Skip to main content

Shareable Lab Report demo

SMA 50/200 CPU vs CUDA proof.

This is the public shape of a Strategy Arena Lab Report: local benchmark proof first, Arena readiness second, chart review last.

Benchmark my Pine strategy Hall of Fame My Lab Reports Open GPU Arena Open in TradingView

Benchmark summary

CPU reference8.2s
CUDA run0.34s
Speedup24.1x
Variants10,000

Hardware and workload

GPUNVIDIA RTX class CUDA device
Backendcuda_sma_batch_v1
Pine scopeSupported subset

This demo is intentionally honest: Strategy Arena Lab is not a TradingView extension and does not claim full Pine runtime parity. It generates native ArenaScript research artifacts and can bridge supported Pine patterns into repeatable CPU/CUDA runs.

Source shape

strategy("SMA 50/200")
fast = ta.sma(close, 50)
slow = ta.sma(close, 200)
strategy.entry("Long", strategy.long, when = ta.crossover(fast, slow))
strategy.close("Long", when = ta.crossunder(fast, slow))

Next product step

The next implementation step is server-side persistence: desktop upload, signed workload hash, verification level, and a public report id linked from GPU Arena rows.

Back to Strategy Arena Lab Open Hall of Fame Build my report library Download Windows Lab