More GPUs or a Smaller Cache? Tensor Parallelism versus KV Compression for Memory-Bound LLM Serving
When an LLM serving deployment runs out of KVcache room, there are two well-established ways out. Tensor parallelism shards the weights and the KV cache across two, four, or eight devices, buying memory headroom at the price of an all-reduce on every layer and a hardware bill that grows with the device count. The algorithms community shrinks the cache in place, with KV quantisation and eviction keeping a single GPU and spending a little quality instead. Compression papers report memory ratios, parallel-scaling papers report throughput curves, and almost nobody puts the two on the same cost axis. We place tensor-parallel configurations (degree 1 to 8) and KV-compressed configurations (16/8...
| Authors | Srikanta Datta Tumkur, Mehar Simhadri, Anshu Bansal et al. |
|---|---|
| Primary category | cs.AI |
| Published | 25 August 2026 |
| Last updated at source | 25 August 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.