# GPU Published articles for 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. ## Experimental Nvidia GPU support on Dell Latitude E6400 variants, plus E6400 XFR support now confirmed DevFeed: [Experimental Nvidia GPU support on Dell Latitude E6400 variants, plus E6400 XFR support now confirmed]() Original publisher: [Read original article]() Author: Leah Rowe Published: 2026-09-17T04:32:50.666044Z Content type: news Language: en Sources: [News about Libreboot releases and development]() Topics: [dell](), [GPU](), [Nvidia](), [Linux]() Tags: [bios](), [canoeboot](), [coreboot](), [dell](), [experimental](), [free-software](), [gpu](), [libre](), [libreboot](), [news](), [nvidia](), [opensource](), [uefi]() ### AI overview The article reports experimental Nvidia GPU support for Dell Latitude E6400 variants in Libreboot, alongside confirmed support for the Intel-graphics Dell Latitude E6400 XFR. It explains that Nvidia variants require the appropriate VGA ROM and identifies Libreboot release versions associated with support. ### Source excerpt Article: Experimental Nvidia GPU support on Dell Latitude E6400 variants, plus E6400 XFR support now confirmed Web link: https://libreboot.org/news/e6400nvidia.html ## Nvidia goes green to keep grid capacity from zapping its revenues DevFeed: [Nvidia goes green to keep grid capacity from zapping its revenues]() Original publisher: [Read original article]() Author: Tobias Mann Published: 2026-09-16T21:37:05Z Content type: news Language: en Sources: [www.theregister.com - Articles]() Topics: [Nvidia](), [datacenter](), [Cloud]() Tags: [ai](), [ai-factory](), [cloud](), [datacenter](), [gpu](), [nvidia](), [systems]() ### AI overview The article reports that Nvidia is courting neocloud providers with more efficient and profitable data centers. ### Source excerpt GPUzilla woos neoclouds into another walled garden, promising smarter, more efficient, and profitable bit barns ## Mesa 26.2.3 Released With AMD GFX1171 Support, Other Fixes DevFeed: [Mesa 26.2.3 Released With AMD GFX1171 Support, Other Fixes]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-09-16T21:18:50Z Content type: news Language: en Sources: [Phoronix]() Topics: [GPU](), [patches](), [Adreno](), [intel](), [Qualcomm](), [Firefox](), [Memory Leaks](), [Flight]() Tags: [adreno](), [amd](), [browser](), [bug](), [bug-fixes](), [desktop-linux](), [firefox](), [gpu](), [intel](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [memory-leak](), [open-source-graphics](), [patches](), [phoronix](), [phoronix-test-suite](), [qualcomm](), [release](), [stable](), [ubuntu-benchmarks](), [ubuntu-hardware](), [vulkan]() ### AI overview Mesa 26.2.3 is a stable bi-weekly point release containing backported bug fixes for RadeonSI, RADV, Freedreno, Turnip, Intel ANV, Asahi, NVK, and other drivers. It also adds AMD GFX1171 support. ### Source excerpt Mesa 26.2.3 is out today as the newest stable point release in this quarter's Mesa 26.2 series... ## Fault tolerant distributed training on Amazon EKS using NVRx DevFeed: [Fault tolerant distributed training on Amazon EKS using NVRx]() Original publisher: [Read original article]() Author: Aravind Neelakantan Published: 2026-09-16T18:59:25Z Content type: tutorial Language: en Sources: [Artificial Intelligence]() Topics: [NCCL](), [Amazon Elastic Kubernetes Service](), [PyTorch](), [distributed-training](), [Resilience](), [GPU]() Tags: [advanced-300](), [amazon-eks](), [amazon-elastic-kubernetes-service](), [async](), [benchmarks](), [distributed-training](), [gpu](), [pytorch](), [resiliency](), [technical-how-to]() ### AI overview This tutorial integrates NVIDIA Resiliency Extension (NVRx) with PyTorch FSDP training on Amazon EKS. It covers asynchronous checkpointing, in-process restart, and in-job restart, and reports H100 benchmarks at 2- to 8-node scale with 99%+ training efficiency and recovery measured in seconds. ### Source excerpt Integrate NVIDIA Resiliency Extension (NVRx) into PyTorch FSDP training on Amazon EKS to overlap checkpoint I/O with training and recover from GPU faults in seconds. This post covers async checkpointing, in-process restart, and ft_launcher in-job restart, with H100 benchmarks at 2 to 8 nodes showing 99%+ training efficiency and second-scale recovery. ## Microsoft Open-Sources TauGrid to Simplify AI Workload Management on Kubernetes DevFeed: [Microsoft Open-Sources TauGrid to Simplify AI Workload Management on Kubernetes]() Original publisher: [Read original article]() Author: Sergio De Simone Published: 2026-09-16T18:00:00Z Content type: news Language: en Sources: [InfoQ]() Topics: [Artificial Intelligence](), [Kubernetes](), [GPU](), [kueue](), [Go]() Tags: [agents](), [ai](), [artificial-intelligence](), [azure](), [development](), [devops](), [gpu](), [kubernetes](), [kueue](), [microsoft](), [microsoft-taugrid-open-source](), [news](), [open-source](), [orchestration]() ### AI overview Microsoft has open-sourced TauGrid, a cloud-native platform for managing, scheduling, and monitoring AI workloads on GPU-enabled Kubernetes clusters. It combines workload submission, Kueue-based queuing, KubeRay orchestration, GPU-node monitoring, and observability, while planned capabilities remain on its roadmap. ### Source excerpt Microsoft has open-sourced TauGrid, a cloud-native platform designed to manage, schedule, and monitor AI workloads on GPU-enabled Kubernetes clusters. By Sergio De Simone ## Translating CUDA Tile Operations from Python to Rust Using Agentic AI DevFeed: [Translating CUDA Tile Operations from Python to Rust Using Agentic AI]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-09-16T16:28:59Z Content type: tutorial Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [CUDA Tile](), [Rust](), [Agent Skill](), [Artificial Intelligence](), [AI Agent](), [GPU](), [benchmarking](), [Compiler]() Tags: [agent-skill](), [agentic-ai](), [agentic-ai-generative-ai](), [ai-agent](), [benchmarking](), [cuda](), [cuda-tile](), [cutile](), [developer-tools-techniques](), [gpu](), [mixture-of-experts-moe](), [python](), [rust]() ### AI overview This NVIDIA developer article explains a multi-agent workflow for translating cuTile Python and Triton-TileIR GPU kernels into cuTile Rust. The team ported 24 public TileGym operators, covering about 40 kernels, and achieved 99.5% of cuTile Python performance on average, with correctness and performance checks at each stage. ### Source excerpt cuTile Rust (cutile-rs) is a tile-based system for safe, idiomatic GPU kernel authoring in the Rust programming language. Extending the Rust ownership model to... ## MLPerf Inference v6.1: 5.7x Per-Accelerator Gains, a 512-GPU Run, and Vera Rubin's First Peer-Reviewed Numbers DevFeed: [MLPerf Inference v6.1: 5.7x Per-Accelerator Gains, a 512-GPU Run, and Vera Rubin's First Peer-Reviewed Numbers]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-09-16T15:00:00Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [Benchmark](), [Inference Performance](), [Retrieval Augmented Generation (RAG)](), [agentic-coding](), [GPU](), [Vera Rubin NVL72](), [Nvidia](), [Vera Rubin]() Tags: [agentic-coding](), [ai](), [benchmark](), [enterprise](), [gpt-oss](), [gpu](), [inference](), [llama-cpp](), [numbers](), [nvidia](), [qwen3](), [rag]() ### AI overview MLCommons published MLPerf Inference v6.1 with record participation, two new inference tests, and peer-reviewed results for several newly covered accelerators. The release reports a 5.7x improvement in the best per-accelerator DeepSeek-R1 server result compared with v5.1. ### Source excerpt MLCommons has published MLPerf Inference v6.1, and the round sets a participation record with 30 submitting organizations and 486 datacenter and edge results. Two new tests join the suite: an End-to-End RAG pipeline for the datacenter and an Edge Agentic Inference benchmark for single-user devices, and the results carry the first peer-reviewed numbers for NVIDIA's The post MLPerf Inference v6.1: 5.7x Per-Accelerator Gains, a 512-GPU Run, and Vera Rubin's First Peer-Reviewed Numbers appeared first on StorageReview.com. ## ADLX 2.0: Extending graphics control to AI agents and agentic apps DevFeed: [ADLX 2.0: Extending graphics control to AI agents and agentic apps]() Original publisher: [Read original article]() Author: Pete Vagiakos; Alexander Blake-Davies Published: 2026-09-16T10:00:00Z Content type: article Language: en Sources: [AMD GPUOpen]() Topics: [GPU](), [Artificial Intelligence](), [Hardware](), [SDK](), [Model Context Protocol]() Tags: [agents](), [ai](), [amd-device-library-extra](), [amd-device-library-extra-adlx](), [announcement](), [developers](), [gpu](), [hardware](), [mcp](), [ml](), [product-blogs](), [product-release](), [tools]() ### AI overview AMD ADLX 2.0 adds an AI extension framework and MCP servers that connect AI applications with AMD graphics technologies, enabling developers to build apps that monitor, manage, and optimize AMD graphics hardware. ### Source excerpt AMD ADLX 2.0 adds AI extensions and MCP servers to help developers build intelligent apps that can monitor, manage, and optimize AMD graphics hardware. ## HP ZBook Ultra G3a 16 Preview: 192GB of Unified Memory Aims for the Top of the Local AI Laptop Leaderboard DevFeed: [HP ZBook Ultra G3a 16 Preview: 192GB of Unified Memory Aims for the Top of the Local AI Laptop Leaderboard]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-09-15T23:15:57Z Content type: article Language: en Sources: [StorageReview.com]() Topics: [Local AI](), [Hardware](), [GPU](), [cpu](), [Large Language Model]() Tags: [consumer](), [cpu](), [deepseek](), [dgx-spark](), [gemma](), [gpu](), [hardware](), [local-ai](), [models](), [ollama](), [windows](), [workstation]() ### AI overview StorageReview previews HP's pre-production ZBook Ultra G3a 16, a local AI laptop with 192GB of unified memory and up to 160GB assignable to its integrated GPU. The article examines its hardware and planned testing while noting that shipping-hardware benchmarks are not yet available. ### Source excerpt HP's ZBook Ultra G1a 14 holds the Best for Large Models spot on our Best Laptops for Local AI leaderboard because its 128GB of unified memory, 96GB of it assignable to the GPU, loaded models no discrete-GPU laptop could touch. The new HP ZBook Ultra G3a 16 raises that pool to 192GB with up to The post HP ZBook Ultra G3a 16 Preview: 192GB of Unified Memory Aims for the Top of the Local AI Laptop Leaderboard appeared first on StorageReview.com. ## Optimizing CPU-side Rendering Code DevFeed: [Optimizing CPU-side Rendering Code]() Original publisher: [Read original article]() Author: Clay John Published: 2026-09-15T17:00:00Z Content type: tutorial Language: en Sources: [Godot Engine Official]() Topics: [Godot](), [Optimization](), [cpu](), [GPU](), [shaders]() Tags: [batching](), [bug](), [cpu](), [gpu](), [optimization](), [performance](), [performance-optimization](), [progress-report](), [shaders]() ### AI overview This article explains how Godot optimizes CPU-side rendering code. It describes balancing CPU and GPU workloads, identifying performance bottlenecks, investigating solutions, measuring results, and repeating the process. ### Source excerpt Optimizing CPU code is a lot of fun. Here's how we do it ## Announcing instance preference lists for Amazon SageMaker AI training jobs DevFeed: [Announcing instance preference lists for Amazon SageMaker AI training jobs]() Original publisher: [Read original article]() Author: Kanwaljit Khurmi Published: 2026-09-15T16:01:47Z Content type: release Language: en Sources: [Artificial Intelligence]() Topics: [Amazon SageMaker AI](), [GPU](), [Model Development](), [data-processing]() Tags: [amazon-sagemaker-ai](), [announcements](), [gpu](), [intermediate-200](), [model-development](), [pipelines](), [scheduled](), [training]() ### AI overview Amazon SageMaker AI introduces instance preference lists for training and processing jobs. Users can specify up to five instance types in priority order, and SageMaker AI launches the job on the first option with available capacity, reducing manual retries and capacity monitoring. ### Source excerpt Amazon SageMaker AI now offers instance preference lists for training and processing jobs. Specify an ordered list of up to five instance types, and SageMaker AI automatically launches on the first type with available capacity, eliminating manual retry loops and capacity-watching scripts. ## Scaling Federated Learning Across Docker, Kubernetes, and Slurm with NVIDIA FLARE DevFeed: [Scaling Federated Learning Across Docker, Kubernetes, and Slurm with NVIDIA FLARE]() Original publisher: [Read original article]() Author: Elizabeth Goodman Published: 2026-09-15T15:00:00Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Federated Learning](), [Docker](), [Kubernetes](), [Nvidia](), [Server](), [dataset](), [Deployment]() Tags: [architecture](), [compute](), [configuration](), [container](), [data-analytics-processing](), [data-science](), [dataset](), [deployment](), [docker](), [docker-container](), [federated-learning](), [gpu](), [job](), [kubernetes](), [nvidia](), [nvidia-flare](), [server]() ### AI overview This article explains how NVIDIA FLARE scales federated learning across sites with different infrastructure, including Docker, Kubernetes, and Slurm. Its two-layer architecture separates persistent federation services from on-demand job execution, while allowing each site to retain local control over compute, data, secrets, and scheduling. ### Source excerpt Federated learning (FL) projects often begin with a straightforward setup: one server, a few clients, and one dataset at each site. As those projects grow, the... ## Announcing Fedora Linux Asahi Remix 45 Beta DevFeed: [Announcing Fedora Linux Asahi Remix 45 Beta]() Original publisher: [Read original article]() Author: Davide Cavalca Published: 2026-09-15T14:05:00Z Content type: release Language: en Sources: [Fedora Magazine]() Topics: [Remix](), [Fedora](), [Linux](), [Hardware](), [GPU](), [dnf]() Tags: [announce](), [apple](), [bugs](), [dnf](), [features](), [fedora-project-community](), [gpu](), [h-264](), [hardware](), [installation](), [linux](), [pre-release](), [release](), [remix](), [report](), [update](), [vp9]() ### AI overview Fedora Asahi Remix 45 Beta is available for Apple Silicon Macs, bringing Fedora Linux 45 Beta and initial out-of-the-box support for Apple M3, M3 Pro, and M3 Max systems. It also adds hardware-accelerated H.264 and VP9 decoding across Apple Silicon, with AV1 decoding on M3 systems. Performance-oriented 3D GPU drivers for M3 systems are deferred to a future Mesa update. ### Source excerpt We are happy to announce the availability of Fedora Linux Asahi Remix 45 Beta. This pre-release will bring the freshly announced Fedora Linux 45 Beta to Apple Silicon Macs. We expect to announce general availability of Fedora Linux Asahi Remix 45 in about a month. This will coincide with the overall Fedora Linux 45 release. [...] ## How Everpure proposes reducing GPU idle time by improving AI data access DevFeed: [How Everpure proposes reducing GPU idle time by improving AI data access]() Original publisher: [Read original article]() Author: Chris Mellor Published: 2026-09-15T08:00:00Z Content type: article Language: en Sources: [www.theregister.com - Articles]() Topics: [AI Infrastructure](), [GPU](), [Large Language Model](), [Retrieval Augmented Generation (RAG)](), [AI Agent](), [data-processing](), [data]() Tags: [agents](), [ai](), [ai-infrastructure](), [ai-ml](), [data-processing](), [gpu](), [hardware](), [rag](), [sponsored]() ### AI overview This sponsored feature describes Everpure's approach to reducing GPU idle time in AI systems by improving access to large-scale insurance data. It discusses central metadata indexing, storage performance, self-describing data, and integration with Nvidia GPU infrastructure for AI agents and retrieval-augmented generation. ### Source excerpt SPONSORED FEATURE: The vendor's AI solutions are dedicated to increasing GPU utilization and avoiding costly GPUs doing nothing while waiting for data ## Monitor TAS and gang scheduling for AI training in Kubernetes DevFeed: [Monitor TAS and gang scheduling for AI training in Kubernetes]() Original publisher: [Read original article]() Author: David Lentz; Kathy Lin Published: 2026-09-15T00:00:00Z Content type: article Language: en Sources: [Datadog | The Monitor blog]() Topics: [Kubernetes](), [kueue](), [distributed-training](), [GPU](), [Monitoring](), [datadog]() Tags: [ai](), [batch](), [containers](), [datadog](), [distributed-training](), [gpu](), [gpu-monitoring](), [kubernetes](), [kueue](), [monitoring](), [scheduler]() ### AI overview This article explains why Kubernetes scheduling is insufficient for distributed AI training workloads and how topology-aware scheduling and gang scheduling address hardware placement and simultaneous startup requirements. It discusses implementing these capabilities with Kueue and the Coscheduling plugin, and monitoring and troubleshooting them with Datadog GPU Monitoring. ### Source excerpt Learn how Datadog helps you correlate Kueue, Coscheduling, GPU, and training framework signals to validate gang scheduling and topology-aware scheduling. ## How and Why We Bought 4x DGX Sparks DevFeed: [How and Why We Bought 4x DGX Sparks]() Original publisher: [Read original article]() Author: Alex Ellis Published: 2026-09-15T00:00:00Z Content type: opinion Language: en Sources: [Alex Ellis' Blog]() Topics: [Local AI](), [GPU](), [qwen](), [Cloud]() Tags: [agents](), [ai](), [cloud](), [gpu](), [llm](), [local-ai](), [localai](), [qwen](), [red-teaming](), [security]() ### AI overview The author explains why OpenFaaS Ltd bought four DGX Sparks and what the team learned from deploying local AI. The article argues that local infrastructure can provide tangible privacy and risk-reduction benefits for business use cases, even though it is not primarily justified by cost per token. ### Source excerpt In June we deployed an RTX 6000 Pro into production, a few weeks later, we're now operating DGX Sparks for the team. Learn how and why. ## MSI XpertStation WS300 Thermals: Why a 1,300W GB300 Doesn't Throttle on a Desk DevFeed: [MSI XpertStation WS300 Thermals: Why a 1,300W GB300 Doesn't Throttle on a Desk]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-09-14T19:41:35Z Content type: article Language: en Sources: [StorageReview.com]() Topics: [Blackwell](), [cpu](), [GPU](), [Grace CPU](), [Nvidia](), [DGX Station]() Tags: [blackwell](), [consumer](), [dgx-station](), [gpu](), [grace-cpu](), [heat](), [memory](), [nvidia](), [performance](), [review](), [workstation]() ### AI overview The article explains why MSI's XpertStation WS300 can sustain a 1,300W GB300 Grace Blackwell Ultra Superchip without throttling on a desk. It attributes the thermal stability to cold plates covering the major heat-producing components, dual 360mm radiators, multiple fans, and a cooling loop rated above the system's nominal CPU and GPU load. ### Source excerpt The most common question we got about the MSI XpertStation WS300 after our review coalesces around one key theme. The GB300 Grace Blackwell Ultra Superchip is a 1,300W part that normally lives in a liquid-cooled rack, so what happens to thermals when you put it in a tower? The concern is fair: a GB300 system The post MSI XpertStation WS300 Thermals: Why a 1,300W GB300 Doesn't Throttle on a Desk appeared first on StorageReview.com. ## Perplexity's new agent runs entirely on your GPU -- with one expensive catch DevFeed: [Perplexity's new agent runs entirely on your GPU -- with one expensive catch]() Original publisher: [Read original article]() Author: Amanda Caswell Published: 2026-09-14T18:21:44Z Content type: news Language: en Sources: [The New Stack]() Topics: [GPU](), [Nvidia](), [Windows](), [Orchestration](), [qwen](), [Cloud](), [browser](), [Ollama]() Tags: [ai-agents](), [ai-infrastructure](), [gpu](), [hardware](), [infrastructure](), [news](), [nvidia](), [ollama](), [orchestration](), [qwen](), [windows]() ### AI overview The article reports that Perplexity's Portable Computer, a local version of its Computer agent, is available in the Perplexity app for Windows on compatible Nvidia GeForce RTX and RTX PRO GPUs. It requires at least 24GB of VRAM and combines local models, orchestration, a browser, tool calling, and a proprietary SPACE sandbox. The article also discusses platform-specific engineering, external service connectors, and the boundary between local and cloud computing. ### Source excerpt Running an LLM on your PC is easy enough, but putting an agent to work there is a different story. The post Perplexity's new agent runs entirely on your GPU -- with one expensive catch appeared first on The New Stack. ## FS Pairs 1.6T Scale-Out Optics With 500 km Coherent Modules and a Handheld Toolkit for AI Fabrics DevFeed: [FS Pairs 1.6T Scale-Out Optics With 500 km Coherent Modules and a Handheld Toolkit for AI Fabrics]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-09-14T17:34:25Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [GPU](), [InfiniBand](), [Networks](), [Ethernet](), [datacenter](), [Nvidia](), [data]() Tags: [announcement](), [data-center](), [enterprise](), [ethernet](), [gpu](), [infiniband](), [networking](), [networks](), [nvidia]() ### AI overview FS presents a two-part optics portfolio for AI networking: 400G, 800G, and 1.6T Scale-Out transceivers for links within GPU clusters, plus 400G and 800G Scale-Across coherent modules for connecting clusters across sites up to 500 km. The announcement also introduces the BOX 5 Ultra handheld toolkit for configuring, validating, and monitoring transceivers from 100M to 1.6T. ### Source excerpt FS has organized its AI optics into a two-part portfolio: Scale-Out transceivers at 400G, 800G, and 1.6T for the links inside a GPU cluster, and Scale-Across coherent modules at 400G and 800G for stitching clusters together across sites at distances up to 500 km. The Scale-Out side covers Ethernet, RoCE, and InfiniBand fabrics between GPU The post FS Pairs 1.6T Scale-Out Optics With 500 km Coherent Modules and a Handheld Toolkit for AI Fabrics appeared first on StorageReview.com. ## Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine DevFeed: [Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-09-14T16:39:15Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Mixture of Experts (MoE)](), [Artificial Intelligence](), [GPU](), [Large Language Model](), [Nvidia](), [Routing (disambiguation)](), [GB200](), [Network](), [Python](), [deepseek](), [mixtral](), [qwen]() Tags: [agentic-ai-generative-ai](), [ai](), [communication](), [deepseek](), [developer-tools-techniques](), [gb200](), [gpu](), [mixtral](), [mixture-of-experts-moe](), [mlops](), [moe](), [networks](), [nvfp4](), [nvidia](), [performance](), [python](), [qwen](), [router](), [routing](), [tensors](), [token](), [tokens](), [training-ai-models]() ### AI overview This article explains how NVIDIA Transformer Engine and JAX kernel optimizations accelerate dropless Mixture of Experts (MoE) training. It describes bottlenecks from token routing, expert dispatch and gathering, all-to-all communication, and ragged expert matrix operations. In DeepSeek-V3 training on NVIDIA GB200, the optimized approach increased performance from 103 to 1,068 TFLOPS per GPU, a 10.4x improvement. ### Source excerpt Mixture of experts (MoE) has become one of the defining architectural trends in large-scale AI model training. DeepSeek, Qwen, and Mixtral are examples of MoE... ## Lightbits Inferra KV Cache Engine Claims 16x Session Density and 10M-Token Contexts DevFeed: [Lightbits Inferra KV Cache Engine Claims 16x Session Density and 10M-Token Contexts]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-09-14T16:23:21Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [Inference](), [Inference Performance](), [Orchestration](), [GPU](), [Multi-tenancy](), [sglang](), [TensorRT](), [vllm](), [Security]() Tags: [ai](), [cache](), [concurrent](), [enterprise](), [gpu](), [inference](), [latency](), [nvme](), [orchestration](), [security]() ### AI overview Lightbits Labs is introducing Inferra, a KV cache orchestration engine for AI inference. It virtualizes GPU memory across DRAM and NVMe storage, preserving attention states for long-context and multi-session workloads. Lightbits claims up to 16 times more concurrent sessions, more than 100 times lower latency than recomputation, and context windows of up to 10 million tokens. Inferra supports vLLM, TensorRT, and SGLang and includes tiering, predictive prefetching, tenant isolation, and encrypted data transfer. ### Source excerpt Lightbits Labs, the company that invented NVMe over TCP, is moving into inference software with Inferra, a KV cache orchestration engine that makes its public debut tomorrow, September 15, at the AI Infra Summit in Santa Clara. The software virtualizes GPU memory across DRAM and NVMe storage tiers and turns the KV cache into a The post Lightbits Inferra KV Cache Engine Claims 16x Session Density and 10M-Token Contexts appeared first on StorageReview.com. ## On-Device AI Series (Part 5): LiteRT-LM DevFeed: [On-Device AI Series (Part 5): LiteRT-LM]() Original publisher: [Read original article]() Author: Oğuzhan Aslan Published: 2026-09-14T05:59:12Z Content type: tutorial Language: en Sources: [ProAndroidDev - Medium]() Topics: [LiteRT](), [On-device AI](), [Large Language Model](), [Inference](), [Streaming](), [GPU](), [LLM evaluation / benchmarking](), [PyTorch]() Tags: [ai](), [android](), [android-development](), [benchmarking](), [gpu](), [inference](), [litert](), [llama](), [llm](), [multimodal](), [on-device](), [on-device-ai](), [programming](), [pytorch](), [qwen](), [software-development](), [streaming]() ### AI overview This tutorial explains LiteRT-LM for running large language models on-device. It covers the Engine/Session API, streaming output, system prompts, tool calling, multimodal inputs, thinking mode, and CPU-versus-GPU benchmarking. The article also discusses tradeoffs involving privacy, network independence, latency, memory, sampling configuration, and model capability compared with cloud APIs. ### Source excerpt Put your phone in airplane mode. Open the app, type a question, and watch the answer arrive one token at a time -- no spinner waiting on a network round-trip, no API key, no per-token bill, and nothing you typed ever leaving the device. LiteRT-LM removes the genuinely hard parts of running an LLM on-device -- KV-cache management, token streaming, backend selection -- but it doesn't remove your job so much as relocate it. What's left on your plate is a short, specific list: sizing a combined input+output token budget, owning your own sampling defaults, hand-building system prompts and tool calling out of raw text, and one native-library collision that presents as a SIGSEGV rather than a build error. Know those going in and the API itself is a clean three-step pattern. We'll get there in that order: Why you'd choose this runtime and what it costs you versus the cloud. The Engine/Session model you need to read the code at all. Real implementation samples -- streaming, system prompts and tool calling, multimodal inputs, thinking mode, and CPU-vs-GPU benchmarking. The anti-patterns to avoid. A developer-friendliness rating on the same rubric as Parts 1-4. Why Use LiteRT-LM? You reach for LiteRT-LM instead of hand-rolling generation on top of raw LiteRT when: You need multi-turn conversation, not single-shot inference -- session state and KV-cache bookkeeping are handled for you, and resetting a conversation is a session swap, not a model reload. You need streaming output -- token-by-token delivery for a responsive chat UI, instead of a blocking call that returns everything at once. You're choosing between CPU and GPU per device -- the explicit backend parameter turns that into a runtime decision instead of a build-time guess. You want a pre-converted model without doing your own PyTorch-to-LiteRT conversion work -- the Model Zoo covers Gemma, Qwen, Llama, and more out of the box. You're willing to own sampling -- the engine won't pick sane decoding defaults for you; that's on the ## Linux 7.3-rc3 Released: "Another Fairly Large RC" DevFeed: [Linux 7.3-rc3 Released: "Another Fairly Large RC"]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-09-13T22:00:23Z Content type: news Language: en Sources: [Phoronix]() Topics: [Linux](), [Large Language Model](), [LLMs]() Tags: [desktop-linux](), [drivers](), [filesystem](), [gpu](), [kernel](), [linux](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [llms](), [networking](), [open-source-graphics](), [phoronix](), [phoronix-test-suite](), [release](), [ubuntu-benchmarks](), [ubuntu-hardware]() ### AI overview Linux 7.3-rc3 is described as a relatively large release candidate, with substantial filesystem, driver, networking, and core-kernel fixes. The update includes an EROFS change disabling LZ4 rolling decompression because of potential data corruption, along with fixes affecting XFS, the SMB client, netfs, AFS, btrfs, sound drivers, GPU drivers, and several architectures. ### Source excerpt Out today is the third weekly release candidate of Linux 7.3 as this next kernel version works its way toward a stable release in the second half of October... ## Enterprise-Grade Precision for Long-Context Multimodal Embedding Inference on Cloud TPU DevFeed: [Enterprise-Grade Precision for Long-Context Multimodal Embedding Inference on Cloud TPU]() Original publisher: [Read original article]() Author: Anthony Su; Injae Kwak Published: 2026-09-12T11:04:33.891311Z Content type: article Language: en Sources: [Google Developers Blog]() Topics: [Embeddings](), [vllm](), [Inference](), [long-context](), [autoscaling](), [Kubernetes](), [Cloud](), [Google](), [Artificial Intelligence](), [model-deployment](), [Open Source]() Tags: [autoscaling](), [embedding](), [github](), [gpu](), [inference](), [kubernetes](), [long-context](), [multimodal](), [open-source](), [retrieval](), [tpu](), [vllm]() ### AI overview This article explains how Google Cloud integrated native TPU support into vLLM to serve long-context, multimodal embedding models at production scale. It describes elastic TPU provisioning with Google Kubernetes Engine, autoscaling across accelerator types, and TPU-specific optimizations for long sequences and chunked prefill. The resulting setup is designed to preserve numerical parity with GPU reference baselines while supporting high-throughput semantic retrieval applications. ### Source excerpt Google Cloud has natively integrated TPU support into the vLLM serving engine, allowing developers to elastically scale high-demand embedding pipelines using Google Kubernetes Engine (GKE). To handle massive 15K+ token contexts for models like Qwen3-Embedding-8B, the engineering team implemented TPU-specific optimizations such as hardware-safe tensor alignment, JAX/XLA compilation pre-warming, and a hybrid StepPool architecture for chunked prefill management. These enhancements achieve near-perfect numerical parity with reference GPU baselines, and developers can immediately leverage the open-sourced setup recipes on the AI-Hypercomputer GitHub to build their own high-throughput semantic retrieval applications. [Next page]()