# AMD GPU

Published articles for AMD GPU.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## AMD ROCm HRX's Loom 0.2 Expands C++ Support, Improved Targeting

DevFeed: [AMD ROCm HRX's Loom 0.2 Expands C++ Support, Improved Targeting](<https://devfeed.tech/articles/amd-rocm-hrx-s-loom-0-2-expands-c-support-improved-targeting-136380.md>)

Original publisher: [Read original article](<https://www.phoronix.com/news/HRX-Loom-0.2>)

Author: Michael Larabel

Published: 2026-11-17T01:37:15Z

Content type: release

Language: en

Sources: [Phoronix](<https://devfeed.tech/sources/phoronix.md>)

Topics: [rocm](<https://devfeed.tech/topics/rocm.md>), [c/c++](<https://devfeed.tech/topics/c-c-plus-plus.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agent-workflows](<https://devfeed.tech/tags/agent-workflows.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [amd](<https://devfeed.tech/tags/amd.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [authoring](<https://devfeed.tech/tags/authoring.md>), [c-plus-plus](<https://devfeed.tech/tags/c-plus-plus.md>), [capabilities](<https://devfeed.tech/tags/capabilities.md>), [compilation](<https://devfeed.tech/tags/compilation.md>), [compiler](<https://devfeed.tech/tags/compiler.md>), [compute](<https://devfeed.tech/tags/compute.md>), [desktop-linux](<https://devfeed.tech/tags/desktop-linux.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [linux-benchmarking](<https://devfeed.tech/tags/linux-benchmarking.md>), [linux-hardware-benchmarks](<https://devfeed.tech/tags/linux-hardware-benchmarks.md>), [linux-hardware-reviews](<https://devfeed.tech/tags/linux-hardware-reviews.md>), [linux-how-to](<https://devfeed.tech/tags/linux-how-to.md>), [linux-performance](<https://devfeed.tech/tags/linux-performance.md>), [linux-server-benchmarks](<https://devfeed.tech/tags/linux-server-benchmarks.md>), [loom](<https://devfeed.tech/tags/loom.md>), [open-source-graphics](<https://devfeed.tech/tags/open-source-graphics.md>), [phoronix](<https://devfeed.tech/tags/phoronix.md>), [phoronix-test-suite](<https://devfeed.tech/tags/phoronix-test-suite.md>), [rocm](<https://devfeed.tech/tags/rocm.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [serving](<https://devfeed.tech/tags/serving.md>), [ubuntu-benchmarks](<https://devfeed.tech/tags/ubuntu-benchmarks.md>), [ubuntu-hardware](<https://devfeed.tech/tags/ubuntu-hardware.md>), [version](<https://devfeed.tech/tags/version.md>)

### AI overview

Loom 0.2 broadens complete-program C++ support and packaged AMD GPU/NPU and Vulkan execution, while expanding a shared target compiler across x86, AIE2P, AMDGPU, SPIR-V, and WebAssembly. The release describes changes aimed at removing fixed and superlinear costs in JIT compilation, graph construction and updates, GPU launch and replay, pipeline planning, and long-context serving. AMD hardware tooling now uses one libamdf-backed HAL via `--device=amdgpu`; the native path supports RDNA/CDNA, shared CPU/GPU/NPU memory contracts, explicit compute-resource selection, cooperative execution, virtual mappings, and reusable command buffers. The former ROCr-backed driver and its separate runtime and memory ownership paths are removed.

### Source excerpt

AMD ROCm's HRX Loom as the compiler, runtime, and serving infrastructure for supporting AI agent workflows is out today with a new version...

## Local AI Weekly #8: All Your RAM Are Belong to Us

DevFeed: [Local AI Weekly #8: All Your RAM Are Belong to Us](<https://devfeed.tech/articles/local-ai-weekly-8-all-your-ram-are-belong-to-us-117756.md>)

Original publisher: [Read original article](<https://feed.itsfoss.com/link/24361/17515249/local-ai-weekly-8>)

Author: Abhishek Prakash

Published: 2026-11-04T14:11:07Z

Content type: opinion

Language: en

Sources: [It's FOSS](<https://devfeed.tech/sources/it-s-foss.md>)

Topics: [Local AI](<https://devfeed.tech/topics/local-ai.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [Instagram](<https://devfeed.tech/topics/instagram.md>)

Tags: [7-4](<https://devfeed.tech/tags/7-4.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-tools](<https://devfeed.tech/tags/ai-tools.md>), [alphabet](<https://devfeed.tech/tags/alphabet.md>), [amazon](<https://devfeed.tech/tags/amazon.md>), [amd](<https://devfeed.tech/tags/amd.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [api](<https://devfeed.tech/tags/api.md>), [data-center](<https://devfeed.tech/tags/data-center.md>), [local](<https://devfeed.tech/tags/local.md>), [local-ai](<https://devfeed.tech/tags/local-ai.md>), [local-ai-weekly](<https://devfeed.tech/tags/local-ai-weekly.md>), [newsletter](<https://devfeed.tech/tags/newsletter.md>), [pc](<https://devfeed.tech/tags/pc.md>), [ram](<https://devfeed.tech/tags/ram.md>), [sam](<https://devfeed.tech/tags/sam.md>), [sam-altman](<https://devfeed.tech/tags/sam-altman.md>), [weekly](<https://devfeed.tech/tags/weekly.md>)

### AI overview

The article argues that idle RAM in a personal computer is a tradeoff for owning it: unlike data-center capacity, it cannot be rented out from under the owner, repriced, or switched off. It highlights Superwhisper's new native Linux app, which offers local CPU/GPU transcription or faster cloud processing but is not open source, and MIT-licensed voiceio, which provides offline dictation for Wayland and X11 and adopts a fine-tuned model only if it scores better. OpenKruise Agents keeps API keys outside agent sandboxes by retrieving them from Kubernetes Secrets through an egress proxy, with network rules and an MCP-tool allowlist. The excerpt also reports that Linux 7.4-rc1 received many AI-found bug fixes, while noting Torvalds was unsure they were all AI-discovered. It defines the KV cache as per-token working data that grows with context and can cause long chats or documents to exceed memory.

### Source excerpt

Sam Altman thinks the RAM in your PC would be better off in a data center.

## Linux 7.4's Merge Window is Huge, and Linus Says it Might Not Just Be AI

DevFeed: [Linux 7.4's Merge Window is Huge, and Linus Says it Might Not Just Be AI](<https://devfeed.tech/articles/linux-7-4-s-merge-window-is-huge-and-linus-says-it-might-not-just-be-ai-116436.md>)

Original publisher: [Read original article](<https://feed.itsfoss.com/link/24361/17515167/linux-7-4-huge-merge-window>)

Author: Sourav Rudra

Published: 2026-11-04T10:55:00Z

Content type: news

Language: en

Sources: [It's FOSS](<https://devfeed.tech/sources/it-s-foss.md>)

Topics: [Linux](<https://devfeed.tech/topics/linux.md>), [Linux Kernel](<https://devfeed.tech/topics/linux-kernel.md>)

Tags: [7-4](<https://devfeed.tech/tags/7-4.md>), [ai](<https://devfeed.tech/tags/ai.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [bug](<https://devfeed.tech/tags/bug.md>), [bug-fixes](<https://devfeed.tech/tags/bug-fixes.md>), [development](<https://devfeed.tech/tags/development.md>), [drivers](<https://devfeed.tech/tags/drivers.md>), [filesystems](<https://devfeed.tech/tags/filesystems.md>), [linux](<https://devfeed.tech/tags/linux.md>), [linux-7-4](<https://devfeed.tech/tags/linux-7-4.md>), [lts](<https://devfeed.tech/tags/lts.md>), [memory-management](<https://devfeed.tech/tags/memory-management.md>), [merge](<https://devfeed.tech/tags/merge.md>), [networking](<https://devfeed.tech/tags/networking.md>), [news](<https://devfeed.tech/tags/news.md>), [patches](<https://devfeed.tech/tags/patches.md>), [window](<https://devfeed.tech/tags/window.md>)

### AI overview

Linux 7.4's merge window set a record in Phoronix's comparison of every window since Linux 3.1: 19,671 commits, up from Linux 7.3's 16,267. Linus Torvalds attributes some of the increase to AI-assisted bug fixes and suggested patches, which are often rewritten; he prefers this work to arrive during the merge window rather than as late release-candidate fixes. He also cites expectations that 7.4 will be the next LTS release as a possible minor factor. AMD GPU register-definition headers account for about a third of the patch, and drivers about half--or roughly two thirds including those headers. Other areas include networking, Rust, BPF, memory management, filesystems, tools, and architecture updates.

### Source excerpt

Another possible factor is that people expect 7.4 to be the next LTS release.

## Linux 7.4-rc1 Released With Nice Performance Improvements, AMD Ryzen Zen 6 Preparations & AMDGPU HDMI 2.1 Default

DevFeed: [Linux 7.4-rc1 Released With Nice Performance Improvements, AMD Ryzen Zen 6 Preparations & AMDGPU HDMI 2.1 Default](<https://devfeed.tech/articles/linux-7-4-rc1-released-with-nice-performance-improvements-amd-ryzen-zen-6-preparations-amdgpu-hdmi-2-1-default-111137.md>)

Original publisher: [Read original article](<https://www.phoronix.com/news/Linux-7.4-rc1>)

Author: Michael Larabel

Published: 2026-11-01T21:51:34Z

Content type: release

Language: en

Sources: [Phoronix](<https://devfeed.tech/sources/phoronix.md>)

Topics: [Linux Kernel](<https://devfeed.tech/topics/linux-kernel.md>), [Linux](<https://devfeed.tech/topics/linux.md>), [Ext4](<https://devfeed.tech/topics/ext4.md>), [hdmi](<https://devfeed.tech/topics/hdmi.md>)

Tags: [7-4](<https://devfeed.tech/tags/7-4.md>), [ai](<https://devfeed.tech/tags/ai.md>), [amd](<https://devfeed.tech/tags/amd.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [amd-ryzen](<https://devfeed.tech/tags/amd-ryzen.md>), [apple](<https://devfeed.tech/tags/apple.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [build](<https://devfeed.tech/tags/build.md>), [desktop-linux](<https://devfeed.tech/tags/desktop-linux.md>), [hdmi](<https://devfeed.tech/tags/hdmi.md>), [holiday](<https://devfeed.tech/tags/holiday.md>), [improvements](<https://devfeed.tech/tags/improvements.md>), [linus-torvalds](<https://devfeed.tech/tags/linus-torvalds.md>), [linux](<https://devfeed.tech/tags/linux.md>), [linux-7-4](<https://devfeed.tech/tags/linux-7-4.md>), [linux-benchmarking](<https://devfeed.tech/tags/linux-benchmarking.md>), [linux-hardware-benchmarks](<https://devfeed.tech/tags/linux-hardware-benchmarks.md>), [linux-hardware-reviews](<https://devfeed.tech/tags/linux-hardware-reviews.md>), [linux-how-to](<https://devfeed.tech/tags/linux-how-to.md>), [linux-performance](<https://devfeed.tech/tags/linux-performance.md>), [linux-server-benchmarks](<https://devfeed.tech/tags/linux-server-benchmarks.md>), [merge](<https://devfeed.tech/tags/merge.md>), [open-source-graphics](<https://devfeed.tech/tags/open-source-graphics.md>), [performance](<https://devfeed.tech/tags/performance.md>), [phoronix](<https://devfeed.tech/tags/phoronix.md>), [phoronix-test-suite](<https://devfeed.tech/tags/phoronix-test-suite.md>), [ryzen](<https://devfeed.tech/tags/ryzen.md>), [schedule](<https://devfeed.tech/tags/schedule.md>), [ubuntu-benchmarks](<https://devfeed.tech/tags/ubuntu-benchmarks.md>), [ubuntu-hardware](<https://devfeed.tech/tags/ubuntu-hardware.md>), [window](<https://devfeed.tech/tags/window.md>), [zen-6](<https://devfeed.tech/tags/zen-6.md>)

### AI overview

Linux 7.4-rc1 arrived on schedule with a broad set of changes, including AMD PerfOpt for Radeon integrated GPUs aimed at improving AI/LLM performance, Zen 6 preparation, AMDGPU HDMI 2.1 enabled by default, and support work for upcoming Intel and Qualcomm laptops and Google Tensor G5 and Pixel 10. It also adds optional buffered IOmap support for EXT4, basic Apple M4 enablement that is not yet ready for end users, and faster kernel build times. Linus Torvalds called this an unusually large merge window, with nearly 20,000 commits versus a more typical roughly 15,000; he attributed much of the growth to AI-related patches, while noting possible contribution from expectations that 7.4 will be the next LTS kernel. The excerpt says a timely release could arrive December 20 and become this year's LTS, but both are projections.

### Source excerpt

Even with Linus Torvalds having been on holiday, Linux 7.4-rc1 is out today on-schedule for capping off the Linux 7.4 merge window...

## vLLM vs. SGLang: Choosing the right inference engine for your workload

DevFeed: [vLLM vs. SGLang: Choosing the right inference engine for your workload](<https://devfeed.tech/articles/vllm-vs-sglang-choosing-the-right-inference-engine-for-your-workload-86632.md>)

Original publisher: [Read original article](<https://developers.redhat.com/articles/2026/10/14/vllm-vs-sglang-choosing-the-right-inference-engine-for-your-workload>)

Author: Cedric Clyburn

Published: 2026-10-14T13:01:50Z

Content type: comparison

Language: en

Sources: [Red Hat](<https://devfeed.tech/sources/red-hat.md>), [Red Hat Developer](<https://devfeed.tech/sources/red-hat-developer.md>)

Topics: [vllm](<https://devfeed.tech/topics/vllm.md>), [jetstream](<https://devfeed.tech/topics/jetstream.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [quantization](<https://devfeed.tech/topics/quantization.md>)

Tags: [ai-applications](<https://devfeed.tech/tags/ai-applications.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [batching](<https://devfeed.tech/tags/batching.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [inference](<https://devfeed.tech/tags/inference.md>), [llms](<https://devfeed.tech/tags/llms.md>), [open-weight-models](<https://devfeed.tech/tags/open-weight-models.md>), [openrouter](<https://devfeed.tech/tags/openrouter.md>), [production](<https://devfeed.tech/tags/production.md>), [sglang](<https://devfeed.tech/tags/sglang.md>), [token-usage](<https://devfeed.tech/tags/token-usage.md>), [vercel](<https://devfeed.tech/tags/vercel.md>), [vllm](<https://devfeed.tech/tags/vllm.md>), [vs](<https://devfeed.tech/tags/vs.md>)

### AI overview

The article compares vLLM and SGLang, two open source engines for serving large language models. It explains how their KV cache approaches differ, reviews shared capabilities and hardware support, and summarizes benchmark results that show performance depends on workload and configuration. It recommends testing with representative traffic to choose an engine.

### Source excerpt

Open models have gone from "worth testing" to powering the majority of AI applications and agents. On Vercel's AI Gateway, open weight models processed almost 80% of production tokens in September (Figure 1), up from less than 10% in December 2025. OpenRouter, a popular gateway for developers using LLMs, shows a similar climb in weekly token usage. The post vLLM vs. SGLang: Choosing the right inference engine for your workload appeared first on Red Hat Developer.

## vime x RL-Kernel x AMD: Bitwise Train-Rollout Consistency on ROCm

DevFeed: [vime x RL-Kernel x AMD: Bitwise Train-Rollout Consistency on ROCm](<https://devfeed.tech/articles/vime-rl-kernel-amd-bitwise-train-rollout-consistency-on-rocm-79486.md>)

Original publisher: [Read original article](<https://vllm.ai/blog/2026-09-14-rl-kernel-v0-1-0-train-rollout-bitwise-consistency>)

Author: RL-Kernel Team, vime Team, and AMD Team

Published: 2026-09-14T00:00:00Z

Content type: article

Language: en

Sources: [vLLM Blog](<https://devfeed.tech/sources/vllm-blog.md>)

Topics: [rocm](<https://devfeed.tech/topics/rocm.md>), [grpo](<https://devfeed.tech/topics/grpo.md>), [vllm](<https://devfeed.tech/topics/vllm.md>), [Kernel](<https://devfeed.tech/topics/kernel.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [Linux Kernel](<https://devfeed.tech/topics/linux-kernel.md>)

Tags: [ai-engineering](<https://devfeed.tech/tags/ai-engineering.md>), [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [attention](<https://devfeed.tech/tags/attention.md>), [ecosystem](<https://devfeed.tech/tags/ecosystem.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [instinct](<https://devfeed.tech/tags/instinct.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [llm-inference](<https://devfeed.tech/tags/llm-inference.md>), [model-serving](<https://devfeed.tech/tags/model-serving.md>), [open-source-ai](<https://devfeed.tech/tags/open-source-ai.md>), [performance](<https://devfeed.tech/tags/performance.md>), [post-training](<https://devfeed.tech/tags/post-training.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [rl](<https://devfeed.tech/tags/rl.md>), [rocm](<https://devfeed.tech/tags/rocm.md>), [token](<https://devfeed.tech/tags/token.md>), [train](<https://devfeed.tech/tags/train.md>), [training](<https://devfeed.tech/tags/training.md>), [vllm](<https://devfeed.tech/tags/vllm.md>), [vllm-blog](<https://devfeed.tech/tags/vllm-blog.md>)

### AI overview

vime coordinates Megatron training, vLLM rollout and weight synchronization, while RL-Kernel aligns numerical execution so selected-token logprobs match bit for bit on AMD Instinct MI300X. The article reports zero mismatches and zero maximum absolute difference across a 200-step Qwen3-8B GRPO run, and describes ROCm-specific kernel, attention, communication and execution changes used to achieve that result.

### Source excerpt

vime and RL-Kernel align selected-token logprobs bit for bit across Megatron training and vLLM rollout on AMD Instinct MI300X, with zero mismatches across 200 GRPO steps.

## Training Flux.1 Dev on MI300X with Massive Batch Sizes

DevFeed: [Training Flux.1 Dev on MI300X with Massive Batch Sizes](<https://devfeed.tech/articles/training-flux-1-dev-on-mi300x-with-massive-batch-sizes-78158.md>)

Original publisher: [Read original article](<https://www.runpod.io/blog/training-flux-mi300x>)

Author: Sean Sube

Published: 2026-09-13T15:06:30Z

Content type: tutorial

Language: en

Sources: [Runpod Blog.](<https://devfeed.tech/sources/runpod-blog.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [AWS Trainium](<https://devfeed.tech/topics/aws-trainium.md>), [peft](<https://devfeed.tech/topics/peft.md>), [Amazon SageMaker HyperPod](<https://devfeed.tech/topics/amazon-sagemaker-hyperpod.md>)

Tags: [amd-gpu](<https://devfeed.tech/tags/amd-gpu.md>), [batch](<https://devfeed.tech/tags/batch.md>), [build-tools](<https://devfeed.tech/tags/build-tools.md>), [cloudflare](<https://devfeed.tech/tags/cloudflare.md>), [comfyui](<https://devfeed.tech/tags/comfyui.md>), [container-image](<https://devfeed.tech/tags/container-image.md>), [cuda](<https://devfeed.tech/tags/cuda.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [gpu](<https://devfeed.tech/tags/gpu.md>)

### AI overview

A tutorial shows how to build a Docker training container and fine-tune Flux.1 Dev LoRAs on Runpod's AMD MI300X GPUs. It covers the training setup, configurable parameters, deployment, and retrieving and testing checkpoints. In the reported experiment, 400 images were used at 1024 target resolution, with peak VRAM measurements used to identify a batch size of about 60 images on one MI300X.

### Source excerpt

Explore what's possible when training Flux.1 Dev on AMD's 192GB MI300X GPU. This post dives into fine-tuning at scale with huge batch sizes and real-world.