Industry Article Content

Sidebar

BLOG CATEGORIES

RECENT ARTICLES

CUSTOM CMS BLOCK
Custom CMS block displayed at the left sidebar on the Catalog Page. Put your own content here: text, html, images, media... or whatever you like. There are many similar sample content placeholders across the store. All editable from Theme Editor.

TAGS
    Featured
    How Does Server RAM Differ from Ordinary Desktop RAM?
    On By Tan cheng / 0 comments
    Key Takeaway The most fundamental difference is Registered (RDIMM) and ECC: server memory carries an extra register chip on the DIMM (Register / RCD) that buffers address and command signals, allowing a single memory channel to carry more DIMMs and reach higher capacities and frequencies; the UDIMMs used in desktops do not have this chip. The two have different PCB notch positions and physically will not fit, so buying the wrong one means a return. Key Comparison Type Full Name Use Case ECC Typical Capacity Per Module UDIMM Unbuffered DIMM... Read more
    Should SMEs Choose a Tower Server or a Rack Server?
    Key Takeaway No server room, placed in the office, buying only 1 to 3 units: choose Tower. Already renting a rack or with a dedicated server room, needing centralised management and expansion: choose Rack (2U is the all-purpose first choice). Hong Kong office rents are high, so the true cost of a Tower is often lower than people imagine, while the expansion and cooling limitations of 1U servers are often underestimated. Key Comparison Comparison Point Tower Server Rack Server Noise About 35 to 50 dB, can be placed in the... Read more
    On By Tan cheng / 0 comments
    RTX 4090, RTX 6000 and Data Center GPU Compared
    Key Takeaway These are not different price points on the same product line but three different engineering approaches: the RTX 4090/5090 is the consumer flagship (no ECC, no NVLink, restricted drivers), the RTX 6000 series is the workstation professional card (ECC, large memory, enterprise drivers), and only the H100/H200/B200 are data centre training cards (HBM, extremely high bandwidth, NVLink interconnect). The cost of choosing wrong is not slowness but a training run crashing halfway through, data corruption or licence violations. Key Comparison Model Architecture Memory ECC Bandwidth NVLink Hong Kong... Read more
    On By Tan cheng / 0 comments
    How Much GPU Memory Does AI Model Training Need?
    Key Takeaway Calculate separately for three scenarios: inference depends on weights plus KV cache; fine-tuning depends on weights, gradients, optimizer states and activations; only training from scratch is a cluster-scale requirement. Quick rule of thumb: full fine-tuning is roughly "parameters (B) x 16 to 20 GB", LoRA roughly "x 1.2", and 4-bit QLoRA roughly "x 0.7". Key Comparison Model Size Full Fine-Tuning (minimum) LoRA (rank 16 to 64) 4-bit QLoRA Inference Only (FP16) 7B 112 to 160 16 to 24 6 to 10 15 to 18 13B to 14B 210... Read more
    On By Tan cheng / 0 comments
    What Should Enterprises Watch Out for When Buying Computers in Bulk?
    Key Takeaway Eighty per cent of the success of bulk procurement depends on manageability, serviceability and warranty terms, not the numbers on a single unit's spec sheet. Standardising models, specifying business series, and locking down warranty terms and the acceptance process matter far more than saving a few hundred dollars per unit. Key Comparison Requirement Tier Typical Roles Recommended Specification Reference Price (Per Unit) Basic Office Tier Administration, customer service, front desk 14-inch, mid-range U-series CPU, 16GB, 512GB NVMe HK$4,500 to 7,000 Mainstream Business Tier Managers, sales, accounting 14-inch, U/P-series... Read more
    On By Tan cheng / 0 comments