# SSD Published articles for SSD. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Pinterest's Manas Search Platform Uses Quantization and SSD-Based Serving DevFeed: [Pinterest's Manas Search Platform Uses Quantization and SSD-Based Serving]() Original publisher: [Read original article]() Author: Olimpiu Pop Published: 2026-09-16T06:06:00Z Content type: news Language: en Sources: [InfoQ]() Topics: [quantization](), [AI search](), [webgpu]() Tags: [algorithms](), [architecture-design](), [development](), [infrastructure](), [news](), [optimization](), [performance](), [pinterest-search](), [quantization](), [search](), [ssd](), [vector]() ### AI overview Pinterest Engineering enhanced its Manas distributed search platform with scalar and product quantization, SSD-based serving, and late-interaction retrieval. The reported evaluations describe trade-offs among index size, recall, throughput, latency, and serving cost. ### Source excerpt Pinterest Engineering has enhanced its Manas search platform to manage vast data, improving efficiency in search and discovery functions. By applying Scalar and Product Quantization, memory usage decreased significantly while maintaining high recall rates. The platform utilizes SSDs for optimized performance, and it is transitioning to multi-vector models for refined relevance matching. By Olimpiu Pop ## Linux server performance: Is disk I/O slowing your application? DevFeed: [Linux server performance: Is disk I/O slowing your application?]() Original publisher: [Read original article]() Author: Hayden James Published: 2026-09-12T16:11:54Z Content type: tutorial Language: en Sources: [Hayden James]() Topics: [Linux](), [IO](), [Server](), [Command-line interface](), [Processes](), [NVMe]() Tags: [blog](), [command-line](), [cpu](), [linux](), [load](), [mysql](), [nvme](), [performance](), [processes](), [request](), [server](), [ssd](), [storage](), [sysadmins](), [terminal]() ### AI overview This tutorial explains how disk I/O bottlenecks can cause high Linux load averages even when CPU usage is low. It shows how to use top, atop, and iotop on a LEMP web server, including per-core I/O wait measurements, to identify storage-related application slowdowns. ### Source excerpt If your Linux server is bogged down by disk I/O, your first step may often be to use the top command in the terminal to check load averages. Continue reading... ## CachyOS vs. Windows 11 vs. Ubuntu 26.04 LTS On Intel Wildcat Lake + 8GB RAM DevFeed: [CachyOS vs. Windows 11 vs. Ubuntu 26.04 LTS On Intel Wildcat Lake + 8GB RAM]() Original publisher: [Read original article]() Author: Michael Larabel Published: 2026-09-11T15:08:55Z Content type: comparison Language: en Sources: [Phoronix]() Topics: [benchmarking](), [Linux](), [Ubuntu](), [Windows](), [cpu](), [Hardware](), [intel](), [Linux performance]() Tags: [benchmarking](), [desktop-linux](), [hardware](), [intel](), [intel-core](), [laptop](), [linux](), [linux-benchmarking](), [linux-hardware-benchmarks](), [linux-hardware-reviews](), [linux-how-to](), [linux-performance](), [linux-server-benchmarks](), [lts](), [memory](), [microsoft](), [nvme](), [open-source-graphics](), [performance](), [phoronix](), [phoronix-test-suite](), [release](), [review](), [ssd](), [testing](), [ubuntu](), [ubuntu-benchmarks](), [ubuntu-hardware](), [windows](), [windows-11]() ### AI overview This comparison measures Ubuntu 26.04 LTS, CachyOS, and Windows 11 performance on a $449 CHUWI UniBook laptop with an Intel Core 3 Wildcat Lake processor, 8GB of memory, and a 256GB NVMe SSD across various workloads. ### Source excerpt Often times when testing different Linux distributions or comparing Windows vs. Linux it's on leading flagship desktop or server hardware, but today we are looking at the Ubuntu vs. CachyOS vs. Windows performance at the opposite end of the spectrum. With the new CHUWI UniBook $449 laptop powered by Intel Core 3 Wildcat Lake and with 8GB of system memory, here is a look at how those three operating systems compare across a variety of workloads. ## Micron 6600 ION 245TB: Swap the Hard Drives, Power an NVL72 for Free DevFeed: [Micron 6600 ION 245TB: Swap the Hard Drives, Power an NVL72 for Free]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-09-03T17:00:39Z Content type: article Language: en Sources: [StorageReview.com]() Topics: [Benchmark](), [benchmarking](), [datacenter](), [GB200](), [GPU](), [Server](), [dell]() Tags: [benchmark](), [compute](), [data-center](), [dell](), [efficiency](), [enterprise](), [gb200](), [gpu](), [hdd](), [measurements](), [nvl72](), [performance](), [servers](), [ssd](), [storage]() ### AI overview This article evaluates replacing eight 30TB nearline HDDs with one 245TB Micron 6600 ION SSD. Its measurements indicate lower power consumption, substantially higher read efficiency, and a reduction from 22 racks of HDD storage to six racks of flash at exabyte scale, potentially freeing enough power for a GB200 NVL72. ### Source excerpt For two decades, the SSD-versus-HDD conversation ended the same way: flash wins on performance, disk wins on price per terabyte, and the size of that price gap settled the argument in favor of bulk storage. As storage technology has matured and AI has taken over, that framing is clearly out of date. The largest data The post Micron 6600 ION 245TB: Swap the Hard Drives, Power an NVL72 for Free appeared first on StorageReview.com. ## KIOXIA CM9-R 15.36TB Review: BiCS8 Flash Hits Full Speed at Low Queue Depths DevFeed: [KIOXIA CM9-R 15.36TB Review: BiCS8 Flash Hits Full Speed at Low Queue Depths]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-09-02T16:13:32Z Content type: article Language: en Sources: [StorageReview.com]() Topics: [datacenter](), [Security](), [Availability](), [configuration](), [.NET Framework](), [dell](), [intel](), [Nvidia]() Tags: [dell](), [enterprise](), [fips-140-3](), [high-availability](), [intel](), [lts](), [nvidia](), [pcie](), [performance](), [review](), [security](), [ssd](), [ubuntu]() ### AI overview This review evaluates the 15.36TB E3.S KIOXIA CM9-R enterprise SSD, built on eighth-generation BiCS FLASH. It highlights strong low-queue-depth performance, including 3,400K random read IOPS, 14,800 MB/s sequential read, 11,000 MB/s sequential write, and 540K random write IOPS, while emphasizing a typical active power rating of 25W. The drive supports PCIe 5.0 x4, dual-port operation, NVMe 2.0, NVMe-MI 1.2c, power-loss protection, end-to-end data protection, and optional security variants. ### Source excerpt The KIOXIA CM9-R is the first enterprise SSD family built on eighth-generation BiCS FLASH, and the 15.36TB E3.S model in our lab is the volume sweet spot of the range: 3,400K random read IOPS and 14,800 MB/s of sequential read from a drive with a 25W typical active rating. Gen5 enterprise drives routinely push into The post KIOXIA CM9-R 15.36TB Review: BiCS8 Flash Hits Full Speed at Low Queue Depths appeared first on StorageReview.com. ## Debugging my new network, when 10 Gigabit Ethernet Runs at 300 Megabits DevFeed: [Debugging my new network, when 10 Gigabit Ethernet Runs at 300 Megabits]() Original publisher: [Read original article]() Published: 2026-08-27T15:25:48Z Content type: article Language: en Sources: [Scott Hanselman]() Topics: [Homelab](), [debugging](), [UniFi UNAS Pro](), [intel](), [Windows](), [NVMe]() Tags: [debugging](), [homelab](), [intel](), [musings](), [nvme](), [smb](), [ssd](), [storage](), [unifi-unas-pro](), [windows]() ### AI overview A Windows desktop connected to a 10GbE homelab network was limited to roughly 100-200 megabits per second despite 10-gigabit link indications. Direct testing traced the problem to the Windows/NIC path: increasing the Intel E610 driver's receive buffers eliminated discarded packets and improved receive throughput, while disabling IPv4 Large Send Offload V2 raised transmit throughput to 7.03 Gbit/sec. ### Source excerpt I've been moving my home storage over to a UniFi UNAS Pro 8 as part of a larger homelab moderization. My main IRONHEART (the ultimate PC from a few years back) desktop now has an Intel E610-XT2 10GbE card, the NAS is on 10GbE, and there's a Minisforum MS-01 miniPC on the same network with a 10GbE SFP+ connection running Immich and Portainer and a few other things. Everything says 10 gigabit. Windows says 10 gigabit. UniFi says 10 gigabit. SMB copies are using the correct NIC but my file copies are running at around 100 to 200 megabits per second which is sad making. Naturally, I blamed the NAS, and the spinning rust within. The UNAS has six 16 TB spinning disks in RAID 6 and a pair of NVMe SSDs being used as cache. I'm also running Immich on the MS-01, with its photo library living on the UNAS, so there are lots of thumbnails, metadata reads, and little background writes happening. All seem like reasonable suspects. I switched the UNAS SSD cache from read-write to read-only. No meaningful difference. I stopped Immich completely. No difference. I looked at iostat; the disks weren't saturated. We looked at SMB signing and Windows Defender network scanning. Still slow. Then I stopped testing the NAS and ran iperf3 directly between the Windows desktop and the MS-01: iperf3 -c 192.168.1.222 -P 4 133 Mbit/sec Oops. The reverse test was better, but still wrong: iperf3 -c 192.168.1.222 -P 4 -R 1.33 Gbit/sec That's weird. Now the disks, SMB, Immich, RAID, and the NAS itself were completely out of the equation. This was a Windows/NIC problem and it's weirdly asymmetrical. Looking at the Intel adapter statistics shows me... Get-NetAdapterStatistics -Name "Ethernet - 10 Gig Intel" There were nearly a million ReceivedDiscardedPackets. During one ten-second iperf3 test, the counter increased by another 268. Why? The E610 driver had its receive buffers at the default 512, although it supported up to 4096. I increased them. I love an increased buffer. Set-NetAdapterAdvancedProperty ## Hot Chips 2026: Applying High Bandwidth Flash (HBF) DevFeed: [Hot Chips 2026: Applying High Bandwidth Flash (HBF)]() Original publisher: [Read original article]() Author: Chester Lam Published: 2026-08-23T22:51:05Z Content type: article Language: en Sources: [Chips and Cheese]() Topics: [Machine Learning & Artificial Intelligence](), [Inference](), [vllm](), [moe](), [Cache]() Tags: [2026](), [dram](), [machine-learning](), [memory](), [moe](), [ssd](), [vllm]() ### AI overview This article examines how High Bandwidth Flash (HBF) could support machine learning workloads. HBF does not yet have products; the discussion uses simulations and projections to explore software strategies, including moving Mixture-of-Experts components or KV cache data between HBF and faster memory, with vLLM as an example. ### Source excerpt Machine learning workloads have an insatiable appetite for DRAM capacity. Flash memory is cheaper per gigabyte of capacity than DRAM. Could it offer a way out? ## Sandisk 600 and 800 NAS SSDs Revealed DevFeed: [Sandisk 600 and 800 NAS SSDs Revealed]() Original publisher: [Read original article]() Author: Rob Andrews Published: 2026-08-20T12:00:27Z Content type: news Language: en Sources: [NAS Compares]() Topics: [Hardware](), [NVMe](), [M.2 NVMe](), [Specifications]() Tags: [asustor-ssd](), [m-2-nvme](), [nas-600](), [nas-800](), [nas-drive-ssd](), [nas-hardware](), [nas-ssd](), [network-attached-storage-nas](), [nvme](), [qnap-ssd](), [sandisk](), [sandisk-1000](), [sandisk-200](), [sandisk-2026](), [sandisk-2027](), [sandisk-400](), [sandisk-600](), [sandisk-800](), [sandisk-nas-600](), [sandisk-nas-800](), [sandisk-nas-ssd](), [sandisk-ssd](), [specifications](), [ssd](), [ssd-for-nas](), [ssd-nas-drive](), [synology-ssd](), [ugreen-ssd](), [uncategorised](), [wd-2026](), [wd-2027](), [wd-nas-ssd](), [wd-red-nas-ssd](), [wd-red-ssd]() ### AI overview Sandisk has introduced the NAS 600 and NAS 800 SSD families for NAS systems. The NAS 600 is a 2.5-inch SATA SSD for storage and all-flash arrays, while the NAS 800 is an M.2 2280 PCIe 5.0 NVMe SSD for caching, tiered storage and primary storage in compatible systems. ### Source excerpt New SSDs for NAS Released by Sandisk - Too late? Sandisk has introduced 2 SSD families designed specifically for NAS use, covering both conventional 2.5-inch SATA storage and higher-performance M.2 NVMe applications. The SANDISK NAS 600 and NAS 800 series are intended for prosumers, creative professionals and small businesses using always-on systems for shared storage, [...] ## How fsync affects data durability, latency, and group commit DevFeed: [How fsync affects data durability, latency, and group commit]() Original publisher: [Read original article]() Author: hello@ankit-rana.com Published: 2026-08-14T00:00:00Z Content type: article Language: en Sources: [Ankit Rana | Mechanical Sympathy]() Topics: [data](), [Latency](), [NVMe](), [Replication]() Tags: [databases](), [durability](), [flush](), [fsync](), [postgresql](), [replication](), [ssd](), [stack](), [storage]() ### AI overview This article explains that a successful write() may leave data only in the kernel page cache, so a power loss can destroy it. It describes fsync as the operation that pushes data to the device, discusses its latency cost, and explains how group commit amortizes the synchronization barrier. It also distinguishes disk durability from replication across failure domains. ### Source excerpt A successful write() only copies your data into the kernel page cache, where a power loss destroys it. Only fsync pushes it to the device, and because fsync is a barrier rather than a write, a system that syncs once per commit is capped at roughly one divided by the fsync latency, no matter how fast the rest of the stack is. Group commit exists to amortise that barrier, and replication solves a different failure domain than fsync does, which is why acks=all does not mean the data is on any disk. ## Monitor Your Drive Health with Performance Co-Pilot on Fedora DevFeed: [Monitor Your Drive Health with Performance Co-Pilot on Fedora]() Original publisher: [Read original article]() Author: Paul Evans Published: 2026-08-10T08:00:00Z Content type: article Language: en Sources: [Fedora Magazine]() Topics: [Fedora](), [Monitoring](), [Grafana](), [telemetry]() Tags: [data](), [diagnostics](), [drive](), [farm](), [fedora-project-community](), [grafana](), [hdd](), [health](), [home-server](), [metrics](), [monitor](), [monitoring](), [nvme](), [open-source](), [performance](), [performance-metrics](), [self-hosted](), [server](), [smart](), [ssd](), [telemetry](), [wwid]() ### AI overview This article explains how to monitor drive health on Fedora using Performance Co-Pilot (PCP). It covers SMART metric collection for HDDs, SSDs, and NVMe drives, NVMe error-log decoding, WWID-based tracking, Seagate FARM telemetry, continuous metric collection, and Grafana dashboards for viewing drive-health trends. ### Source excerpt You wake up one morning to find your home server unresponsive, after some investigation you discover a failed NVMe drive taking your self-hosted services and data with it. Perhaps you're a system administrator and a workstation's SSD has been silently accumulating errors for months, and now a user is reporting corrupted files. Drive failures are [...] ## How to Write to SSDs - Co-Designing DBMS and Flash Storage DevFeed: [How to Write to SSDs - Co-Designing DBMS and Flash Storage]() Original publisher: [Read original article]() Published: 2026-06-22T00:00:00Z Content type: article Language: en Sources: [Gaurav Sarma's Blog]() Topics: [Database](), [Databases](), [systems](), [optimize](), [Parallelism]() Tags: [b-tree](), [checkpoint](), [cycles](), [database](), [databases](), [latency](), [mysql](), [optimize](), [optimizing](), [postgresql](), [ssd](), [storage](), [systems](), [wal](), [write-amplification]() ### AI overview The article explains how database management systems and SSDs jointly amplify writes. It reports that a 4 KiB logical page write can become about 18.85 KiB of flash writes on a Samsung PM9A3, and presents DBMS-SSD co-design, including avoiding in-place updates, as a way to address the combined amplification. ### Source excerpt . [How to Write to SSDs](optimising-ssd-writes-for-dbms-cover... ## Security Week 2623: слежка за пользователями через SSD DevFeed: [Security Week 2623: слежка за пользователями через SSD]() Original publisher: [Read original article]() Author: Kaspersky\_Lab ("Лаборатория Касперского") Published: 2026-06-01T17:31:16Z Content type: news Language: ru Sources: ["Лаборатория Касперского" RU]() Topics: [Security](), [Filesystems](), [Google Chrome](), [JavaScript](), [Linux](), [macOS]() Tags: [google-chrome](), [javascript](), [linux](), [mac](), [macos](), [security](), [side-channel](), [ssd](), [tag-9fe8963de219]() ### AI overview Researchers from Graz University of Technology described FROST, a browser-based contention side-channel attack that uses repeated SSD access through the Origin Private File System to infer running programs and visited websites from storage-performance delays. The attack runs in JavaScript without user intervention and was tested with Google Chrome on Linux and macOS systems. ### Source excerpt На прошлой неделе исследователи из Грацского технического университета в Австрии опубликовали научную работу, в которой предложили метод отслеживания пользовательской активности через браузер. Варианты атак, в которых вкладка с вредоносным сайтом может, например, определять, какие страницы открыты по соседству, предлагались и ранее, но полагались в основном на особенности работы браузеров. Новая атака FROST также использует особенности браузеров, но основным каналом утечки оказываются не они, а производительность накопителя данных. Общая схема атаки выглядит следующим образом: пользователя заманивают на вредоносную страницу, которая при помощи стандартных технологий браузеров реализует активный и регулярный обмен данными на SSD. Производительность SSD ниже, чем у оперативной памяти, что вызывает определенные задержки при обмене, когда к накопителю также обращаются другие программы и веб-сайты. Как выяснилось, паттерны этих задержек позволяют с достаточно высокой надежностью выявлять не только запущенные на ПК программы, но и посещаемые веб-сайты. Читать далее ## Security Week 2621: три уязвимости нулевого дня в Windows DevFeed: [Security Week 2621: три уязвимости нулевого дня в Windows]() Original publisher: [Read original article]() Author: Kaspersky\_Lab ("Лаборатория Касперского") Published: 2026-05-18T15:51:44Z Content type: news Language: ru Sources: ["Лаборатория Касперского" RU]() Topics: [Security](), [Windows](), [Windows 11](), [Microsoft]() Tags: [bitlocker](), [microsoft](), [plasma](), [security](), [ssd](), [tag-9fe8963de219](), [windows](), [windows-11](), [yellow]() ### AI overview The article reports three Windows vulnerabilities publicly disclosed with proof-of-concept code before a manufacturer patch was available. YellowKey reportedly allows BitLocker encryption to be bypassed with physical access on Windows 11 systems using default TPM-based settings; GreenPlasma may enable privilege escalation but its exploit is incomplete; and MiniPlasma affects cldflt.sys and may enable privilege escalation even on fully patched systems. ### Source excerpt На прошлой неделе стало известно о трех новых уязвимостях нулевого дня в Windows. Обычно термин zero-day применяется в случае, если уязвимость на момент обнаружения используется в реальных атаках, но в данном случае имеет место другая ситуация. Данные о проблемах в открытом доступе опубликовал аноним, называющий себя Nightmare Eclipse. Уязвимости снабжены демонстрационным кодом, который вполне возможно применить для реальных атак, -- и это при отсутствии патча от производителя. Отсюда и квалификация zero-day, намеренно созданная человеком, который, судя по его риторике, затаил некую обиду на Microsoft. Наибольший интерес представляет проблема YellowKey -- это ошибка в логике работы системы шифрования BitLocker (подробное описание в издании Ars Technica, новость на Хабре). При наличии физического доступа к компьютеру или ноутбуку YellowKey позволяет полностью обойти шифрование и получить прямой доступ к данным. Эксплойт работает так: нужно записать набор файлов на флешку, подключить ее к компьютеру, загрузить его в режиме Recovery. Дальнейшая последовательность действий открывает консоль, в которой будет предоставлен полный доступ к зашифрованному хранилищу данных. Читать далее ## 【分布式系统百科】分布式 KV 存储对比：etcd、TiKV 与 FoundationDB DevFeed: [【分布式系统百科】分布式 KV 存储对比：etcd、TiKV 与 FoundationDB]() Original publisher: [Read original article]() Author: Liao Tonglang Published: 2026-04-13T00:00:00Z Content type: comparison Language: zh Sources: [土法炼钢 - 系统与基础设施]() Topics: [etcd](), [Raft](), [Kubernetes]() Tags: [bbolt](), [boltdb](), [compaction](), [coreos](), [distributed](), [distributed-systems](), [errcompacted](), [etcd](), [go](), [k8s](), [kubernetes](), [lease](), [mmap](), [mvcc](), [raft](), [readindex](), [revision](), [single-raft-group](), [ssd](), [v3](), [wal](), [watch]() ### AI overview This Chinese-language comparison examines etcd, TiKV, and FoundationDB as distributed key-value stores for metadata. It focuses on architecture, implementation, performance, and limitations, with detailed coverage of etcd's single Raft group, MVCC, Watch, Lease, bbolt storage engine, linearizable reads, compaction, and defragmentation. ### Source excerpt 一个典型的架构评审场景：团队需要一个分布式键值存储来承载新系统的元数据。候选方案三个----etcd、TiKV、FoundationDB。三者都提供强一致性保证，都经过大规模生产验证，但设计目标截然不同。选错了，轻则性能不达标，重则数据丢失后无法恢复。这篇文章从架构、实现、性能和局限四个维度，把三个系统拆开来看。 ## Introducing no-cost trials for Data Connect DevFeed: [Introducing no-cost trials for Data Connect]() Original publisher: [Read original article]() Author: Cynthia Wang Published: 2026-03-12T00:00:00Z Content type: release Language: en Sources: [Firebase Blog]() Topics: [Firebase](), [PostgreSQL](), [Database](), [Cloud]() Tags: [billing](), [firebase](), [free](), [news](), [postgresql](), [quotas](), [sql-connect](), [ssd](), [storage](), [updates]() ### AI overview Firebase announces no-cost trials for Firebase Data Connect. Spark-plan projects can provision one PostgreSQL database instance for 90 days without a credit card, while Blaze-plan projects qualify for a no-cost Cloud SQL for PostgreSQL trial for three months. ### Source excerpt News, tutorials, and updates from the Firebase team. ## How Tinybird's storage architecture works: S3, local caching, and zero-copy replication DevFeed: [How Tinybird's storage architecture works: S3, local caching, and zero-copy replication]() Original publisher: [Read original article]() Author: Daniel Pozo, Irene Martínez Published: 2026-03-02T10:00:00Z Content type: article Language: en Sources: [Tinybird]() Topics: [clickhouse](), [Amazon S3](), [Caching](), [Replication](), [cloud-infrastructure]() Tags: [architecture](), [caching](), [clickhouse](), [engineering-excellence](), [object-storage](), [replication](), [run](), [s3](), [ssd](), [storage](), [zero-copy]() ### AI overview This article explains how Tinybird runs ClickHouse on object storage using local SSD caching, zero-copy replication, and a packed part format. It states that the format reduces S3 costs by 30-40%. ### Source excerpt How we run ClickHouse on top of object storage with local SSD caching, zero-copy replication, and a packed part format that cuts S3 costs by 30-40%. ## My Private Mini Kubernetes Cluster - Secured by Tailscale DevFeed: [My Private Mini Kubernetes Cluster - Secured by Tailscale]() Original publisher: [Read original article]() Author: Techno Tim Published: 2025-11-07T13:00:00Z Content type: tutorial Language: en Sources: [Techno Tim]() Topics: [Kubernetes](), [Virtual Private Network](), [5G](), [rancher](), [GitOps](), [Ubuntu](), [networking]() Tags: [5g](), [dark-mode](), [firewall](), [gitops](), [homelab](), [intel](), [kubernetes](), [latency](), [rancher](), [ssd](), [tailscale](), [ubuntu](), [unifi-cloud-gateway](), [vpn]() ### AI overview A tutorial documenting a private, three-node RKE2 Kubernetes cluster built on Beelink Mini PCs and connected securely through Tailscale over a 5G WAN. It covers UniFi networking, Ubuntu installation, PiKVM access, cluster bootstrapping, and Flux-based GitOps management. ### Source excerpt A compact Kubernetes cluster, completely private and connected over 5G - powered by Tailscale. 📺 Watch the video Why Build a Private Kubernetes Cluster? I wanted a way to run a full Kubernetes cluster that I could access securely from anywhere - no open ports, no complicated VPN setup, no public exposure. Tailscale made that possible by creating a secure network for exposing and connect... ## I Rebuilt My Home Server -- Quieter, Faster, and Ready for 2025 DevFeed: [I Rebuilt My Home Server -- Quieter, Faster, and Ready for 2025]() Original publisher: [Read original article]() Author: Techno Tim Published: 2025-04-12T13:00:00Z Content type: opinion Language: en Sources: [Techno Tim]() Topics: [Homelab](), [Hardware](), [servers](), [cpu](), [GPU](), [intel](), [Nvidia](), [NVIDIA RTX]() Tags: [cpu](), [files](), [gpu](), [hardware](), [home-server](), [homelab](), [intel](), [lighting](), [noctua](), [nvidia](), [nvidia-rtx](), [pwm](), [server](), [ssd](), [upgrades]() ### AI overview Techno Tim describes rebuilding a home server with upgraded CPU, GPU, RAM, fans, SSDs, and power supply. The article says the rebuilt system is quieter and faster, and lists hardware including Noctua cooling components, Intel products, NVIDIA GPUs, ECC RAM, and SSDs. ### Source excerpt I rebuilt my home server from the ground up: CPU, GPU, RAM, fans, SSDs, and power supply... even some "accessory lighting". This thing is now quieter, faster, and built to handle just about anything I throw at it in 2025 and hopefully beyond. Thank you to Noctua for helping to make this server quiet! Thank you to Patrick from Serve the Home for the CPU upgrades! (Both Noctua and Patrick gifted... ## LLM Inference Machine for $300 DevFeed: [LLM Inference Machine for $300]() Original publisher: [Read original article]() Published: 2024-12-28T04:17:14Z Content type: article Language: en Sources: [Ariya Hidayat]() Topics: [Inference](), [Large Language Model](), [GPU](), [datacenter](), [data]() Tags: [3d-printed](), [amd](), [cost](), [ddr4](), [gpu](), [inference](), [llama](), [llm](), [memory](), [models](), [nvidia](), [performance](), [qwen](), [ssd](), [threads](), [vision]() ### AI overview A $300 used-hardware build runs quantized LLMs including Qwen-2.5 32B, Llama-3.1 8B, and Llama-3.2 Vision 11B. It uses an NVIDIA Tesla M40 with 24GB of VRAM and achieves model-dependent speeds from 7 to 47 tokens per second. The article compares its performance, cost, and upgrade options with newer GPUs and Apple Silicon. ### Source excerpt You can absolutely run Qwen-2.5 32B. And of course, Llama-3.1 8B and Llama-3.2 Vision 11B are no problem at all. ## We're Cutting Storage Costs (Automatically) DevFeed: [We're Cutting Storage Costs (Automatically)]() Original publisher: [Read original article]() Author: Anna Stepanyan Published: 2024-11-21T00:26:43Z Content type: article Language: en Sources: [Blog -- Neon Docs]() Topics: [cloud-infrastructure](), [Database](), [Cache](), [Amazon S3](), [Availability](), [Developer experience](), [Azure]() Tags: [amazon-s3](), [availability](), [azure](), [cache](), [cloud-storage](), [company](), [data](), [developer-experience](), [performance](), [ssd](), [storage]() ### AI overview Neon is reducing storage costs for inactive database branches automatically. Branches that have been idle for at least one day and are older than two weeks are archived by evicting their data from the Pageserver's high-performance SSD cache while retaining it in durable cloud object storage such as Amazon S3 or Azure Blob Storage. When accessed again, a branch is automatically reactivated, with a small first-query delay while its data is reloaded into the cache. ### Source excerpt Starting December 1st, many Neon users will see lower storage costs. Inactive branches--those idle for at least a day and older than 2 weeks--will now be charged at a reduced per-GB rate. If this applies to you, there's nothing you need to do--it all happens automatically. Simply en... ## Building an All-in-One Homelab Server DevFeed: [Building an All-in-One Homelab Server]() Original publisher: [Read original article]() Author: Techno Tim Published: 2024-05-10T13:00:00Z Content type: opinion Language: en Sources: [Techno Tim]() Topics: [Homelab](), [Server](), [Hardware]() Tags: [ethernet](), [hardware](), [home-server](), [homelab](), [ssd](), [video]() ### AI overview A personal build report about assembling an all-in-one homelab server. The article lists components including a Sliger CX4150a case, Samsung SSDs, a 10G Ethernet adapter, and an LSI HBA, and discloses that Sliger provided the case without requesting anything in return. ### Source excerpt Today I built the ultimate, all in one, HomeLab Home Server to handle everything. 📺 Watch Video Disclosures: Sliger did send this case to me however asked for nothing in return. Other 4u Cases Sliger CX4150a - https://www.sliger.com/products/rackmount/4u/cx4150a/ SilverStone RM44 4U - https://amzn.to/3K0wpmk RackChoice 4U - https://amzn.to/3UB8bEf Other Parts Samsung SSD... ## UniFi Dream Machine Pro Max: Features, Shadow Mode, and Network Throughput Testing DevFeed: [UniFi Dream Machine Pro Max: Features, Shadow Mode, and Network Throughput Testing]() Original publisher: [Read original article]() Author: Techno Tim Published: 2024-04-23T13:00:00Z Content type: comparison Language: en Sources: [Techno Tim]() Topics: [Network](), [Hardware]() Tags: [cpu](), [drive](), [hardware](), [homelab](), [network](), [ssd](), [testing](), [unifi](), [upgrades](), [video]() ### AI overview A video review of the UniFi Dream Machine Pro Max examines its hardware upgrades, configures and tests Shadow Mode, and measures network throughput while comparing it with predecessor devices. ### Source excerpt The UDM Pro Max is here and it's packed with upgrades like a faster CPU, more RAM, an internal SSD, more eMMC, Dual Drive bays and more! Today we check on the new UniFi Dream Machine Pro Max, configure and test Shadow Mode, and test network throughput to see if this really is the fastest UniFi Dram Machine yet. 📺 Watch Video Get your UDM Pro Max here: https://l.technotim.com/udm-pro-max... ## HomeLab Hardware Tour! (Late 2023) DevFeed: [HomeLab Hardware Tour! (Late 2023)]() Original publisher: [Read original article]() Author: Techno Tim Published: 2023-12-08T13:00:00Z Content type: article Language: en Sources: [Techno Tim]() Topics: [Homelab](), [Hardware](), [Server](), [networking]() Tags: [hardware](), [homelab](), [jellyfin](), [networking](), [nvme](), [power-management](), [server](), [ssd](), [unifi](), [ups]() ### AI overview A late-2023 tour of a home lab and server rack, covering servers, networking equipment, UPS units, power management, storage, lighting, TV gear, and an Intel NUC mini cluster. ### Source excerpt Well, here it is! My Late 2023 Server Rack and HomeLab tour! I've upgraded, replaced, added, and consolidated quite a bit since my last tour! New servers, new networking, UPS, cabling, power management, and more new tech on the wall! 📺 Watch Video In case you missed it, check out my HomeLab Services Tour (2024)! Join the conversation It's time of year again! Time for my Server Rack... ## Geth v1.13.0 DevFeed: [Geth v1.13.0]() Original publisher: [Read original article]() Author: Péter Szilágyi Published: 2023-09-12T00:00:00Z Content type: release Language: en Sources: [Ethereum Foundation Blog]() Topics: [Ethereum](), [Database]() Tags: [database](), [ethereum](), [offline](), [release](), [research-development](), [ssd](), [technical]() ### AI overview Geth v1.13.0 introduces a new database model for storing Ethereum state. The model is faster than the previous scheme and supports proper pruning, addressing indefinite accumulation of unused data. State trie nodes are stored by path rather than hash to enable pruning. ### Source excerpt Geth v1.13 comes fairly close on the heels of the 1.12 release family, which is funky, considering it's main feature has been in development for a cool 6 years now. 🤯 This post will go into a number of technical and historical details, but if you just want the... [Next page]()