MirrorFrog AI Compute Cards Wiki
The most comprehensive AI accelerator chip encyclopedia. 100+ cards, side-by-side comparison.
Welcome to MirrorFrog AI Compute Cards Wiki — your go-to reference for AI accelerator specifications, benchmarks, and industry insights.
:::info English Version in Progress This English version is being actively translated. For the complete content, please visit the Chinese version or switch language using the dropdown above. :::
What You Can Find Here
| Section | Description |
|---|---|
| 100+ Card Comparison | Side-by-side specs for NVIDIA, AMD, Intel, Google TPU, Huawei Ascend, and more |
| Architecture Deep Dives | GPU, TPU, NPU, LPU, WSE, ASIC — how they work and when to choose each |
| Industry News | Latest launches, acquisitions, roadmaps, and market analysis |
| Selection Guides | Training vs inference, datacenter vs edge, budget vs performance |
Browse by Category
| Category | Representative Cards |
|---|---|
| Datacenter Training | H100 / H200 / B200 / MI300X / Ascend 910C |
| Datacenter Inference | L4 / L40S / RTX 6000 Ada / TPU v6e |
| Workstation | RTX 5090 / RTX 4090 / W7900 |
| Edge AI | Jetson Orin / Edge TPU / Ascend 310 |
| Innovative | Groq LPU / Cerebras WSE-3 / Graphcore IPU |
Quick Links
Last updated: June 2026