Predicting model behavior before release by simulating deployment

The KV Cache Compression Race: TurboQuant vs OSCAR vs EpiCache

Long-context large language models (LLMs) face a memory bottleneck that has nothing to do with model weights. During decoding, transformers cache the key and value (KV) vectors for every token at every layer so they don’t have to recompute attention. This cache grows linearly with sequence length and batch size, and at long context with high concurrency it can dwarf the model’s own footprint. Consider Llama-3.1-70B in BF16. Its KV cache costs about 0.31 MB per token (80 layers ×