VALYNOX
VALYNOXSTRATEGY RESEARCH TECHNOLOGY
AlphaPrimeAlphaPrime
GPU-NATIVE / MEASURED

1.268 billion strategy evaluations.
18 hours, nonstop.

Measured on one RTX 3090 with a real research setup. Each completed backtest of a candidate strategy counts as one evaluation, and the same candidate can be evaluated more than once.

RTX 3090 / REAL RESEARCH SETUP

Under the test conditions below, AlphaPrime completed 1,268,200,000 strategy evaluations.

1,268,200,000Strategy evaluations completed
18:23:11Run time: one continuous run (h:min:s)
RTX 3090One NVIDIA graphics card (24 GB)
114,371Bars per evaluation
  • Test date: 2026-10-05.
  • Data: TY# (10-Year Treasury Note futures), 37-minute bars, 2015-01-01 to 2026-09-25, fees and slippage included.
  • Strategy setup: single entry and exit conditions, simple stop-loss and take-profit, one direction only; correlation filter on.
  • What is counted: the search can evaluate the same candidate more than once; this run generated 367,567,675 candidates. The run time is the wall-clock time of the whole run.

This figure is the number of strategy evaluations under the stated conditions. It is not the number of strategies that passed validation, and it does not indicate live trading profitability. Different data, settings and hardware affect the results.

AlphaPrime Chronograph as recorded
The Chronograph as recorded (RTX 3090, 2026-10-05): ② 1,268,200,000 strategy evaluations ③ 18 h 23 min 11 s elapsed; the other figures on screen are live readings at that moment. Strategy names and money figures are blurred.
CPU / GPU · ARCHITECTURE

The same research question.A different compute scale.

CPU

Dispatch to worker cores

Multiple CPU cores can work in parallel on assigned candidates.

ALPHAPRIME / GPU

Candidate batches, evaluated in parallel

CPU prepares the research task; GPU batches generation, backtesting and scoring.

Blocks on the left only illustrate CPU task distribution, not core counts or speedup.