Accelerated functions setup¶

The easiest way to start, is by looking at the examples provided for each HUGenomic core. You can find the examples also in the accelerated functions section of this guide.

To execute one of the accelerated HUGenomic core you first need to configure the FPGA with the specific HUGenomic core you want to use.

Once the computation is over, or whenever you want to reconfigure the FPGA with a new HUGenomic core, remember to release and reconfigure the FPGA.

On this page you can find the methods in hugenomic.hpp to perform the aforementioned operations.

namespace hugenomic

Enums

enum hugenomic_core¶

The HUGenomic cores available within the HUGenomic library.

Values:

smith_waterman¶

dynamic_time_warping¶

ksw2_score¶

ksw2_cigar¶

Functions

void fpga_configure(hugenomic_core core, uint32_t fpga_id)¶

Configure an HUGenomic core onto one of the available FPGAs. The number of available FPGAs varies according to the AWS F1 instance being selected (https://aws.amazon.com/it/ec2/instance-types/f1/):

f1.2xlarge: 1 FPGA (fpga id: 0)
f1.4xlarge: 2 FPGAs (fpga id from 0 to 1)
f1.16xlarge 8 FPGAs (fpga id from 0 to 7)

Parameters

core: The HUGenomic core to configure on the FPGA
fpga_id: Identifier of the FPGA

Exceptions

invalid_fpga_identifier: if an invalid FPGA identifier is specified

void fpga_release(uint32_t fpga_id)¶

Release the resources allocated to a specific FPGA.

Parameters

fpga_id: The identifier of the FPGA (from 0 to 7)

Exceptions

invalid_fpga_identifier: If an invalid FPGA identifier is specified

uint32_t get_available_fpgas()¶

Return the number of available FPGAs on the current instance.

Note: When using the software emulator, this method always returns MAX_NUM_FPGA_PER_INSTANCE.
Return: The number of available FPGAs

Variables

const uint32_t MAX_NUM_FPGA_PER_INSTANCE = 8¶: The maximum number of FPGAs per AWS F1 instance.