Generated common-QP profiles
Compare construction-plus-solve timing with objective-gap and feasibility gates on explicitly generated workloads.
Inspect profiles & timingsResearch / original source
Read the method.
Inspect the evidence.
The original PRISM paper, by Debdoot Ghosh.
Submitted 22 June 2026.
Evidence register
Two tests.
Different acceptance gates.
Executed-trade comparisons
Use the frozen economic acceptance criteria for the relevant full-workflow record. Numerical agreement alone is a different claim.
Read the measurement protocolPublic materials
Follow the record
to its source.
01Public evaluation
Evaluation materials associated with the PRISM technical work.
02Direct-indexing evaluationThe separately scoped direct-indexing public evaluation repository.
03Historical studiesRead earlier workload records alongside their original scope.
04Benchmark dataDownload the public chart data and reported numerical gates.
Scope & evaluation
Generated numerical workloads, executed-trade economic parity, historical real-data studies, and fleet extrapolations answer different questions. Use the scope and acceptance criteria attached to the relevant record. Public reports do not expose PRISM’s proprietary implementation.
Sources / Citation
A reference you
can follow.
The original PRISM paper
Ghosh, Debdoot. Asymmetry PRISM: A CPU/GPU Portfolio Optimization Engine for Deadline-Bounded Institutional Rebalancing. arXiv:2606.23367, June 2026.
Open the paper record Review the current benchmark dataset
The original paper and the current website evidence are separately scoped records. Cite the result and methodology that support the specific claim, including the measurement conditions. The product overview identifies the company and its portfolio workflows.
Work with Asymmetry