Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs Article Swipe

PDF

Bodhisatwa Chatterjee , Neeraj Jadhav , Sharjeel Khan , Santosh Pande ·

YOU? · · 2024 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2412.06994

Application profiling is an indispensable technique for many software development tasks, such as code and memory layout optimizations, where optimization decisions are tailored to specific program profiles. Unfortunately, modern application codebases exhibit highly variant behavior across different inputs, creating challenges for conventional profiling approaches that rely on a single representative execution instance. In this paper, we propose \textbf{Phaedrus}, a new \textit{compiler-assisted deep learning framework} designed to predict dynamic program behaviors across varied execution instances, specifically focusing on dynamic function call prediction.Such predicted call sequences are then used for producing optimized code pertinent to a given input. Traditional profile-guided optimization methods struggle with the input-dependent variability of modern applications, where profiling on different inputs yields divergent application behaviors. To address this, Phaedrus proposes two new approaches: \textit{Application Behavior Synthesis}, a profile-less approach where Large Language Models (LLMs) directly infer dynamic functions based on source code \& static compiler analysis, bypassing the need for traditional profiling, and \textit{Application Profile Generalization}, which uses generative models trained on compressed and augmented \textit{Whole Program Path} (WPP) based function profiles to predict application behavior under unseen inputs. Our experiments show that \textit{Phaedrus} can achieve upto $10^7X$ reduction in WPP function profile sizes, can predict most frequently executed functions that cover upto 85-99\% of the execution time, along with an average of 13.19\% (upto 65\%) reduction in application binary size, and an average of 6.08\% (upto 20\%) performance improvement over the traditional profile-guided optimization, without any execution.

Related Topics

Generative Grammar

Computer Science

Artificial Intelligence

Concepts

Generative grammar Computer science Generative model Artificial intelligence

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2412.06994
PDF: https://arxiv.org/pdf/2412.06994
OA Status: green
Related Works: 10
OpenAlex ID: https://openalex.org/W4405254837

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4405254837

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2412.06994

Digital Object Identifier
Title: Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2024

Year of publication
Publication date: 2024-12-09

Full publication date if available
Authors: Bodhisatwa Chatterjee, Neeraj Jadhav, Sharjeel Khan, Santosh Pande

List of authors in order
Landing page: https://arxiv.org/abs/2412.06994

Publisher landing page
PDF URL: https://arxiv.org/pdf/2412.06994

Direct link to full text PDF
Open access: Yes

Whether a free full text is available
OA status: green

Open access status per OpenAlex
OA URL: https://arxiv.org/pdf/2412.06994

Direct OA link when available
Concepts: Generative grammar, Computer science, Generative model, Artificial intelligence

Top concepts (fields/topics) attached by OpenAlex
Cited by: 0

Total citation count in OpenAlex
Related works (count): 10

Other works algorithmically related by OpenAlex

Full payload

id	https://openalex.org/W4405254837
doi	https://doi.org/10.48550/arxiv.2412.06994
ids.doi	https://doi.org/10.48550/arxiv.2412.06994
ids.openalex	https://openalex.org/W4405254837
fwci
type	preprint
title	Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T10215
topics[0].field.id	https://openalex.org/fields/17
topics[0].field.display_name	Computer Science
topics[0].score	0.934499979019165
topics[0].domain.id	https://openalex.org/domains/3
topics[0].domain.display_name	Physical Sciences
topics[0].subfield.id	https://openalex.org/subfields/1702
topics[0].subfield.display_name	Artificial Intelligence
topics[0].display_name	Semantic Web and Ontologies
topics[1].id	https://openalex.org/T12720
topics[1].field.id	https://openalex.org/fields/33
topics[1].field.display_name	Social Sciences
topics[1].score	0.9028000235557556
topics[1].domain.id	https://openalex.org/domains/2
topics[1].domain.display_name	Social Sciences
topics[1].subfield.id	https://openalex.org/subfields/3312
topics[1].subfield.display_name	Sociology and Political Science
topics[1].display_name	Multimedia Communication and Technology
is_xpac	False
apc_list
apc_paid
concepts[0].id	https://openalex.org/C39890363
concepts[0].level	2
concepts[0].score	0.7601996660232544
concepts[0].wikidata	https://www.wikidata.org/wiki/Q36108
concepts[0].display_name	Generative grammar
concepts[1].id	https://openalex.org/C41008148
concepts[1].level	0
concepts[1].score	0.5965513586997986
concepts[1].wikidata	https://www.wikidata.org/wiki/Q21198
concepts[1].display_name	Computer science
concepts[2].id	https://openalex.org/C167966045
concepts[2].level	3
concepts[2].score	0.42773938179016113
concepts[2].wikidata	https://www.wikidata.org/wiki/Q5532625
concepts[2].display_name	Generative model
concepts[3].id	https://openalex.org/C154945302
concepts[3].level	1
concepts[3].score	0.2872454524040222
concepts[3].wikidata	https://www.wikidata.org/wiki/Q11660
concepts[3].display_name	Artificial intelligence
keywords[0].id	https://openalex.org/keywords/generative-grammar
keywords[0].score	0.7601996660232544
keywords[0].display_name	Generative grammar
keywords[1].id	https://openalex.org/keywords/computer-science
keywords[1].score	0.5965513586997986
keywords[1].display_name	Computer science
keywords[2].id	https://openalex.org/keywords/generative-model
keywords[2].score	0.42773938179016113
keywords[2].display_name	Generative model
keywords[3].id	https://openalex.org/keywords/artificial-intelligence
keywords[3].score	0.2872454524040222
keywords[3].display_name	Artificial intelligence
language	en
locations[0].id	pmh:oai:arXiv.org:2412.06994
locations[0].is_oa	True
locations[0].source.id	https://openalex.org/S4306400194
locations[0].source.issn
locations[0].source.type	repository
locations[0].source.is_oa	True
locations[0].source.issn_l
locations[0].source.is_core	False
locations[0].source.is_in_doaj	False
locations[0].source.display_name	arXiv (Cornell University)
locations[0].source.host_organization	https://openalex.org/I205783295
locations[0].source.host_organization_name	Cornell University
locations[0].source.host_organization_lineage	https://openalex.org/I205783295
locations[0].license
locations[0].pdf_url	https://arxiv.org/pdf/2412.06994
locations[0].version	submittedVersion
locations[0].raw_type	text
locations[0].license_id
locations[0].is_accepted	False
locations[0].is_published	False
locations[0].raw_source_name
locations[0].landing_page_url	http://arxiv.org/abs/2412.06994
locations[1].id	doi:10.48550/arxiv.2412.06994
locations[1].is_oa	True
locations[1].source.id	https://openalex.org/S4306400194
locations[1].source.issn
locations[1].source.type	repository
locations[1].source.is_oa	True
locations[1].source.issn_l
locations[1].source.is_core	False
locations[1].source.is_in_doaj	False
locations[1].source.display_name	arXiv (Cornell University)
locations[1].source.host_organization	https://openalex.org/I205783295
locations[1].source.host_organization_name	Cornell University
locations[1].source.host_organization_lineage	https://openalex.org/I205783295
locations[1].license	cc-by
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id	https://openalex.org/licenses/cc-by
locations[1].is_accepted	False
locations[1].is_published
locations[1].raw_source_name
locations[1].landing_page_url	https://doi.org/10.48550/arxiv.2412.06994
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5080183088
authorships[0].author.orcid	https://orcid.org/0000-0003-3098-6256
authorships[0].author.display_name	Bodhisatwa Chatterjee
authorships[0].author_position	first
authorships[0].raw_author_name	Chatterjee, Bodhisatwa
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5115087507
authorships[1].author.orcid
authorships[1].author.display_name	Neeraj Jadhav
authorships[1].author_position	middle
authorships[1].raw_author_name	Jadhav, Neeraj
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5009383888
authorships[2].author.orcid	https://orcid.org/0000-0002-4563-4619
authorships[2].author.display_name	Sharjeel Khan
authorships[2].author_position	middle
authorships[2].raw_author_name	Khan, Sharjeel
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5061235810
authorships[3].author.orcid	https://orcid.org/0000-0001-6723-8062
authorships[3].author.display_name	Santosh Pande
authorships[3].author_position	last
authorships[3].raw_author_name	Pande, Santosh
authorships[3].is_corresponding	False
has_content.pdf	False
has_content.grobid_xml	False
is_paratext	False
open_access.is_oa	True
open_access.oa_url	https://arxiv.org/pdf/2412.06994
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2024-12-12T00:00:00
display_name	Phaedrus: Predicting Dynamic Application Behavior with Lightweight Generative Models and LLMs
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T10215
primary_topic.field.id	https://openalex.org/fields/17
primary_topic.field.display_name	Computer Science
primary_topic.score	0.934499979019165
primary_topic.domain.id	https://openalex.org/domains/3
primary_topic.domain.display_name	Physical Sciences
primary_topic.subfield.id	https://openalex.org/subfields/1702
primary_topic.subfield.display_name	Artificial Intelligence
primary_topic.display_name	Semantic Web and Ontologies
related_works	https://openalex.org/W4365211920, https://openalex.org/W3014948380, https://openalex.org/W4391584540, https://openalex.org/W4380551139, https://openalex.org/W4317695495, https://openalex.org/W4395044357, https://openalex.org/W4287117424, https://openalex.org/W4387506531, https://openalex.org/W2087346071, https://openalex.org/W2967848559
cited_by_count	0
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2412.06994
best_oa_location.is_oa	True
best_oa_location.source.id	https://openalex.org/S4306400194
best_oa_location.source.issn
best_oa_location.source.type	repository
best_oa_location.source.is_oa	True
best_oa_location.source.issn_l
best_oa_location.source.is_core	False
best_oa_location.source.is_in_doaj	False
best_oa_location.source.display_name	arXiv (Cornell University)
best_oa_location.source.host_organization	https://openalex.org/I205783295
best_oa_location.source.host_organization_name	Cornell University
best_oa_location.source.host_organization_lineage	https://openalex.org/I205783295
best_oa_location.license
best_oa_location.pdf_url	https://arxiv.org/pdf/2412.06994
best_oa_location.version	submittedVersion
best_oa_location.raw_type	text
best_oa_location.license_id
best_oa_location.is_accepted	False
best_oa_location.is_published	False
best_oa_location.raw_source_name
best_oa_location.landing_page_url	http://arxiv.org/abs/2412.06994
primary_location.id	pmh:oai:arXiv.org:2412.06994
primary_location.is_oa	True
primary_location.source.id	https://openalex.org/S4306400194
primary_location.source.issn
primary_location.source.type	repository
primary_location.source.is_oa	True
primary_location.source.issn_l
primary_location.source.is_core	False
primary_location.source.is_in_doaj	False
primary_location.source.display_name	arXiv (Cornell University)
primary_location.source.host_organization	https://openalex.org/I205783295
primary_location.source.host_organization_name	Cornell University
primary_location.source.host_organization_lineage	https://openalex.org/I205783295
primary_location.license
primary_location.pdf_url	https://arxiv.org/pdf/2412.06994
primary_location.version	submittedVersion
primary_location.raw_type	text
primary_location.license_id
primary_location.is_accepted	False
primary_location.is_published	False
primary_location.raw_source_name
primary_location.landing_page_url	http://arxiv.org/abs/2412.06994
publication_date	2024-12-09
publication_year	2024
referenced_works_count	0
abstract_inverted_index.a	47, 58, 93, 128
abstract_inverted_index.In	52
abstract_inverted_index.To	117
abstract_inverted_index.an	3, 212, 224
abstract_inverted_index.as	12
abstract_inverted_index.in	191, 219
abstract_inverted_index.is	2
abstract_inverted_index.of	105, 206, 214, 226
abstract_inverted_index.on	46, 76, 110, 141, 163
abstract_inverted_index.to	23, 65, 92, 174
abstract_inverted_index.we	55
abstract_inverted_index.Our	181
abstract_inverted_index.WPP	192
abstract_inverted_index.and	14, 154, 165, 223
abstract_inverted_index.any	238
abstract_inverted_index.are	21, 84
abstract_inverted_index.can	186, 196
abstract_inverted_index.for	6, 40, 87, 151
abstract_inverted_index.new	59, 123
abstract_inverted_index.the	102, 149, 207, 233
abstract_inverted_index.two	122
abstract_inverted_index.call	79, 82
abstract_inverted_index.code	13, 90, 143
abstract_inverted_index.deep	61
abstract_inverted_index.many	7
abstract_inverted_index.most	198
abstract_inverted_index.need	150
abstract_inverted_index.over	232
abstract_inverted_index.rely	45
abstract_inverted_index.show	183
abstract_inverted_index.such	11
abstract_inverted_index.that	44, 184, 202
abstract_inverted_index.then	85
abstract_inverted_index.this	53
abstract_inverted_index.upto	188, 204
abstract_inverted_index.used	86
abstract_inverted_index.uses	159
abstract_inverted_index.with	101, 211
abstract_inverted_index.(WPP)	170
abstract_inverted_index.(upto	216, 228
abstract_inverted_index.20\%)	229
abstract_inverted_index.65\%)	217
abstract_inverted_index.Large	132
abstract_inverted_index.Path}	169
abstract_inverted_index.along	210
abstract_inverted_index.based	140, 171
abstract_inverted_index.cover	203
abstract_inverted_index.given	94
abstract_inverted_index.infer	137
abstract_inverted_index.size,	222
abstract_inverted_index.this,	119
abstract_inverted_index.time,	209
abstract_inverted_index.under	178
abstract_inverted_index.where	18, 108, 131
abstract_inverted_index.which	158
abstract_inverted_index.(LLMs)	135
abstract_inverted_index.6.08\%	227
abstract_inverted_index.Models	134
abstract_inverted_index.\&	144
abstract_inverted_index.across	35, 70
abstract_inverted_index.binary	221
abstract_inverted_index.highly	32
abstract_inverted_index.input.	95
abstract_inverted_index.inputs	112
abstract_inverted_index.layout	16
abstract_inverted_index.memory	15
abstract_inverted_index.models	161
abstract_inverted_index.modern	28, 106
abstract_inverted_index.paper,	54
abstract_inverted_index.single	48
abstract_inverted_index.sizes,	195
abstract_inverted_index.source	142
abstract_inverted_index.static	145
abstract_inverted_index.tasks,	10
abstract_inverted_index.unseen	179
abstract_inverted_index.varied	71
abstract_inverted_index.yields	113
abstract_inverted_index.$10^7X$	189
abstract_inverted_index.13.19\%	215
abstract_inverted_index.85-99\%	205
abstract_inverted_index.Profile	156
abstract_inverted_index.Program	168
abstract_inverted_index.achieve	187
abstract_inverted_index.address	118
abstract_inverted_index.average	213, 225
abstract_inverted_index.dynamic	67, 77, 138
abstract_inverted_index.exhibit	31
abstract_inverted_index.inputs,	37
abstract_inverted_index.inputs.	180
abstract_inverted_index.methods	99
abstract_inverted_index.predict	66, 175, 197
abstract_inverted_index.profile	194
abstract_inverted_index.program	25, 68
abstract_inverted_index.propose	56
abstract_inverted_index.trained	162
abstract_inverted_index.variant	33
abstract_inverted_index.without	237
abstract_inverted_index.Behavior	126
abstract_inverted_index.Language	133
abstract_inverted_index.Phaedrus	120
abstract_inverted_index.approach	130
abstract_inverted_index.behavior	34, 177
abstract_inverted_index.compiler	146
abstract_inverted_index.creating	38
abstract_inverted_index.designed	64
abstract_inverted_index.directly	136
abstract_inverted_index.executed	200
abstract_inverted_index.focusing	75
abstract_inverted_index.function	78, 172, 193
abstract_inverted_index.learning	62
abstract_inverted_index.profiles	173
abstract_inverted_index.proposes	121
abstract_inverted_index.software	8
abstract_inverted_index.specific	24
abstract_inverted_index.struggle	100
abstract_inverted_index.tailored	22
abstract_inverted_index.analysis,	147
abstract_inverted_index.augmented	166
abstract_inverted_index.behaviors	69
abstract_inverted_index.bypassing	148
abstract_inverted_index.codebases	30
abstract_inverted_index.decisions	20
abstract_inverted_index.different	36, 111
abstract_inverted_index.divergent	114
abstract_inverted_index.execution	50, 72, 208
abstract_inverted_index.functions	139, 201
abstract_inverted_index.instance.	51
abstract_inverted_index.optimized	89
abstract_inverted_index.pertinent	91
abstract_inverted_index.predicted	81
abstract_inverted_index.producing	88
abstract_inverted_index.profiles.	26
abstract_inverted_index.profiling	1, 42, 109
abstract_inverted_index.reduction	190, 218
abstract_inverted_index.sequences	83
abstract_inverted_index.technique	5
abstract_inverted_index.approaches	43
abstract_inverted_index.behaviors.	116
abstract_inverted_index.challenges	39
abstract_inverted_index.compressed	164
abstract_inverted_index.execution.	239
abstract_inverted_index.framework}	63
abstract_inverted_index.frequently	199
abstract_inverted_index.generative	160
abstract_inverted_index.instances,	73
abstract_inverted_index.profiling,	153
abstract_inverted_index.Application	0
abstract_inverted_index.Synthesis},	127
abstract_inverted_index.Traditional	96
abstract_inverted_index.application	29, 115, 176, 220
abstract_inverted_index.approaches:	124
abstract_inverted_index.development	9
abstract_inverted_index.experiments	182
abstract_inverted_index.improvement	231
abstract_inverted_index.performance	230
abstract_inverted_index.traditional	152, 234
abstract_inverted_index.variability	104
abstract_inverted_index.conventional	41
abstract_inverted_index.optimization	19, 98
abstract_inverted_index.profile-less	129
abstract_inverted_index.specifically	74
abstract_inverted_index.\textit{Whole	167
abstract_inverted_index.applications,	107
abstract_inverted_index.indispensable	4
abstract_inverted_index.optimization,	236
abstract_inverted_index.Unfortunately,	27
abstract_inverted_index.optimizations,	17
abstract_inverted_index.profile-guided	97, 235
abstract_inverted_index.representative	49
abstract_inverted_index.input-dependent	103
abstract_inverted_index.prediction.Such	80
abstract_inverted_index.Generalization},	157
abstract_inverted_index.\textit{Phaedrus}	185
abstract_inverted_index.\textbf{Phaedrus},	57
abstract_inverted_index.\textit{Application	125, 155
abstract_inverted_index.\textit{compiler-assisted	60
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	4
citation_normalized_percentile