#cuda

GitHub - bertaye/agentic-cuda-optimizer: A langgraph based workflow with a C++ CUDA suite to optimize CUDA kernels

🖼️

0

GitHub - bertaye/agentic-cuda-optimizer: A langgraph based workflow with a C++ CUDA suite to optimize CUDA kernels

Hacker News·11 days ago

#3KVY7XXs

#github #cuda #langgraph #gpu #nvrtc #photo

A langgraph based workflow with a C++ CUDA suite to optimize CUDA kernels - bertaye/agentic-cuda-optimizer

15s

GitHub - Speedstu/CUDA-for-AMD-Windows: Run CUDA-targeted Windows applications on AMD GPUs with ZLUDA + ROCm/HIP.

🖼️

0

GitHub - Speedstu/CUDA-for-AMD-Windows: Run CUDA-targeted Windows applications on AMD GPUs with ZLUDA + ROCm/HIP.

Hacker News·23 days ago

#QEfsSbTd

#github #cuda #windows #runtime #zluda #scripts

Run CUDA-targeted Windows applications on AMD GPUs with ZLUDA + ROCm/HIP. - Speedstu/CUDA-for-AMD-Windows

15s

🖼️

0

CUDA Shared Memory Swizzling

Hacker News·about 2 months ago

#XWZKywFA

#leimao #include #define #val #cuda #swizzling

View the full article

Create a free account to read full articles inline — no redirect to the original site.

Create account Log in

GitHub - PJHkorea/photonic-mesh-fng-router: A humble exploratory PoC for a hardware-native, optical timing-frozen control plane engine. Fuses inline CUDA PTX, RAII memory tunnels, and 4D JAX shard_map structures to investigate 0ns-overhead fault-tolerant routing for hyperscale distributed AI.

📰

0

GitHub - PJHkorea/photonic-mesh-fng-router: A humble exploratory PoC for a hardware-native, optical timing-frozen control plane engine. Fuses inline CUDA PTX, RAII memory tunnels, and 4D JAX shard_map structures to investigate 0ns-overhead fault-tolerant routing for hyperscale distributed AI.

Hacker News·about 2 months ago

#sZXcuiUa

#github #cuda #article #englishlanguage

A humble exploratory PoC for a hardware-native, optical timing-frozen control plane engine. Fuses inline CUDA PTX, RAII memory tunnels, and 4D JAX shard_map structures to investigate 0ns-overhead f...

15s

GitHub - graham33/nixos-dgx-spark: Use Nix and NixOS on your DGX Spark!

🖼️

0

GitHub - graham33/nixos-dgx-spark: Use Nix and NixOS on your DGX Spark!

Hacker News·2 months ago

#um255Hf0

#github #cuda #usb #sec #generate #dgx

Use Nix and NixOS on your DGX Spark! Contribute to graham33/nixos-dgx-spark development by creating an account on GitHub.

15s

📰

293

What happens when you run a CUDA kernel

Hacker News·3 months ago

#weGJiMor

#fergusfinn #define #include #cuda #article #discussion

Tracing one vector-add kernel from nvcc all the way down to the warps that execute it.

15s

GitHub - NVlabs/cutile-rs: cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.

🖼️

106

GitHub - NVlabs/cutile-rs: cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.

Hacker News·4 months ago

#tTZipU3j

#github #cutile #cuda #rust #tile #kernel

cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions. -...

15s

AMD shipped Nvidia's new AI laptop over a year ago, and the software is finally catching up

🖼️

0

AMD shipped Nvidia's new AI laptop over a year ago, and the software is finally catching up

XDA ·Adam Conway·4 months ago

#HshR6ahn

#xda-developers #rocm #nvidia #cuda #halo #tikoo

Nvidia's RTX Spark is competing in a space that AMD kickstarted over a year ago.

15s

Our LiDAR detector spent 40% of its time in voxelization, not convs

🖼️

0

Our LiDAR detector spent 40% of its time in voxelization, not convs

DEV Community: machinelearning·Elise Moreau·4 months ago

#mggrPY87

#dev #cuda #backbone #voxelization #part #frame

TL;DR: We profiled a LiDAR object detector expecting the 3D backbone to dominate. It didn't....

15s

Nvidia's long-awaited N1/N1X SoC specs leak ahead of Computex launch — N1 to feature up to 20 Arm-based cores, standard N1 equipped with 12- and 10-core configs

📰

0

Nvidia's long-awaited N1/N1X SoC specs leak ahead of Computex launch — N1 to feature up to 20 Arm-based cores, standard N1 equipped with 12- and 10-core configs

Latest from Tom's Hardware ·Hassam Nasir·4 months ago

#xpN10TnF

#tomshardware #core #cuda #cores #option #reportedly

The N1X reportedly comes in two SKUs: a top-end 20-core option with 6,144 CUDA cores matching the desktop RTX 5070, and a cut-down 18-core option with 5,120 CUDA cores.…

15s

NVIDIA CUDA 13.3 Enhances GPU Development with Tile Programming in C++, Compiler Autotuning, and Python

🖼️

0

NVIDIA CUDA 13.3 Enhances GPU Development with Tile Programming in C++, Compiler Autotuning, and Python

NVIDIA Technical Blog·Jonathan Bentz·4 months ago

#MjyCwVQT

#developer #include #cuda #cccl #import #python

NVIDIA CUDA 13.3 brings new capabilities and performance optimizations to developers across the CUDA ecosystem. The launch of NVIDIA CUDA Tile programming in…

15s

Develop High-Performance GPU Kernels in C++ with NVIDIA CUDA Tile

🖼️

0

Develop High-Performance GPU Kernels in C++ with NVIDIA CUDA Tile

NVIDIA Technical Blog·Jonathan Bentz·4 months ago

#VjvQQoBg

#developer #include #define #tile #auto #float

Developers can now use NVIDIA CUDA Tile programming within large existing C++ GPU codebases to develop highly optimized GPU kernels using tile-based…

15s

Tracing torch.cuda.empty_cache() on an RTX 4090 - Where Do the 53 MB Go?

🖼️

0

Tracing torch.cuda.empty_cache() on an RTX 4090 - Where Do the 53 MB Go?

DEV Community: pytorch·Ingero Team·4 months ago

#iPThd87N

#dev #empty_cache #pytorch #cuda #memory #allocator

TL;DR After del tensor; torch.cuda.empty_cache(), PyTorch's caching allocator still...

15s

What a GPU Actually Is (and Why ML Stole It)

🖼️

0

What a GPU Actually Is (and Why ML Stole It)

DEV Community·Abhishek Gautam·5 months ago

#VrA583cZ

#section #three #demo #why #cuda #memory

Introduction You've written model.to('cuda') a hundred times. You've celebrated when...

15s

RTX 5090, LLaMA.cpp TurboQuant, & Blackwell CUDA Scheduling Boosts GPU Performance

🖼️

0

RTX 5090, LLaMA.cpp TurboQuant, & Blackwell CUDA Scheduling Boosts GPU Performance

DEV Community·soy·5 months ago

#uhqyEBXj

#nvidia #gpu #hardware #software #performance #cuda

From Dev Community: RTX 5090, LLaMA.cpp TurboQuant, & Blackwell CUDA Scheduling Boosts GPU Performance

15s

Why CUDA kernels silently corrupt memory and how to catch the bug

🖼️

0

Why CUDA kernels silently corrupt memory and how to catch the bug

DEV Community·Alan West·5 months ago

#pAAsZBHU

#ifndef #cuda #rust #kernel #scratch #compute

A practical guide to debugging silent memory corruption in CUDA kernels, with compute-sanitizer workflows and a look at Rust-on-GPU tooling.

15s

How to Deploy Llama 3.2 90B with vLLM + Speculative Decoding on a $16/Month DigitalOcean GPU Droplet: 2.5x Faster Inference at 1/110th Claude Cost

🖼️

0

How to Deploy Llama 3.2 90B with vLLM + Speculative Decoding on a $16/Month DigitalOcean GPU Droplet: 2.5x Faster Inference at 1/110th Claude Cost

DEV Community·RamosAI·5 months ago

#pACPdpCa

#programming #tutorial #ai #fullscreen #llama #vllm

From Dev.to - tutorial: How to Deploy Llama 3.2 90B with vLLM + Speculative Decoding on a $16/Month DigitalOcean GPU Droplet: 2.5x Faster Inference at 1/110th Claude Cost

15s

CUDA Proves Nvidia Is a Software Company - Slashdot

🖼️

0

CUDA Proves Nvidia Is a Software Company - Slashdot

hardware.slashdot.org·5 months ago

#ZvzQBICT

#comments #modal_box #cuda #nvidia #gpus #single

Nvidia's real AI moat isn't "a piece of hardware," writes Wired's Sheon Han. It's CUDA: a mature, deeply optimized software ecosystem that keeps machine-learning workloads tied to Nvidia GPUs.…

15s

RTX 5080 Launched, Rust for CUDA, & LLM GPU Scheduling Deep Dive

🖼️

0

RTX 5080 Launched, Rust for CUDA, & LLM GPU Scheduling Deep Dive

DEV Community·soy·5 months ago

#sw7vv6rK

#nvidia #gpu #hardware #software #cuda #rust

From Dev RSS Feed: RTX 5080 Launched, Rust for CUDA, & LLM GPU Scheduling Deep Dive

15s

Nvidia libera CUDA-oxide v0.1.0: kernels GPU en Rust puro sin DSL ni FFI

🖼️

0

Nvidia libera CUDA-oxide v0.1.0: kernels GPU en Rust puro sin DSL ni FFI

DEV Community·lu1tr0n·5 months ago

#dfr0am45

#releasenotesv010 #qu #programming #tutorial #cuda #para

NVlabs publica el primer compilador oficial que convierte Rust estándar en kernels CUDA. Mismo archivo .rs para CPU y GPU, sin C++ de por medio.

15s

Menu

GitHub - bertaye/agentic-cuda-optimizer: A langgraph based workflow with a C++ CUDA suite to optimize CUDA kernels

GitHub - Speedstu/CUDA-for-AMD-Windows: Run CUDA-targeted Windows applications on AMD GPUs with ZLUDA + ROCm/HIP.

CUDA Shared Memory Swizzling

GitHub - PJHkorea/photonic-mesh-fng-router: A humble exploratory PoC for a hardware-native, optical timing-frozen control plane engine. Fuses inline CUDA PTX, RAII memory tunnels, and 4D JAX shard_map structures to investigate 0ns-overhead fault-tolerant routing for hyperscale distributed AI.

GitHub - graham33/nixos-dgx-spark: Use Nix and NixOS on your DGX Spark!

What happens when you run a CUDA kernel

GitHub - NVlabs/cutile-rs: cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.

AMD shipped Nvidia's new AI laptop over a year ago, and the software is finally catching up

Our LiDAR detector spent 40% of its time in voxelization, not convs

Nvidia's long-awaited N1/N1X SoC specs leak ahead of Computex launch — N1 to feature up to 20 Arm-based cores, standard N1 equipped with 12- and 10-core configs

NVIDIA CUDA 13.3 Enhances GPU Development with Tile Programming in C++, Compiler Autotuning, and Python

Develop High-Performance GPU Kernels in C++ with NVIDIA CUDA Tile

Tracing torch.cuda.empty_cache() on an RTX 4090 - Where Do the 53 MB Go?

What a GPU Actually Is (and Why ML Stole It)

RTX 5090, LLaMA.cpp TurboQuant, & Blackwell CUDA Scheduling Boosts GPU Performance

Why CUDA kernels silently corrupt memory and how to catch the bug

How to Deploy Llama 3.2 90B with vLLM + Speculative Decoding on a $16/Month DigitalOcean GPU Droplet: 2.5x Faster Inference at 1/110th Claude Cost

CUDA Proves Nvidia Is a Software Company - Slashdot

RTX 5080 Launched, Rust for CUDA, & LLM GPU Scheduling Deep Dive

Nvidia libera CUDA-oxide v0.1.0: kernels GPU en Rust puro sin DSL ni FFI