2023-09-01_tpuv5e-the-new-benchmark-in-cost
Canonical Target Counts
| Target | Occurrences |
|---|
| 04_knowledge_base/TPUv5e | 39 |
| 02_companies/GOOG | 26 |
| 04_knowledge_base/TPU | 10 |
| 04_knowledge_base/TPUv5 | 10 |
| 04_knowledge_base/GPU | 9 |
| 02_companies/OpenAI | 7 |
| 04_knowledge_base/GPT-3 | 7 |
| 04_knowledge_base/H100 | 6 |
| 02_companies/NVDA | 5 |
| 04_knowledge_base/Tensor Core | 5 |
| 04_knowledge_base/TPUv4 | 5 |
| 04_knowledge_base/A100 | 5 |
| 04_knowledge_base/400G | 3 |
| 02_companies/Together AI | 2 |
| 04_knowledge_base/Microsoft Azure | 2 |
| 04_knowledge_base/NIC | 2 |
| 04_knowledge_base/XLA | 2 |
| 04_knowledge_base/CPU | 2 |
| 04_knowledge_base/BF16 | 2 |
| 04_knowledge_base/100G | 2 |
| 04_knowledge_base/Ethernet | 1 |
| 02_companies/AVGO | 1 |
| 04_knowledge_base/GCP | 1 |
| 04_knowledge_base/Quantization | 1 |
| 04_knowledge_base/HBM | 1 |
| 02_companies/Amazon | 1 |
| 04_knowledge_base/PyTorch | 1 |
| 04_knowledge_base/Pre-training | 1 |
| 02_companies/AMD | 1 |
| 02_companies/MSFT | 1 |
| 04_knowledge_base/1.6T | 1 |
| 04_knowledge_base/Trainium | 1 |
| 04_knowledge_base/Large language model | 1 |
| 02_companies/META | 1 |
| 04_knowledge_base/TOPS | 1 |
| 04_knowledge_base/SXM | 1 |
| 04_knowledge_base/Pufferfish | 1 |
| 04_knowledge_base/MTIA | 1 |
| 04_knowledge_base/Systolic Array | 1 |
| 04_knowledge_base/Twisted Torus | 1 |
| 04_knowledge_base/Matrix Multiply Unit | 1 |
| 04_knowledge_base/Contact | 1 |
| 04_knowledge_base/Inter-chip interconnect | 1 |
| 04_knowledge_base/Batch size | 1 |
| 04_knowledge_base/Speculative Decoding | 1 |
| 04_knowledge_base/HBM2E | 1 |
| 04_knowledge_base/GPT-4 | 1 |
| 04_knowledge_base/AWS Inferentia | 1 |
| 04_knowledge_base/TFLOPS | 1 |
| 04_knowledge_base/Parametric Yield | 1 |