SynerDiff: Synergetic Continuous Batching for Fast and Parallel Diffusion Model Inference

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Zhou, Ziqi, Yang, Peng, Liang, Yuxin, Liu, Mingliu, Lu, Jia
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Flash Communication: Reducing Tensor Parallelization Bottleneck for Fast Large Language Model Inference
by: Li, Qingyuan, et al.
Published: (2024)

DiffAgent: Fast and Accurate Text-to-Image API Selection with Large Language Model
by: Zhao, Lirui, et al.
Published: (2024)

BlockBatch: Multi-Scale Consensus Decoding for Efficient Diffusion Language Model Inference
by: Wu, Xiaoyou, et al.
Published: (2026)

AdaDiff: Adaptive Step Selection for Fast Diffusion Models
by: Zhang, Hui, et al.
Published: (2023)

BatchLLM: Optimizing Large Batched LLM Inference with Global Prefix Sharing and Throughput-oriented Token Batching
by: Zheng, Zhen, et al.
Published: (2024)

DiffScale: Continuous Downscaling and Bias Correction of Subseasonal Wind Speed Forecasts using Diffusion Models
by: Springenberg, Maximilian, et al.
Published: (2025)

DiffGRM: Diffusion-based Generative Recommendation Model
by: Liu, Zhao, et al.
Published: (2025)

FairBatching: Fairness-Aware Batch Formation for LLM Inference
by: Lyu, Hongtao, et al.
Published: (2025)

DiffLoad: Uncertainty Quantification in Electrical Load Forecasting with the Diffusion Model
by: Wang, Zhixian, et al.
Published: (2023)

DiffCAP: Diffusion-based Cumulative Adversarial Purification for Vision Language Models
by: Fu, Jia, et al.
Published: (2025)

Threshold-Based Exclusive Batching for LLM Inference
by: Zhang, Weifang, et al.
Published: (2026)

Diff-PIC: Revolutionizing Particle-In-Cell Nuclear Fusion Simulation with Diffusion Models
by: Liu, Chuan, et al.
Published: (2024)

AnimateDiff-Lightning: Cross-Model Diffusion Distillation
by: Lin, Shanchuan, et al.
Published: (2024)

Stage-Diff: Stage-wise Long-Term Time Series Generation Based on Diffusion Models
by: Hou, Xuan, et al.
Published: (2025)

EPIC: Efficient and Parallel Inference under CFG Constraints for Diffusion Language Models
by: Jin, Hyundong, et al.
Published: (2026)

DisCo-Diff: Enhancing Continuous Diffusion Models with Discrete Latents
by: Xu, Yilun, et al.
Published: (2024)

Fast On-device LLM Inference with NPUs
by: Xu, Daliang, et al.
Published: (2024)

PerturbDiff: Functional Diffusion for Single-Cell Perturbation Modeling
by: Yuan, Xinyu, et al.
Published: (2026)

Fast Inference for Augmented Large Language Models
by: Shahout, Rana, et al.
Published: (2024)

SimDiff: Simpler Yet Better Diffusion Model for Time Series Point Forecasting
by: Ding, Hang, et al.
Published: (2025)

AnchorDiff: Topology-Aware Masked Diffusion with Confidence-based Rewriting for Radiology Report Generation
by: Yu, Shiying, et al.
Published: (2026)

Why Diffusion Language Models Struggle with Truly Parallel (Non-Autoregressive) Decoding?
by: Li, Pengxiang, et al.
Published: (2026)

RealignDiff: Boosting Text-to-Image Diffusion Model with Coarse-to-fine Semantic Re-alignment
by: Jiang, Zutao, et al.
Published: (2023)

DepCap: Adaptive Block-Wise Parallel Decoding for Efficient Diffusion LM Inference
by: Xia, Xiang, et al.
Published: (2026)

Resource-Efficient Generative AI Model Deployment in Mobile Edge Networks
by: Liang, Yuxin, et al.
Published: (2024)

Fast Monte Carlo Tree Diffusion: 100x Speedup via Parallel Sparse Planning
by: Yoon, Jaesik, et al.
Published: (2025)

MacDiff: Unified Skeleton Modeling with Masked Conditional Diffusion
by: Wu, Lehong, et al.
Published: (2024)

DiffGraph: Heterogeneous Graph Diffusion Model
by: Li, Zongwei, et al.
Published: (2025)

Edge Intelligence Optimization for Large Language Model Inference with Batching and Quantization
by: Zhang, Xinyuan, et al.
Published: (2024)

Phys-Diff: A Physics-Inspired Latent Diffusion Model for Tropical Cyclone Forecasting
by: Liu, Lei, et al.
Published: (2026)

Iterative Learning Control-Informed Reinforcement Learning for Batch Process Control
by: Lin, Runze, et al.
Published: (2026)

GlobeDiff: State Diffusion Process for Partial Observability in Multi-Agent Systems
by: Yang, Yiqin, et al.
Published: (2026)

LocDiff: Identifying Locations on Earth by Diffusing in the Hilbert Space
by: Wang, Zhangyu, et al.
Published: (2025)

DiffSpectra: Molecular Structure Elucidation from Spectra using Diffusion Models
by: Wang, Liang, et al.
Published: (2025)

Elastic Multi-Gradient Descent for Parallel Continual Learning
by: Lyu, Fan, et al.
Published: (2024)

AlignDiff: Aligning Diverse Human Preferences via Behavior-Customisable Diffusion Model
by: Dong, Zibin, et al.
Published: (2023)

DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers
by: Shi, Minglei, et al.
Published: (2025)

DiffCSS: Diverse and Expressive Conversational Speech Synthesis with Diffusion Models
by: wu, Weihao, et al.
Published: (2025)

An Aligned Constraint Programming Model For Serial Batch Scheduling With Minimum Batch Size
by: Huertas, Jorge A., et al.
Published: (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	"4b47a5e42abf584addcc355555fee815"
ui	"standard"
language	"en"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790953183.9897 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "4b47a5e42abf584addcc355555fee815" ]`]
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=4b47a5e42abf584addcc355555fee815; ui=standard; language=en"
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	"216.73.217.142"
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	"64294"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2605_08835/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2605_08835/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790953183.9812
REQUEST_TIME	1790953183

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
VUFIND_SESSION	"4b47a5e42abf584addcc355555fee815"
ui	"standard"
language	"en"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790953183.9897 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "4b47a5e42abf584addcc355555fee815" ]`]
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=4b47a5e42abf584addcc355555fee815; ui=standard; language=en"
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	"216.73.217.142"
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	"64294"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2605_08835/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2605_08835/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790953183.9812
REQUEST_TIME	1790953183

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler

Similar Items

Similar Items

Environment & details:

Environment & details: