We lifted 1 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| kvcache-ai/Mooncake | canonical | 1 of 1 |
| Function | Status | Where it lives |
|---|---|---|
| dequantize_fp8 | Ran | kvcache-ai/Mooncake/mooncake-ep/benchmarks/ep_benchmark/run_ep_benchmark.py code served (permissive licence) · get_code("94fe019e51e290c4") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI. It features a KVCache-centric disaggregated architecture that separates the prefill and decoding clusters. It also leverages the underutilized CPU, DRAM, and SSD resources of the GPU cluster to implement a disaggregated cache of KVCache. The core of Mooncake is its KVCache-centric scheduler, which balances maximizing overall effective throughput while meeting latency-related Service Level Objectives (SLOs). Unlike traditional studies that assume all requests will be processed, Mooncake faces challenges due to highly overloaded scenarios. To mitigate these, we developed a prediction-based early rejection policy. Experiments show that Mooncake excels in long-context scenarios. Compared to the baseline method, Mooncake can achieve up to a 525% increase in throughput in certain simulated scenarios while adhering to SLOs. Under real workloads, Mooncake's innovative architecture enables Kimi to handle 75% more requests.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2407.00079")
get_code_for_paper("2407.00079")
have("2407.00079")
Connect an agent — have() is free.