2024-02-21_groq-inference-tokenomics-speed-but
Canonical Target Counts
| Target | Occurrences |
|---|
| 02_companies/Groq | 58 |
| 04_knowledge_base/H100 | 23 |
| 02_companies/NVDA | 20 |
| 02_companies/GOOG | 11 |
| 04_knowledge_base/GPU | 11 |
| 04_knowledge_base/Speculative Decoding | 11 |
| 04_knowledge_base/Mixtral | 10 |
| 04_knowledge_base/HGX | 5 |
| 04_knowledge_base/Google Gemini | 5 |
| 04_knowledge_base/CPU | 5 |
| 04_knowledge_base/SRAM | 4 |
| 04_knowledge_base/KV cache | 4 |
| 04_knowledge_base/Large language model | 3 |
| 02_companies/Together AI | 3 |
| 02_companies/Marvell | 3 |
| 02_companies/OpenAI | 2 |
| 02_companies/2330 台積電 | 2 |
| 02_companies/AMD | 2 |
| 04_knowledge_base/Mixture of Experts | 2 |
| 04_knowledge_base/Groq LPU | 2 |
| 04_knowledge_base/Quantization | 2 |
| 04_knowledge_base/Llama | 2 |
| 04_knowledge_base/DRAM | 2 |
| 04_knowledge_base/B100 | 2 |
| 04_knowledge_base/Batch size | 2 |
| 04_knowledge_base/FP16 | 2 |
| 04_knowledge_base/VLIW | 2 |
| 04_knowledge_base/TFLOPS | 2 |
| 04_knowledge_base/Prefill | 2 |
| 04_knowledge_base/14nm | 2 |
| 04_knowledge_base/Hyperscaler | 1 |
| 04_knowledge_base/InfiniBand | 1 |
| 04_knowledge_base/ASIC | 1 |
| 04_knowledge_base/封裝 | 1 |
| 04_knowledge_base/Pipeline parallelism | 1 |
| 04_knowledge_base/Yield Harvesting | 1 |
| 04_knowledge_base/CoWoS | 1 |
| 02_companies/000660 | 1 |
| 04_knowledge_base/Tokenomics | 1 |
| 04_knowledge_base/HBM | 1 |
| 02_companies/SMCI | 1 |
| 04_knowledge_base/MI300X | 1 |
| 02_companies/GlobalFoundries | 1 |
| 04_knowledge_base/GPT-4 | 1 |
| 04_knowledge_base/Medusa | 1 |
| 04_knowledge_base/5nm | 1 |
| 04_knowledge_base/Process node | 1 |
| 04_knowledge_base/Parametric Yield | 1 |
| 04_knowledge_base/NVIDIA B200 | 1 |
| 04_knowledge_base/矽晶圓 | 1 |