Blog
GPU guides and engineering notes
How much memory a model really needs, which GPU to pick, what a month of compute costs — written by the engineers who run our servers.
No article matches your search.
Try fewer or different words, or browse the categories above.
Latest articles
AMD MI355X for AI: specs, ROCm and the workloads it suits
288 GB of HBM3e at 8 TB/s, FP8 and MXFP4, the real state of ROCm, and the workloads that benefit most. Read before you buy.
Best GPU for LLM inference in 2026: a sizing guide by model size
Use our 2026 sizing table, from RTX 5090 to 8× B300, to choose.
Dedicated vs shared GPU: what changes when you share
How vGPU, MIG, time-slicing and shared hosts affect memory, consistency, security and debugging, and when sharing is fine.
DeepSeek R1 hardware requirements: GPUs and memory for R1 and V3
About 671 GB of FP8 weights plus KV cache. See which 8-GPU servers fit R1 and V3, and the GPUs for distilled models.
FP8 vs FP4 vs BF16: number formats for AI, explained
Bit layouts, NVFP4 and MXFP4 block scaling, which GPUs support each format, the memory saved and the accuracy you trade.
GPU shortage 2026: what buyers face and how to secure GPUs
On-demand stock sells out, providers push1-year contracts and prices are up. Why it happens, and how teams still secure GPUs.
H100 rental cost in 2026: what you actually pay
SXM vs PCIe, monthly vs hourly billing, egress and storage, plus a break-even formula to compare any quote before you sign.
How to choose a GPU server for AI
A framework for memory, bandwidth, precision, interconnect, CPU and storage, software and billing, plus a checklist.
How to fine-tune a 70B model: GPU requirements and workflowQLoRA fits one 80 GB GPU, BF16 LoRA about two, full fine-tuning over 1 TB (8x H200 class). Memory math and setup.
How to self-host LLM inference with vLLMSize the GPU, run an OpenAI-compatible server, tune FP8, prefix caching and parallelism, and secure it for production.
NVLink vs PCIe: bandwidth, topology and when it matters
Per-GPU bandwidth, NVSwitch and NVL72, whichmulti-GPU workloads need NVLink and which run fine on PCIe, with worked numbers.
Deploy a GPU server
Get your GPU server this month.
Choose a configuration, create your account in the same form and pay the first month from your balance. Your server is delivered in under 10 minutes.
- Delivered in under 10 minutes
- Cancel any month
- Pay with crypto, up to +50% bonus