Throwing this into bt or backtesting.py to accelerate backtesting?

[BradKML]

Currently there are a few well-established backtesters out there that are not necessarily optimized for large-sized backtesting, but his repo seems promising with GPU acceleration. Ditto for technical analysis libraries that runs on pandas instead of optimized on a lower level.

[Heerozh]

this project mainly pursues performance.
squeeze performance often requires redesigning the architecture,
design acceleration for other project has limited effect

[BradKML]

I think the projects in question are also hoping for performance increases by shaking up their own architecture due to slowing development. They do however cover a broader range of backtesting techniques and have established userbase.

[Heerozh]

highly impossible, gpu acceleration has strict restrictions, and all algorithms need to be vectorized, vram is also very limited.

u can use spectre to generate factors, and then use other backtesting libraries to perform backtests based on spectre results, which can take advantage of gpu acceleration.