open-source multi-lang GPGPU eng. @nvidia. @rapidsai, @apachearrow, @reactivex. https://nitter.gq/t.co/LzvJHVPdJO
atx
Joined March 2008
- Tweets5.5K
- Following132
- Followers1.2K
- Likes1.2K
paul taylor retweeted
Excited to be (virtually) back at #GTC23! Join me for a deep dive into concurrent #CUDA code with stream-ordered allocation. Learn how to avoid common pitfalls and achieve safe, lightning-fast performance. C++ examples included! Register now: nvidia.com/gtc/session-catal…
paul taylor retweeted
Everyone knows GPUs speed up ML training but did you know they can also turbocharge your analytics workflow? Our friends from @nvidia explain how to get started with @RAPIDSai cuDF.
medium.com/google-colab/rapi…
paul taylor retweeted
RAPIDS available via pip on an experimental basis. Send tweet. rapids.ai/pip.html
I just pushed some updates to download the datasets + instructions for running the demos (needs Ubuntu20.04+ and an NVIDIA GPU)
I put the code up on GitHub here: github.com/trxcllnt/rapids-c…, and will post links to data sources when I get a chance (on vacation now)
paul taylor retweeted
Kicking off part 2 of #openvizsummit22 we have @inlineptx giving us a great live demo of GPU-accelerated Geospatial Analytics with @nvidia RAPIDS, working with less common #geospatial formats like CSVs, parquet and torque 🤩
paul taylor retweeted
1B Uint32 values loaded from CPU RAM to GPU RAM in 235ms (17GB/s) while sharing my PCIe 4.0 bus (64GB/s) with three other VMs. Nice! Many thanks to @inlineptx for his help. @NVIDIA rocks!
paul taylor retweeted
I had to pleasure of having @coding_matt intern with us on the @RAPIDSai Viz team this last year. Check out his blog post detailing one of neat things he helped us with for @nodejs + RAPIDS - mzegar.github.io/posts/20211…
I'm sure the scientists at ANL would love to hear how they didn't factor in mining and manufacturing costs in their analysis where they explicitly did just that.
The libraries do run on Jetson, but I haven't built the containers from the jetpack base images in the last few months. Possible a few x86-only dependencies slipped in w/ addition of NVENC WebRTC and SQL libs.
paul taylor retweeted
I'll be speaking today about GPU accelerated Data Science with Node.js at nodeconfremote.com/#speakers Come check it out, its free!
paul taylor retweeted
Take time to check out the talk @exactlyAllan will give today!! Did you know we have RAPIDS Node.js? If you did, do you want to know great ways to put it to use? The talk is free! And check out his blog post medium.com/rapids-ai/acceler…
I'll be speaking today about GPU accelerated Data Science with Node.js at nodeconfremote.com/#speakers Come check it out, its free!
seriously considering maintaining a chromium fork with the "tab hover cards" and "tab search" commits reverted
paul taylor retweeted
Here it is - node-RAPIDS - an #OpenSource, technical preview of modular @RAPIDSai bindings in @nodejs built to enable high-performance, server-side #visualizations and #Analytics. medium.com/rapids-ai/acceler…