TY  - RPRT
TI  - Scaling LLM Inference via Parallelism: Throughput-Latency Trade-off
AU  - Javad Mirzaei
AU  - Jeebak Mitra
PY  - 2026
UR  - https://arxiv.org/abs/2610.19405
ID  - 2610.19405
ER  -