SuperInfer: SLO-Aware Rotary Scheduling and Memory Management for LLM Inference on Superchips

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Yu, Jiahuan, Hu, Mingtao, Lin, Zichao, Zhang, Minjia
Format:	Preprint
Publié:	2026
Sujets:	Distributed, Parallel, and Cluster Computing Artificial Intelligence Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

SuperOffload: Unleashing the Power of Large-Scale LLM Training on Superchips
par: Lian, Xinyu, et autres
Publié: (2025)

VoltanaLLM: Feedback-Driven Frequency Control and State-Space Routing for Energy-Efficient LLM Serving
par: Yu, Jiahuan, et autres
Publié: (2025)

CoLLM: Continuous Adaptation for SLO-Aware LLM Serving on Shared GPU Clusters
par: Huang, Shaoyuan, et autres
Publié: (2026)

A Framework for SLO, Carbon, and Wastewater-Aware Sustainable FaaS Cloud Platform Management
par: Qi, Sirui, et autres
Publié: (2024)

SLO-Aware Scheduling for Large Language Model Inferences
par: Huang, Jinqi, et autres
Publié: (2025)

Taming Request Imbalance: SLO-Aware Scheduling for Disaggregated LLM Inference
par: Wang, Qipeng
Publié: (2026)

FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving
par: Ye, Zihao, et autres
Publié: (2025)

AdaServe: Accelerating Multi-SLO LLM Serving with SLO-Customized Speculative Decoding
par: Li, Zikun, et autres
Publié: (2025)

TAPAS: Thermal- and Power-Aware Scheduling for LLM Inference in Cloud Platforms
par: Stojkovic, Jovan, et autres
Publié: (2025)

SLO-aware GPU Frequency Scaling for Energy Efficient LLM Inference Serving
par: Kakolyris, Andreas Kosmas, et autres
Publié: (2024)

SLICE: SLO-Driven Scheduling for LLM Inference on Edge Computing Devices
par: Chow, Will
Publié: (2025)

Optimizing LLM Inference: Fluid-Guided Online Scheduling with Memory Constraints
par: Ao, Ruicheng, et autres
Publié: (2025)

NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference
par: Jiang, Xuanlin, et autres
Publié: (2024)

Fairness-Aware Job Scheduling for Multi-Job Federated Learning
par: Shi, Yuxin, et autres
Publié: (2024)

Sustainable Carbon-Aware and Water-Efficient LLM Scheduling in Geo-Distributed Cloud Datacenters
par: Moore, Hayden, et autres
Publié: (2025)

Prompt-Aware Scheduling for Low-Latency LLM Serving
par: Tao, Yiheng, et autres
Publié: (2025)

MACE: A Hybrid LLM Serving System with Colocated SLO-aware Continuous Retraining Alignment
par: Li, Yufei, et autres
Publié: (2025)

Harli: SLO-Aware Co-location of LLM Inference and PEFT-based Finetuning on Model-as-a-Service Platforms
par: Xu, Ao, et autres
Publié: (2025)

TRAIL: Trust-Aware Client Scheduling for Semi-Decentralized Federated Learning
par: Hu, Gangqiang, et autres
Publié: (2024)

SLO-Aware Compute Resource Allocation for Prefill-Decode Disaggregated LLM Inference
par: Li, Luchang, et autres
Publié: (2026)

Beluga: A CXL-Based Memory Architecture for Scalable and Efficient LLM KVCache Management
par: Yang, Xinjun, et autres
Publié: (2025)

Llamas on the Web: Memory-Efficient, Performance-Portable, and Multi-Precision LLM Inference with WebGPU
par: Levine, Reese, et autres
Publié: (2026)

SCOOT: SLO-Oriented Performance Tuning for LLM Inference Engines
par: Cheng, Ke, et autres
Publié: (2024)

HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling
par: Yousefijamarani, Zahra, et autres
Publié: (2025)

Justitia: Fair and Efficient Scheduling of Task-parallel LLM Agents with Selective Pampering
par: Yang, Mingyan, et autres
Publié: (2025)

Cost-Efficient Multimodal LLM Inference via Cross-Tier GPU Heterogeneity
par: Yu, Donglin
Publié: (2026)

FairBatching: Fairness-Aware Batch Formation for LLM Inference
par: Lyu, Hongtao, et autres
Publié: (2025)

Stochastic Sparse Attention for Memory-Bound Inference
par: Lee, Kyle, et autres
Publié: (2026)

KunServe: Parameter-centric Memory Management for Efficient Memory Overloading Handling in LLM Serving
par: Cheng, Rongxin, et autres
Publié: (2024)

LLM-PQ: Serving LLM on Heterogeneous Clusters with Phase-Aware Partition and Adaptive Quantization
par: Zhao, Juntao, et autres
Publié: (2024)

Token-Budget-Aware Pool Routing for Cost-Efficient LLM Inference
par: Chen, Huamin, et autres
Publié: (2026)

Dooly: Configuration-Agnostic, Redundancy-Aware Profiling for LLM Inference Simulation
par: Kim, Joon Ha, et autres
Publié: (2026)

Efficient MoE Inference with Fine-Grained Scheduling of Disaggregated Expert Parallelism
par: Pan, Xinglin, et autres
Publié: (2025)

Application of Machine Learning Optimization in Cloud Computing Resource Scheduling and Management
par: Zhang, Yifan, et autres
Publié: (2024)

ELIS: Efficient LLM Iterative Scheduling System with Response Length Predictor
par: Choi, Seungbeom, et autres
Publié: (2025)

ExpertFlow: Adaptive Expert Scheduling and Memory Coordination for Efficient MoE Inference
par: Shen, Zixu, et autres
Publié: (2025)

ProTrain: Efficient LLM Training via Memory-Aware Techniques
par: Yang, Hanmei, et autres
Publié: (2024)

LLM-42: Enabling Determinism in LLM Inference with Verified Speculation
par: Gond, Raja, et autres
Publié: (2026)

Semantic Parallelism: Redefining Efficient MoE Inference via Model-Data Co-Scheduling
par: Li, Yan, et autres
Publié: (2025)

Memory Offloading for Large Language Model Inference with Latency SLO Guarantees
par: Ma, Chenxiang, et autres
Publié: (2025)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
VUFIND_SESSION	"74cbe74f4ffa95a2e096cd3c146b3f49"
language	"fr"
ui	"standard"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790867570.4864 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "74cbe74f4ffa95a2e096cd3c146b3f49" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"VUFIND_SESSION=74cbe74f4ffa95a2e096cd3c146b3f49; language=fr; ui=standard"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.216.147"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"36968"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2601_20309/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2601_20309/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790867570.4818
REQUEST_TIME	1790867570

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler