Bibliographic record
Abstract
Managed language runtimes such as the Java virtual machine (JVM) rely on just-in-time (JIT) compilers to improve application performance by converting bytecodes into optimized machine code. Unfortunately, JIT compilation introduces significant CPU and memory runtime overheads. These overheads are especially significant for modern cloud workloads where JVM instances are often short-lived and/or memory-constrained. JIT compiler disaggregation is a technique that decouples the JIT from the JVM and ships compilation to a separate remote process in order to improve application warm-up performance and reduce memory footprint. JITServer is an implementation of this approach in the Eclipse OpenJ9 JVM. Several optimizations proposed in this thesis improve JITServer performance in the multi-client setting and in environments with higher network latency. The performance evaluation presented in this thesis is the first study of remote JIT compilation in the context of cloud computing, demonstrating significant improvements in application warm-up speed and memory footprint. Caching compiled native code at the JITServer and its transparent reuse in multiple client JVMs running on different machines further improves warm-up performance and system-wide resource utilization, effectively increasing application density in the cloud. JITServer with caching compiled methods reduces overall CPU cost by up to 77%, overall memory usage by up to 62%, application start time by up to 58%, and warm-up time by up to 87% compared to local JIT compilation for a set of realistic web applications. Caching profiling data at the JITServer enables eager JIT compilation and prefetching server-cached compiled code, further improving application warm-up performance. Evaluation on a wide variety of application benchmarks shows reductions in runtime of short-running workloads of up to 24% and warm-up time of long-running ones of up to 80%.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".