BonTechLabs is an independent technology site led by Gavin Bonshor—former Senior Editor at AnandTech and systems reporter at The Register—bringing rigorous, data-driven coverage of CPUs, GPUs, SSDs, memory, motherboards, handheld PCs, and the platforms that connect them.
What BonTechLabs is (and isn’t)
We publish analytical, lab-first pieces and fast, well-sourced news that help builders, enthusiasts, and creators make smarter decisions. Our lane is the nuts and bolts: cache, memory latency, storage behavior under mixed workloads, board design and firmware, driver policy shifts, and the business signals that inform where performance and pricing go next.
Who’s behind it: Gavin Bonshor
I’m Gavin Bonshor, a UK-based technology journalist and editor. Before launching BonTechLabs, I served as Senior Editor at AnandTech, where I directed CPU/platform coverage and wrote in-depth reviews. I later joined The Register as a systems reporter covering client-side processors, GPUs, and semiconductor developments. I’ve also contributed to outlets such as XDA Developers and others. BonTechLabs is my home for the kind of deep, methodical coverage readers know me for—now with the freedom to move faster and go broader.
Our editorial promises
- People-first content: We write for readers, not algos. We explain jargon and we show our work—assumptions, methodology, and limitations—so you can judge the conclusions yourself.
- Primary sourcing over hearsay: Where possible we cite original vendor docs, press materials, or driver notes; rumor-only items are clearly labeled and treated cautiously.
- Separation of church and state: Editorial decisions are independent of commercial relationships, affiliate economics, or vendor loans.
- Corrections welcome: If we miss something, we fix it fast and note the change. See the “Corrections & Contact” section below.
What we cover
Our day-to-day beats include:
- CPUs & platforms: architecture changes, cache/latency behavior, boosting, scheduling, microcode/AGESA, and practical buying advice. Start with our CPU pricing snapshot: Ryzen 7 9800X3D at all-time lows.
- GPUs & gaming tech: raster vs. RT tradeoffs, upscalers, frame gen, board design, and power delivery. See: RTX 5090 finally back at MSRP.
- Storage: controllers, DRAM vs HMB, SLC caching, thermal throttling, endurance, and what “real-world” actually means. Our anchor piece: The NVMe SSD Mega Guide.
- Memory & I/O: DDR5 tuning that matters (and what doesn’t), USB4 v2 / Thunderbolt 5 device realities, and display plumbing. Deep dives: DDR5 at 13,020 MT/s—what it means for you and USB4 v2 vs Thunderbolt 5 vs DP 2.1.
- Semiconductor ecosystem: foundry/packaging capacity, HBM supply, and how big industry moves ripple into consumer parts. A current example: TSMC’s N2 early adopters and NVIDIA–OpenAI supply impact.
How we work: methodology in plain English
We design tests to reveal practical behavior, not just peak numbers. Typical storage runs include sustained mixed read/write spans at realistic queue depths, cache-exhaustion and thermal soak tests, and power-limited scenarios that emulate laptops or compact builds. CPU testing emphasizes 1%/0.1% lows, memory sensitivity (timings and topology), and background load tolerance (streaming, capture, browser tabs). GPU tests favor consistent, repeatable runs with a focus on frametime stability and image-quality parity across vendor features. Every review calls out firmware/BIOS versions, driver branches, game builds, and test rig specs.
Conflict disclosures & samples
If a product is loaned or provided by a vendor, we say so in the piece. Loaners go back; samples may be retained only when clearly allowed by the vendor and when it serves long-term comparison testing. Access never buys coverage or a conclusion.
Corrections & contact
See an error? Want to add missing context? We’re reachable via the site’s Contact page. For sensitive issues (security disclosures, embargo concerns), please mention it in the subject line; we’ll prioritize accordingly. We update articles promptly with a timestamped note when factual changes are made.
Why you can trust us
- Track record: Years of front-line reporting and long-form analysis at the most demanding enthusiast outlets. The bar for accuracy and transparency set our defaults here, too.
- Repeatability: We publish enough test detail for others to reproduce our results where practical.
- Source hygiene: We prefer primary documents, driver notes, and vendor briefings—and link them. Aggregations are labeled as such and corroborated.
House style: clarity first
We write conversationally but precisely: explain the acronyms once, define the workloads, and show why a number matters. When we say “this setting is free performance,” we show the delta and the side effects. When a feature is marketing gloss, we say that, too.
Start here (reader favorites)
- Which SSD Should You Buy? The NVMe Mega Guide
- RTX 5090 at MSRP — what it means
- 9800X3D price floor and AM5 value
- DDR5 records vs daily settings
- NVIDIA–OpenAI supply impact, explained
For PR, reviewers’ guides & briefings
Please send briefing invites, driver notes, and sampling inquiries via the Contact page with embargo timing, regional availability, and any known firmware/driver requirements. We prefer at least 72 hours lead time for reviews that require multi-pass testing, thermals, and power characterization.
Affiliate links & advertising
Where we use affiliate links, we disclose them in the article. Affiliate participation does not influence our coverage, product selection, or conclusions. Sponsored posts are labeled “Sponsored” at the top—no exceptions.
Legal & reuse
All original content © BonTechLabs. Short quotations with attribution are welcome; for full or commercial reuse, please contact us.
FAQ
Do you accept guest posts? Not currently. We may commission pieces from veteran engineers, reviewers, or academics; pitches are welcome via Contact.
Will you cover mobile and AI beyond PCs? Yes—when there’s a clear tie-in to compute, memory, storage, displays, or developer workflows we already cover (e.g., NPU blocks, USB4/DP 2.1 docks, HBM supply).
How can I support BonTechLabs? Share our work, link to our deep dives, and subscribe to our newsletter when it launches. If you find a bug in our test harness or methodology, tell us—we’ll fix it fast.