/web/status/1379355919851646976
https://twitter.com/CuPy_Team/status/1382205987768070145
Watch sessions featuring CuPy by @nvidia and users at #GTC21!
https://www.nvidia.com/en-us/gtc/catalog/?search=CuPy
[Twitter]
https://twitter.com/CuPy_Team/status/1385113170344169477
:tada: We have just released CuPy v9.0.0! Say hi to the JIT, more new functions, the latest CUDA features, and much more! Thanks to all our contributors and specially @NVIDIA and @BrookhavenLabCS.
:scroll: Release Notes: https://t.co/Se4KOSu5XS
:bookmark: Highlights: https://t.co/zkiyEuFsCC
[Twitter]
https://twitter.com/CuPy_Team/status/1385113517003444224
With CuPy v9 comes official @AMD #ROCm HIP support, install it now with pip install cupy-rocm-4-0 and let’s make those Radeons crunch some numbers!
https://t.co/fUNT31BST9
:tada: We have just released CuPy v9.0.0! Say hi to the JIT, more new functions, the latest CUDA features, and much more… https://twitter.com/i/web/status/1385113170344169477
[Twitter]
https://twitter.com/phoronix/status/1387815915954360323
CuPy retweeted:
#CuPy 9.0 Brings @AMD @Radeon GPU Support To This #Numpy-Compatible #Python Library
https://www.phoronix.com/scan.php?page=news_item&px=CuPy-9.0-Released
https://twitter.com/CuPy_Team/status/1397823695905726465
Get your hands now on CuPy v10.0.0a1 / v9.1.0! As usual thanks to our incredible contributors at @nvidia @RAPIDSai @BrookhavenLabCS and many other places!
CUDA 11.0 and 11.1 wheels are not available on PyPI yet. Refer to the release notes for instructions: https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1397824179328622597
We have released binary wheels CuPy v10.0.0a1 / v9.1.0 for #ROCm 4.0! Install them now by piping the cupy-rocm-4-0 package =D
Thanks to @AMD for their support in making these releases possible!
Get your hands now on CuPy v10.0.0a1 / v9.1.0! As usual thanks to our incredible contributors at @nvidia @RAPIDSai… https://twitter.com/i/web/status/1397823695905726465
https://twitter.com/r_c_thomas/status/1399177938759602176
CuPy retweeted:
Daniel’s going to talk (@SciPyConf) about what he did to accelerate the code 20x porting it to @nvidia GPUs using C… https://twitter.com/i/web/status/1399177938759602176
[Unknown, chainer] Available commands (see https://sameroom.io/faq for more info):
^^^ <message>* -- Post <message> to your side of Tube only (hush command)
*-sameroom ping -- Ping all connected rooms
-sameroom map -- Show map of all connected rooms
-sameroom disable -- Delete all Tubes connected to a room
-sameroom format -- Show the current posting format for a room
-sameroom format <format option>* -- Set posting format for a room (see https://sameroom.io/blog/introducing-posting-options/ for details). <format option> can be one of the following: *off, dispname, username, dispname team, username team, dispname username, dispname username team
-sameroom portal -- Generate a Portal URL
-sameroom open -- Initiate a connection. Return a code to be used with -sameroom connect
*-sameroom connect <code>* -- Create a Tube to room where code was issued by -sameroom open
*-sameroom connect <Portal URL>* -- Create a Tube via the Portal URL issued by -sameroom portal
[Twitter]
https://twitter.com/pyconil/status/1405862105492856832
CuPy retweeted:
"Python bottleneck optimization progression from lists to cupy arrays" by Yair Beer from Pycon Israel 2021
https://twitter.com/CuPy_Team/status/1407980387112873988
Summer is here so let's heat up things a bit with the new CuPy releases.
v10.0.0a2/v9.2.0 come with official support for @NVIDIA CUDA 11.3.
Thanks to our awesome contributors.
Check out the release notes for important info:
https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1407980626817417232
We bring you support for ROCm 4.2 in the new CuPy v10.0.0a2/v9.2.0 releases.
Thanks to @AMD and pip it now with
pip install cupy-rocm-4-2.Summer is here so let's heat up things a bit with the new CuPy releases.
v10.0.0a2/v9.2.0 come with official suppor… https://twitter.com/i/web/status/1407980387112873988
[Twitter]
https://twitter.com/CuPy_Team/status/1415216735909679107
CuPy uses Optuna for kernel launch parameter optimization: https://docs.cupy.dev/en/latest/reference/generated/cupyx.optimizing.optimize.html
Looking forward to see the evolution of great hyperparameter optimization framework :relaxed:
Help us create the next version of Optuna!
Please take a few minutes to fill in this survey, and let us know how y… https://twitter.com/i/web/status/1414784107574222852
https://twitter.com/labs_mann/status/1419017799163432964
CuPy retweeted:
Today, we open up our GitHub repository for AlphaPept, our brand new framework for open, fast, #Python-based next g… https://twitter.com/i/web/status/1419017799163432964
https://twitter.com/CuPy_Team/status/1423199806789324802
CuPy v10.0.0b1/v9.3.0 releases are in time for the Olympic games! Added CUDA 11.4 support and other exciting features. CuPy also participates in GSoC’21! Check the release notes here and don't forget to star on GitHub!:hugging_face:
https://t.co/Fu0ranE13s
https://twitter.com/miguelusque/status/1424759483591503880
CuPy retweeted:
Interested in ML Frameworks Interoperability? @koalalesque and I have been working on a set of blogs covering this… https://twitter.com/i/web/status/1424759483591503880
[Twitter]
https://twitter.com/CuPy_Team/status/1427374613244387328
We are excited to announce that we just merged the initial support for CUDA Python in CuPy! :tada:
Having the unified, official Python bindings of the CUDA API is surely helpful for the entire ecosystem :$
https://t.co/b1pZLIjCgk
The NVIDIA CUDA Python runtime, driver, and NVRTC API bindings are now in a public Github repo:… twitter.com/i/web/status/1…
https://twitter.com/adamlikesai/status/1427396766178086913
CuPy retweeted:
I built #cusignal this afternoon with the development version of cupy and cuda-python, and I'm pleased to report th… twitter.com/i/web/status/1…
We are excited to announce that we just merged the initial support for CUDA Python in CuPy! :tada:
Having the unified, o… twitter.com/i/web/status/1…
https://twitter.com/CuPy_Team/status/1430799187025698819
CuPy v10.0.0b2/v9.4.0 has just been released. We bring support for CUDA Python in the v10 series and increased SciPy/NumPy compatibility thanks to GSoC’21!
Check the release notes here and don't forget to star on GitHub!
https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1430799475740606468
Starting in v10.0.0b2, pre-releases binary packages will not be uploaded to @pypi in order to keep a clean ecosystem. Install the new pre-releases with:
pip install cupy-cuda*** -f https://t.co/suK8Ma8kgN
CuPy v10.0.0b2/v9.4.0 has just been released. We bring support for CUDA Python in the v10 series and increased SciP… twitter.com/i/web/status/1…
We are going to remove CuPy's pre-release binary wheels (
cupy-cuda*** ) from PyPI on 2021-09-20. Please consider using the stable releases, or follow the updated installation instructions if you depend on pre-releases. See cupy/cupy#5667 for details. Thank you for your understanding. :pray:
[Twitter]
https://twitter.com/CuPy_Team/status/1433309172687245312
You can integrate CuPy code with Numba, mpi4py, PyTorch, TensorFlow, RAPIDS, or any libraries supporting DLPack!
Read more on here: https://t.co/uDe3hpyger
3 part Machine Learning Framework Interoperability guide:
Memory Layouts and Memory Pools: https://developer.nvidia.com/blog/machine-learning-frameworks-interoperability-part-1-memory-layouts-and-memory-pools… https://twitter.com/i/web/status/1433094935985770496
https://twitter.com/CuPy_Team/status/1435465647178149895
We are thinking of dropping support for CUDA 10.1 or earlier in CuPy v10 planned on November. Any thoughts?
https://twitter.com/CuPy_Team/status/1443503060370616322
We have released CuPy v10.0.0b3/v9.5.0! v10 has started to implement the Python Array API for interoperability with other libraries! Check the release notes here: https://t.co/Fu0ranE13s
https://twitter.com/CuPy_Team/status/1443503238892851200
Thanks to those who participated in the poll! As the majority of users already use CUDA 10.2+, we finally decided to drop CUDA ~10.1 in CuPy v10. For those in need, please use CuPy v9 − but we encourage upgrading CUDA as it will provide better performance.
We are thinking of dropping support for CUDA 10.1 or earlier in CuPy v10 planned on November. Any thoughts?
https://twitter.com/whostolehonno/status/1446103972272553984
CuPy retweeted:
Soon libraries like @scikit_learn will be able to support @numpy_team, @PyTorch, @CuPy_Team, etc. all at once than… https://twitter.com/i/web/status/1446103972272553984
https://twitter.com/CuPy_Team/status/1449036624365428742
:man-raising-hand: Feedbacks wanted: we plan to add more sparse graph algorithms (
scipy.sparse.csgraph equivalent) to CuPy.If you have any suggestion/wishlist to help us prioritize, please feel free to leave your comments in this thread!
https://t.co/3wqEDn4wwj
https://twitter.com/haesleinhuepf/status/1451669700065341446
CuPy retweeted:
Hi @CuPy_Team and @napari_imaging fans,
I just made another #GPUaccelerated #BioImageAnalysis plugin and would love… https://twitter.com/i/web/status/1451669700065341446
https://twitter.com/CuPy_Team/status/1458694913143771136
CuPy v10.0.0rc1/9.6.0 released! Get a preview of the new cupyx.distributed module for easy NCCL use, Array API coverage increased and more! Check the release notes at: https://t.co/Fu0ranWahA
[Twitter]
https://twitter.com/CuPy_Team/status/1460530569058856960
The world's fastest supercomputer runs with ARM and CuPy also adds support for ARM architectures (bicep emoji)
Download it for your NVIDIA Jetson devices!
https://t.co/RZGVJbHP62
#SC21
Congratulations to the TOP500 Supercomputers as announced at #SC21! Some highlights:
:one: Supercomputer Fugako (:jp:)… https://twitter.com/i/web/status/1460339331017625602
https://twitter.com/CuPy_Team/status/1468835301208760325
:mega: We have released CuPy v10.0.0! This version packs so many features and improvements that it is hard to describe them in a few characters. Thanks to our wonderful contributors and happy holidays!
:scroll: Release Notes: https://t.co/sV8TuSVXra
:bookmark: Highlights: https://t.co/QTp0tDKFmd