Question 1

What is PRISM by Asymmetry Computing?

Accepted Answer

PRISM is Asymmetry Computing's GPU-native optimization stack for institutional finance. Public April 2026 evidence spans a corrected 20.3x cold-start result on 274 real S&P equities, a repeated-run transition family delivering 21×–125× speedup at 5,000 assets, GPU proof at 82,103 real assets, and a 100,004 real-asset structured benchmark solved in 2.16 seconds.

Question 2

How does PRISM perform against published baselines?

Accepted Answer

The public speed story now has multiple layers. Fresh April 2026 evidence shows PRISM at a corrected 20.3x cold-start result on 274 real S&P equities. The repeated-run 5,000-asset transition family reports 21×–125× speedups over a commercial reference baseline across four institutional workflows. At 82,103 real assets the GPU lane reports a 4.51×–42.14× speedup band, and the 100,004 real-asset structured benchmark solves in 2.16 seconds.

Question 3

How does PRISM compare to other solver options?

Accepted Answer

PRISM is positioned as a routed production stack rather than a single solver point result. The preserved transition suite, the corrected cold-start result, the 82,103-asset GPU proof lane, and the 100,004 real-asset structured benchmark together show where PRISM wins on latency, scale, and delivery. In the large-N field story, PRISM completes the 100k path while other accessible solvers time out well before that ceiling.

Question 4

What is the solution quality of PRISM?

Accepted Answer

The repeated-run institutional transition family stays inside a production-quality band below 0.01% versus the commercial reference baseline across all four workflows. The corrected April 2026 cold-start result on 274 real S&P equities shows effectively identical quality with a 0.0000% reported gap. PRISM is marketed around auditable, production-ready outputs rather than speed-only tradeoffs.

Question 5

What asset scale can PRISM handle?

Accepted Answer

Public April 2026 evidence now reaches a 100,004 real-asset benchmark cache spanning 47+ exchanges. The structured 100k benchmark solves in 2.16 seconds, and the walk-forward 100k suite reports Sharpe 0.74 with 1.82-second average solve time. The GPU proof lane is separately documented at 82,103 real assets.

Question 6

What is direct indexing and how does PRISM support it?

Accepted Answer

Direct indexing is the account-level construction of a personalized portfolio from individual securities rather than a pooled index fund. PRISM supports that workflow with the repeated-run transition family at 5,000 assets, 1.4-second headline account timing on the preserved rebalance lane, and the broader April 2026 scale evidence showing the stack can extend from account-level personalization up through very large benchmark universes.

Question 7

How is PRISM deployed and integrated?

Accepted Answer

PRISM is delivered as a REST API with dedicated cloud or VPC deployment options. The stack is described publicly as routed across CPU, GPU, and bounded discrete-refinement lanes, with audit-friendly outputs, replayable run records, and benchmark packs available through the matched-workload pilot program.

Question 8

What is PRISM-Q?

Accepted Answer

PRISM-Q is Asymmetry Computing's bounded discrete refinement layer for hard cardinality and combinatorial cleanup inside the broader PRISM stack. Public positioning now frames it as the premium refinement lane on top of the classical production core, with 15×–28× speedup versus commercial reference baseline MIQP on public cardinality-style subproblem benchmarks.

Question 9

What is a matched-workload pilot with Asymmetry Computing?

Accepted Answer

A matched-workload pilot benchmarks PRISM against your current solver on your own asset universe, constraint set, and institutional risk mapping inputs under the same timing and quality methodology used in the public benchmark materials. The deliverables include a benchmark pack, technical walk-through, and the option to map your workload onto the preserved 5,000-asset transition family, the corrected cold-start result, the 82,103-asset GPU lane, or the 100,004 real-asset structured benchmark.

PRISM Stack — Institutional Portfolio Optimization | Asymmetry Computing