XY-Serve is an Ascend-native serving system that uses virtual padding and unified meta-kernels to smooth dynamic workloads, reporting up to 89% throughput gains over Ascend-vLLM.
https://flashinfer.ai/ 2024/02/02/introduce-flashinfer.html
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Tackling the Dynamicity in a Production LLM Serving System with SOTA Optimizations via Hybrid Prefill/Decode/Verify Scheduling on Efficient Meta-kernels
XY-Serve is an Ascend-native serving system that uses virtual padding and unified meta-kernels to smooth dynamic workloads, reporting up to 89% throughput gains over Ascend-vLLM.