2026-09-21T08:37:27Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1568152024-09-17T03:22:34Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Amarasinghe, Saman
Sollee III, Richard P.
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2024-09-16T13:50:53Z
2024-09-16T13:50:53Z
2024-05
2024-07-11T14:37:22.882Z
https://hdl.handle.net/1721.1/156815
0009-0007-4460-9597
The field of Lattice Quantum Chromodynamics faces massive scaling problems because of the large iteration spaces of the sums required which scale with the factorial of the number of atoms represented. The LQCD IR and rewrite system from this thesis allows tackling these scaling problems quicker and more effectively. The IR allows representing both mathematical concepts such as products and sums as well as algorithmic concepts such as precomputations. Our system requires minimal code to initialize the naive algorithm and apply effective rewrites to increase performance. This development time speedup allows trying various approaches with ease. The rewrite system allows correctness to be maintained at each step while being able to drastically change the algorithmic approach in search of better asymptotic bounds. Our approaches lead to up to 5x speedups and at worse 2x slowdowns for our most important problem, but with a better development cycle, requiring only 100s of SLOC compared to 1000s of SLOC.
M.Eng.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright retained by author(s)
https://rightsstatements.org/page/InC-EDU/1.0/
An Intermediate Representation for Expressing and Optimizing Computations in Lattice Quantum Chromodynamics
Thesis
application/pdf
Master
Master of Engineering in Electrical Engineering and Computer Science
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="b1998e41-84cc-4d10-916e-893b0a413a60">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>An Intermediate Representation for Expressing and Optimizing Computations in Lattice Quantum Chromodynamics</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2024-05</PublicationDate>
<Authors>
<Author>
<DisplayName>Sollee III, Richard P.</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>The field of Lattice Quantum Chromodynamics faces massive scaling problems because of the large iteration spaces of the sums required which scale with the factorial of the number of atoms represented. The LQCD IR and rewrite system from this thesis allows tackling these scaling problems quicker and more effectively. The IR allows representing both mathematical concepts such as products and sums as well as algorithmic concepts such as precomputations. Our system requires minimal code to initialize the naive algorithm and apply effective rewrites to increase performance. This development time speedup allows trying various approaches with ease. The rewrite system allows correctness to be maintained at each step while being able to drastically change the algorithmic approach in search of better asymptotic bounds. Our approaches lead to up to 5x speedups and at worse 2x slowdowns for our most important problem, but with a better development cycle, requiring only 100s of SLOC compared to 1000s of SLOC.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>