Top all-time · #272
NVIDIA Technical Blog
Company blog · devblogs.nvidia.comprogrammingai
NVIDIA Technical Blog is the company blog at devblogs.nvidia.com, writing about programming and ai. Since 2013, 177 different Hacker News readers have linked it 313 times, most often “Volta: Advanced Data Center GPU”. It is still writing: the last post was on 2026-10-30, and you can follow it by feed.
Rarely recommended · 2 of 40 sampled comments recommend it?
Why HN links it
Linked as a technical reference for GPU work, including architecture posts on Volta, Pascal and Turing, TensorRT, tensor cores and CUDA learning. Most comments cite posts as sources rather than judging them.
Summarised by Claude from a sample of 40 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
News and tutorials for developers, scientists, and IT admins
Most-linked posts?
- Volta: Advanced Data Center GPU8 people · 8 mentions · first linked 2017-05-10
- Inside Pascal: Nvidia 's Newest Computing Platform7 people · 8 mentions · first linked 2016-04-05
- GOAI: Open GPU-Accelerated Data Analytics5 people · 5 mentions · first linked 2017-06-13
- End-To-End Deep Learning for Self-Driving Cars4 people · 5 mentions · first linked 2016-08-19
- Understanding Aesthetics with Deep Learning4 people · 5 mentions · first linked 2016-03-01
- /inside-volta/4 people · 4 mentions · first linked 2018-02-23
- Nvidia Announces Jetson TX24 people · 4 mentions · first linked 2017-03-08
- Nvidia Ampere Architecture In-Depth – Nvidia Developer Blog4 people · 4 mentions · first linked 2020-05-14
- Mocha.jl: Deep Learning for Julia3 people · 4 mentions · first linked 2015-09-01
- Nvidia CUDA 8 Release Candidate Now Available3 people · 4 mentions · first linked 2016-04-05
What Hacker News says?
The link I shared wasn't UE5, and their other videos specifically indicate that they are running on PC - unbelievably modest PC hardware at that. > it's not a trivial thing to get a demo like this running well on PCs This old tired argument. The current theory for how the meshing is done is something similar to mesh shaders (available on commodity PC hardware since 2018)[1]. This "PS5 platform specific feature" runn…
Comment linking /introduction-turing-mesh-shaders/
on “A first look at Unreal Engine 5”Submitted Volta: Advanced Data Center GPU as a story · 278 points
HN discussionThe topmost of notches. Very reproducible and user-friendly. https://devblogs.nvidia.com/docker-compatibility-singularity...
Comment linking /docker-compatibility-singularity-hpc/
on “Docker cannot be downloaded without logging into Docker Store”Submitted Numba: High-Performance Python with CUDA Acceleration as a story · 276 points
HN discussionYou could just use LoopVectorization on the CPU side. It's been shown to match well-tuned C++ BLAS implementations, for example with the pure Julia Gaius.jl ( https://github.com/MasonProtter/Gaius.jl ), so you can follow that as an example for getting BLAS-speed CPU side kernels. For the GPU side, there's CUDAnative.jl and KernelAbstractions.jl, and indeed benchmarks from NVIDIA show that it at least rivals directly…
Comment linking /gpu-computing-julia-programming-language/
on “Swift: Google’s Bet on Differentiable Programming”Submitted CUDA on Windows Subsystem for Linux 2 as a story · 268 points
HN discussionThese guys are using deep learning/convolutional neural networks to train their models to detect beautiful/thoughtful photographic images. https://www.eyeem.com/ They talk about their curation algorithms here: https://devblogs.nvidia.com/parallelforall/understanding-aes...
Comment linking /parallelforall/understanding-aesthetics-deep-learning/
on “Ask HN: Where can I find high-end stock images for a website?”Submitted Real-Time Noise Suppression Using Deep Learning as a story · 256 points
HN discussionAlso worth checking out OmniSci (formerly MapD), which features an LLVM query compiler to gain large speedups executing SQL on both CPU and GPU: https://github.com/omnisci/omniscidb . And here's a link to a blog post giving a high level overview of the advantages of JIT compilation of queries over an interpreter: https://devblogs.nvidia.com/mapd-massive-throughput-database... .
Comment linking /mapd-massive-throughput-database-queries-llvm-gpus/
on “Weld: Accelerating numpy, scikit and pandas as much as 100x with Rust and LLVM”Submitted An Even Easier Introduction to CUDA as a story · 247 points
HN discussionSeems to lose in Mobilenet and win in other benches according to Nvidia's benchmarks. I'll wait for independent testing before I drop $100. https://devblogs.nvidia.com/jetson-nano-ai-computing/
Comment linking /jetson-nano-ai-computing/
on “Nvidia's $99 Jetson Nano Is an AI Computer for DIY Enthusiasts”Submitted How Jet Built a GPU-Powered Fulfillment Engine with F# and CUDA as a story · 203 points
HN discussionhttps://developer.nvidia.com/embedded/buy/jetson-nano-devkit https://devblogs.nvidia.com/jetson-nano-ai-computing/
Comment linking /jetson-nano-ai-computing/
on “Nvidia's $99 Jetson Nano Is an AI Computer for DIY Enthusiasts”Submitted Hybridizer: High-Performance C# on GPUs as a story · 169 points
HN discussionNo, the A100 has a 19.5 TFLOP theoretical peak for SGEMM[1], real world benchmarks will likely achieve 93% of that, and so the DGX A100 will be 145 TFlops of FP32 SGEMM performance or 0.145 FP32 PFLOPS. Maybe in 72 FP64 TFLOPS. FP64 is what the TOP500 benchmaks count.[2] The 5 "petaflops" number is a creatively constructed marketing number based on FP16 TensorCore "flops", sparse matrix calculations, and then multip…
Comment linking /nvidia-ampere-architecture-in-depth/
on “Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU”Submitted Accelerated “Ray Tracing in One Weekend” in CUDA as a story · 165 points
HN discussionEven more details: https://devblogs.nvidia.com/nvidia-ampere-architecture-in-de...
Comment linking /nvidia-ampere-architecture-in-depth/
on “Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU”Submitted Using Deep Learning to model personal visual aesthetics as a story · 144 points
HN discussionNVLink works fine between GPUs even if the CPUs are x86. Just look at the NVIDIA DGX boxes/stations. https://devblogs.nvidia.com/wp-content/uploads/2017/04/NVLin...
Comment linking /wp-content/uploads/2017/04/NVLink-topology-e1491278374674.png
on “Nvidia announces RTX 2000 GPU series with ‘6x more performance’ and ray-tracing”Submitted High-Performance GPU Computing in the Julia Programming Language as a story · 139 points
HN discussionThis is a great intro into the subject. We use LLVM to at MapD ( http://mapd.com ) to compile SQL queries to CPU and GPU machine code - it has given us a major boost over an interpreter based approach. For more see here - http://devblogs.nvidia.com/parallelforall/mapd-massive-throu... . If you have a background in LLVM or compilers in general and are interested in tackling problems like this please reach out at jobs…
Comment linking /parallelforall/mapd-massive-throughput-database-queries-llvm-gpus/
on “LLVM for Grad Students”Submitted High-Performance GPU Computing in the Julia Programming Language (2017) as a story · 122 points
HN discussionIn the paper they say it took 3.14e23 flops to train. They used v100s to do it. This is an insane energy cost (and financial cost). Nvidia's v100 product page [0] says that it gets about 15 (single precision) - 125 ("deep learning") teraflop/s at 250-300 watts (joules per second). That means that if everything's as perfectly efficient as a marketing product page, it gets about 250/125-300/15 = 2-25 joules per terafl…
Comment linking /nvidia-ampere-architecture-in-depth/
on “GPT-3: Language Models Are Few-Shot Learners”Submitted Inside Pascal: Nvidia 's Newest Computing Platform as a story · 122 points
HN discussionIf you dont know already: Azure has started investing heavily in offloading the task from CPU to GPU and FPGA. There are really good framework on .NET to achieve the same. NVIDIA CUDA GPU - https://devblogs.nvidia.com/hybridizer-csharp/ GPU scripting and CUDA - http://www.aleagpu.com/release/3_0_4/doc/ FPGA - https://hastlayer.com/project , https://netfpga.org/site/#/
Comment linking /hybridizer-csharp/
on “.NET Core 3.0”Submitted Nvidia Accelerates Real Time Speech to Text Transcription 3500x with Kaldi as a story · 98 points
HN discussionGraphistry is pretty under the radar as we're focusing more on big enterprise/federal security users and certain data science users before opening to others, but we've been investing big on Arrow for our V2. Ex: We recently released https://github.com/apache/arrow/tree/master/js and are generally using Arrow as a way to compose interactive-time columnar CPU/GPU visual analytics technologies, including our own: https…
Comment linking /parallelforall/goai-open-gpu-accelerated-data-analytics/
on “Apache Arrow and the “Things I Hate About Pandas””Submitted Thinking Parallel, Part I: Collision Detection on the GPU (2012) as a story · 93 points
HN discussionmy recent blog post had some details about the tech behind krisp: https://devblogs.nvidia.com/nvidia-real-time-noise-suppressi...
Comment linking /nvidia-real-time-noise-suppression-deep-learning/
on “Krisp.ai – Mute background noise during your calls”Submitted Deep Learning in a Nutshell: History and Training as a story · 87 points
HN discussion
