pub fn stacked_matrix(
l_skip: usize,
n_stack: usize,
traces: &[&DeviceMatrix<F>],
device_ctx: &GpuDeviceCtx,
) -> Result<(PleMatrix<F>, StackedLayout), ProverError>Expand description
The traces must already be in height-sorted order.
This function is generic in F and only relies on CUDA memory operations.