REVIEW 9 cited by
ConServe: Fine-Grained GPU Harvesting for LLM Online and Offline Co-Serving
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
ConServe: Fine-Grained GPU Harvesting for LLM Online and Offline Co-Serving
read the original abstract
Large language model (LLM) serving demands low latency and high throughput, but high load variability makes it challenging to achieve high GPU utilization. In this paper, we identify a synergetic but overlooked opportunity to co-serve latency-critical online requests alongside latency-tolerant offline tasks such as model benchmarking. While promising, existing serving systems fail to co-serve them efficiently, as their coarse-grained resource management at the request or iteration level cannot harvest millisecond-level GPU idle cycles without introducing interference that violates online latency objectives. ConServe is a new LLM co-serving system that achieves high throughput and strong online latency guarantees by managing resources at finer granularities. ConServe introduces three techniques: (1) a latency-aware token-level scheduler that precisely sizes offline batches and tokens to fit within online latency objectives; (2) sub-iteration, layer-wise preemption that allows offline tasks to yield to online load spikes; and (3) incremental KV cache management that enables preempting and resuming offline requests at near-zero cost. Evaluations with Llama-3.1 and Qwen-2.5 models on real-world workloads show that ConServe delivers an average of 2.2$\times$ higher throughput and reduces online serving tail latency by 2.9$\times$ on average compared to state-of-the-art systems.
Forward citations
Cited by 9 Pith papers
-
RW-TTT: Batched Serving for Request-Owned Test-Time Training State
RW-TTT enables batched serving for request-owned TTT state via owner/version tagging and selective batching of compatible phases, reaching 274.61 tok/s and 9.31x speedup on one GPU with eight InPlace-TTT streams while...
-
Beyond Greedy Chunking: SLO-Aware Sliding-Window Scheduling for LLM Inference
SlidingServe achieves up to 30% higher service capacity and 16-53% fewer SLO violations in LLM inference by using dynamic chunking and priority-based batch construction.
-
ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
ROSE delivers 1.2-3.3x higher end-to-end throughput for agentic RL by safely co-using underutilized serving GPUs for rollouts while meeting serving SLOs.
-
ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
ROSE is a system for cooperative elasticity that co-locates serving and rollout models on shared GPUs, delivering 1.3-3.3x higher end-to-end throughput than fixed-resource baselines while preserving serving SLOs.
-
Valve: Production Online-Offline Inference Colocation with Jointly-Bounded Preemption Latency and Rate
Valve jointly bounds preemption latency and rate for online-offline LLM colocation on GPUs, delivering 34.6% higher cluster utilization and a 2,170-GPU saving in a production deployment of 8,054 GPUs with under 5% TTF...
-
Blink: CPU-Free LLM Inference by Delegating the Serving Stack to GPU and SmartNIC
Blink enables CPU-free LLM inference via SmartNIC offload and persistent GPU kernel, delivering up to 8.47x lower P99 TTFT, 3.4x lower P99 TPOT, 2.1x higher decode throughput, and 48.6% lower energy per token while re...
-
Foundry: Template-Based CUDA Graph Context Materialization for Fast LLM Serving Cold Start
Foundry uses template-based CUDA graph context materialization to reduce LLM serving cold-start latency by up to 99% while preserving CUDA graph throughput gains.
-
The Energy Cost of Execution-Idle in GPU Clusters
Execution-idle accounts for 19.7% of GPU execution time and 10.7% of energy in a large cluster, motivating power management that treats it as a distinct operating state.
-
Parallelizing Tool Execution and LLM Generation for Low-Latency Agent Serving
Pattern-aware speculative tool execution cuts agent end-to-end latency by roughly half and observed tool latency by about 1.8× by overlapping predicted tools with LLM generation.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.