Ashok Bhat

Prod Mgr at Arm - Tweets are my own and are not the views of my employer.

[Arm Software Developers @ArmSoftwareDev]

We've been working to improve @PyTorch inference performance. 💪

In fact, on Arm Neoverse, integration with Arm Kleidi brings performance gains of up to 2x in Eager & Graph execution modes on @awscloud Graviton. 

Discover how we make this possible: https://t.co/QbIA8wINmf

[Ashok Bhat @AshokIVBhat]

Excellent blog on Accelerated ML on Arm Neoverse based Alibaba's Yitian 710 https://t.co/Wg8pgvoO27

[Ashok Bhat @AshokIVBhat]

@filippospiga Yes :-)

[What's New on AWS (Unofficial) @awswhatsnew]

Amazon SageMaker adds eight new Graviton-based instances for model deployment

Amazon SageMaker expands access to eight new Graviton2 and Graviton3-based machine learning (ML) instance families so that customers have more options for optimizing thei... https://t.co/56mch09esB

[Arm @Arm]

👋 Meet Neoverse V2, our new V-series core with:

- Market-leading integer performance
- CXL & UCIe-ready CMN-700 mesh interconnect
- Armv9 security enhancements

Rooted in scalable efficiency, this helps build the future of infrastructure #onArm: https://t.co/eB1s6KDOgG

[Ashok Bhat @AshokIVBhat]

With TensorFlow 2.10 adding 64-bit Arm support, now you can install using "pip install tensorflow" on Arm hardware. This has been possible thanks to many months of effort by @LinaroOrg, @awscloud , @Google and @Arm teams.

[TensorFlow @TensorFlow]

🎉TensorFlow 2.10 has been released!

Highlights of this release include new user-friendly features in Keras, performance enhancements, expanded GPU support in Windows, and more.

💡Learn more here → https://t.co/WXScQxCjj3

[Ashok Bhat @AshokIVBhat]

Official TensorFlow Nightly build for 64-bit Arm now available - thanks to the @awscloud and @Google team. Do try the package and give us feedback. It contains oneDNN+ACL optimizations that Arm contributes to. You can soon expect an official TF release build.

[Ashok Bhat @AshokIVBhat]

Atos showcasing Ampere Altra (with Neoverse N1) performance for life sciences https://t.co/bLoukFWVmK via @YouTube

[Ashok Bhat @AshokIVBhat]

If you want to develop using SVE intrinsics on any platform (Arm without SVE support, or x86_64), have a look at the Farm-SVE project.

[Frank @jedisct1]

Farm-SVE: A scalar C++ implementation of the ARM® Scalable Vector Extension (SVE) https://t.co/3mLebG7a6p

[Ashok Bhat @AshokIVBhat]

Learn what's new in Arm Compiler for Linux 22.0, Arm provided HPC toolchain (with C/C++/Fortran Compiler and BLAS/LAPACK/FFT libraries). https://t.co/k1XGShz1Sc

[Arm @Arm]

AWS Graviton3 is officially here! Based on @Arm Neoverse CPU, it’ll help you:
💲 Reduce costs
♻️ Improve efficiency
⚡️ Enhance performance

Congrats @awscloud 🎉🎉

What workloads do you want to try on Amazon’s EC2 C7g instances based #onArm?

[Jeff Barr ☁️ @jeffbarr]

New – Amazon EC2 C7g Instances, Powered by AWS Graviton3 Processors - https://t.co/V7ehhP3pJb #AWS #EC2

C7g instances are available in 8 sizes with 1-64 vCPUs, 2-128 GiB of memory, up to 30 Gbps of network bandwidth, and up to 20 Gbps of EBS bandwidth.

[Ashok Bhat @AshokIVBhat]

@gnutools Congrats on the release. Tamar from Arm has put out a blog covering what's new and improved on Arm https://t.co/Y7gipBfuwO

[Ashok Bhat @AshokIVBhat]

We have now made Arm Compiler for Linux free-to-use. HPC apps can now benefit from Arm's optimized libraries and compilers on Arm Neoverse-based servers. Further details in my blog https://t.co/BRuFCqaFTd

[Ashok Bhat @AshokIVBhat]

I came across an interview with Jim Keller https://t.co/6NE0lFgVcL. Amazing! So, many life and technology lessons. I am sure many of you already know it. Just in case :-)

[Ashok Bhat @AshokIVBhat]

Anaconda packages now available for @Arm Neoverse-N1

[Jeff Underhill @JeffUnderhill]

Anaconda support now available on AWS Graviton2 - check it out and let us know what you think! details: https://t.co/RZ9gsEpWAB #AWS #Graviton2 @anacondainc

[Ashok Bhat @AshokIVBhat]

Curious about what GCC11 can do for SVE code gen in the current and future Arm-based hardware - Read the blog by Kyrylo Tkachov https://t.co/gGQST355Fs

[Ashok Bhat @AshokIVBhat]

To get started, take a look at the blog by Jason Andrews @Arm https://t.co/GQyFOGAhTb

[Ashok Bhat @AshokIVBhat]

Starting Apr 21, Arm plans to release monthly snapshots of TensorFlow and PyTorch containers for Arm Neoverse N1 cores https://t.co/i4BKbgFx3n

[Ashok Bhat @AshokIVBhat]

Blog from Arm on the performance improvements in GCC 11 https://t.co/vG3m20blzz

[Ashok Bhat @AshokIVBhat]

GCC 11.1 has now been released with better Arm support - Better auto-vectorization, new CPU support including Arm Neoverse-N2 and Neoverse-V1. See https://t.co/GZcyTAS0Dk for details.

[GCC - GNU Toolchain @gnutools]

GCC 11.1 Released with huge C++ and optimization improvements! Continually improving. https://t.co/jW3yHD9z7j