seqan3::align_cfg::vectorised Class Reference

Enables the vectorised alignment computation if possible for the current configuration. More...

#include <seqan3/alignment/configuration/align_config_vectorised.hpp>

Inheritance diagram for seqan3::align_cfg::vectorised:

Public Member Functions
Constructor, destructor and assignment
constexpr	vectorised ()=default
	Defaulted.
constexpr	vectorised (vectorised const &)=default
	Defaulted.
constexpr	vectorised (vectorised &&)=default
	Defaulted.
constexpr vectorised &	operator= (vectorised const &)=default
	Defaulted.
constexpr vectorised &	operator= (vectorised &&)=default
	Defaulted.
	~vectorised ()=default
	Defaulted.

Detailed Description

Enables the vectorised alignment computation if possible for the current configuration.

In the vectorised alignment computation several pairwise sequence alignments are processed simultaneously in one invocation. To do so, we pack the alignments in so called extended SIMD registers which allow to compute a single instruction on multiple data at the same time. Depending on your processor architecture you can gain a significant speed-up, e.g. by running up to 64 alignments in parallel on the latest intel CPUs. In our mode we vectorise multiple alignments and not a single alignment. This means that you should provide many sequences to compute as one batch rather than computing them separately as there won't be performance gains.

See also

For further information on SIMD see https://en.wikipedia.org/wiki/SIMD.

Example

// SPDX-FileCopyrightText: 2006-2024 Knut Reinert & Freie Universität Berlin
// SPDX-FileCopyrightText: 2016-2024 Knut Reinert & MPI für molekulare Genetik
// SPDX-License-Identifier: CC0-1.0
 
#include <seqan3/alignment/configuration/align_config_vectorised.hpp>
 
int main()
{
    // Enable SIMD vectorised alignment computation.
    auto cfg = seqan3::align_cfg::vectorised{};
}

---

The documentation for this class was generated from the following file:

• align_config_vectorised.hpp