How-To
157 articles
-
Your S$2,000 Laptop Can't Run Two Monitors? Here's the Real Reason
Why a laptop won't drive two monitors: display pipes, macOS and MST docks, base Apple chips, USB-C Alt Mode and DisplayLink trade-offs, with fixes.
Wei Fei
·
-
Why a UPS Rated at 1500VA Can't Always Support a 1500W PC
VA vs watts explained: why 1500VA UPS models range from 865 W to 1,500 W, why PCs rarely draw their PSU rating, and how to size a UPS...
Wei Fei
·
-
Video Conferencing & Meeting Rooms
Why Touchscreen Displays Are Often Used Like Ordinary TVs
Why interactive displays end up as plain screens, and how home-screen setup, short training and usage checks get the touch features used.
Wei Fei
·
-
Why IT Departments Are Buying Business Laptops Without Testing the Docking Stations
MAC pass-through, PXE, Wake-on-LAN, firmware and monitor limits vary by laptop and dock. A pilot test plan to run before a fleet order.
Wei Fei
·
-
Why Your Headset Works on Zoom but Fails on Microsoft Teams
Why a headset works in Zoom but fails in Microsoft Teams: device selection, MS vs UC variants, firmware, Bluetooth pairing and call-control fixes.
Wei Fei
·
-
ASUS ProArt GR1X: The 128GB AI Memory Trap Nobody Talks About
Why 128GB on the ASUS ProArt GR1X doesn't mean fast AI: ~300 GB/s bandwidth, decode maths, GPU memory split and which models suit RTX Spark.
Wei Fei
·
-
Your GPU Has 20 Petaflops of Power. So Why Is Your AI Model Still Slow?
Why a 20-petaflop GB300 can still run slow: memory bandwidth, CPU-GPU links, storage, software and network bottlenecks, with figures to diagnose each.
Wei Fei
·
-
ASUS ProArt GR1X vs MSI EdgeMesa: The Buying Mistake Nobody Warns AI Developers About
Before choosing ASUS ProArt GR1X or MSI EdgeMesa, check Windows on Arm AI software, GPU memory share, Linux needs and multi-user serving.
Wei Fei
·
-
ASUS ProArt GR1X: Why 1 Petaflop Doesn't Tell You How Fast Your AI Will Run
ASUS ProArt GR1X 1 petaflop FP4 explained: sparsity, memory bandwidth, tokens per second, prefill and how to judge real local AI speed.
Wei Fei
·
-
Why ASUS ProArt GR1X Might Be Better for AI Inference Than AI Training
Why the ASUS ProArt GR1X suits AI inference better than training: memory needs for 8B, 32B and 70B models, LoRA, QLoRA and full fine-tuning.
Wei Fei
·
-
ASUS ProArt GR1X for Local LLMs: Why Running a 70B Model Is Only Half the Story
Running a 70B model on the ASUS ProArt GR1X: what fits in 64GB or 128GB, expected tokens per second, context limits, quantisation and faster MoE options.
Wei Fei
·
-
ASUS ProArt GR1X: Powerful Enough for AI, But Is It Powerful Enough for Multiple Users?
Can one ASUS ProArt GR1X serve a team? Concurrency, per-user context memory, Windows serving limits and when to move to a GB10 or GPU workstation.
Wei Fei
·
-
The ASUS ProArt GR1X Windows on ARM Gamble: What Developers Need to Know
ASUS ProArt GR1X on Windows on Arm: CUDA preview, PyTorch, vLLM, WSL2, Docker and driver issues developers should check before buying.
Wei Fei
·
-
ASUS ProArt GR1X vs MSI EdgeMesa: Which One Will Throttle First?
No public sustained test of the GR1X or EdgeMesa N AI+ exists yet. What drives RTX Spark throttling, what laptops show, and how to test it.
Wei Fei
·
-
Line-Interactive vs Online UPS: Which Does Your Office Server Rack Need?
Standby, line-interactive or online double-conversion? How to pick the right UPS topology, size it in VA and watts, and keep batteries alive in Singapore's climate.
Wei Fei
·
-
PoE Budget Planning: How to Power Access Points, Cameras and Phones Without Overloading Your Switch
PoE, PoE+ and PoE++ explained, how to add up a real power budget, and the sizing mistakes that leave cameras and access points rebooting at random.
Wei Fei
·
-
800Gbps InfiniBand vs 130TB/s NVLink: Why These Numbers Don't Tell the Whole Story
Convert 800Gb/s and 130TB/s to GB/s per GPU per direction: bits vs bytes, aggregate vs per GPU, bidirectional, bisection, NCCL busbw and latency.
Wei Fei
·
-
RTX PRO 6000 vs GB300: Is More Memory Always Better for AI?
Why memory bandwidth and compute matter as much as VRAM: token-speed maths for RTX PRO 6000, RTX 5090, DGX Spark and GB300, with published results.
Wei Fei
·
Wei Fei
