GPU acceleration options for the current version of Huygens.

This page gives an overview of NVIDIA cards that can be used in combination with Huygens Small, Medium, Large or Extreme GPU options for GPU acceleration. This page refers to the GPU options of the most current Huygens version. For a category overview of cards working with older versions of Huygens, please contact us.

For cards that are not included in this page: the option that is required in your license will depend on the amount of CUDA cores and amount of RAM that the card has according to the following specifications:

Small GPU option: for cards that have up to 1024 CUDA cores and up to 6 GB of video RAM (included with every Huygens license Free of Charge)
Medium GPU option: for cards that have up to 3072 CUDA cores and up to 8 GB of video RAM.
Large GPU option: for cards that have up to 8192 CUDA cores and up to 24 GB of video RAM.
Extreme GPU option: for cards that have up to 24576 CUDA cores and up to 64 GB of video RAM.
Ultimate GPU option: for cards that have up to 49152 CUDA cores and up to 256 GB of video RAM.

For the ideal combination with your CPU power check the various Performance Options in which we bundle your GPU and CPU power to let Huygens run extremely fast through your biggest data.

Huygens GPU acceleration is supported for Nvidia cards. These are supported by Windows and Linux, but not by Mac OS. If you have a valid GPU card but it's not recognized by Huygens check out GPU checklist.

Deprecation notice

Huygens 20.04

Huygens versions up to and including 20.04 support NVidia graphics cards with a Compute Capability of 3.0 or higher and a Cuda Toolkit version of 7.0 or higher.

Huygens 20.10

Compute Capability lower than 3.5 and Cuda Toolkit versions older than 8.0 are now deprecated. Huygens 20.10 does no longer support these for GPU acceleration. CPU computation and display on a monitor via these cards will continue to be supported.

Affected cards that are still supported in Huygens 20.04 but not in Huygens 20.10:
GeForce GTX 770, GeForce GTX 760, GeForce GT 740, GeForce GTX 690, GeForce GTX 680, GeForce GTX 670, GeForce GTX 660 Ti, GeForce GTX 660, GeForce GTX 650 Ti BOOST, GeForce GTX 650 Ti, GeForce GTX 650, GeForce GTX 880M, GeForce GTX 780M, GeForce GTX 770M, GeForce GTX 765M, GeForce GTX 760M, GeForce GTX 680MX, GeForce GTX 680M, GeForce GTX 675MX, GeForce GTX 670MX, GeForce GTX 660M, GeForce GT 750M, GeForce GT 650M, GeForce GT 745M, GeForce GT 645M, GeForce GT 740M, GeForce GT 730M, GeForce GT 640M, GeForce GT 640M LE, GeForce GT 735M, GeForce GT 730M;
Quadro K5000, Quadro K4200, Quadro K4000, Quadro K2000, Quadro K2000D, Quadro K600, Quadro K420, Quadro K500M, Quadro K510M, Quadro K610M, Quadro K1000M, Quadro K2000M, Quadro K1100M, Quadro K2100M, Quadro K3000M, Quadro K3100M, Quadro K4000M, Quadro K5000M, Quadro K4100M, Quadro K5100M, NVS 510, Quadro 410;
Tesla K10, GRID K340, GRID K520.

Huygens 23.10

Compute Capability lower than 5.0 and Cuda Toolkit versions older than 11.0 are deprecated as of Huygens 23.04. Huygens 23.10 will no longer support these for GPU acceleration. CPU computation and display on a monitor via these cards will continue to be supported.

Affected cards that are still supported in Huygens 23.04 but not in Huygens 23.10:
GeForce GTX Titan Z, GeForce GTX Titan Black, GeForce GTX Titan, GeForce GTX 780 Ti, GeForce GTX 780, GeForce GT 640 (GDDR5), GeForce GT 630 v2, GeForce GT 730, GeForce GT 720, GeForce GT 710, GeForce GT 740M (64-bit, DDR3), GeForce GT 920M; Quadro K6000, Quadro K5200, Tesla K40; Tesla K20x, Tesla K20, Tesla K80.

See also the wikipedia page on CUDA.

Multi-GPU acceleration with Performance options

Besides the GPU options for Huygens, SVI also offers Performance options. For the use of multi-GPU acceleration the Performance Plus, Mega or Extreme option is needed. With Huygens version 16.10.0p8, multi-GPU support has been introduced in the Batch Processor (now Workflow Processor), and Huygens Core to perform deconvolution of a queue of images on multiple GPU devices simultaneously. With Huygens 17.10 (Linux) and Huygens 18.04 (Windows), support is added for running a single image deconvolution on multiple GPUs in Huygens Professional Operation Window. We offer the following Performance Packages to get the most out of your workstation:

Performance Option: Standard included in Huygens for using up to 16 CPU cores (32 logical when hyper-threaded) and 1 Small GPU card.
Performance Plus: Allows you to use up to 32 CPU cores (64 logical cores when hyper-threaded) and 2 Large GPU cards.
Performance Mega: Allows you to use up to 64 CPU cores (128 logical cores when hyper-threaded) and 4 Large GPU cards.
Performance Extreme: Allows you to use up to 128 CPU cores (256 logical cores when hyper-threaded) and 8 Large GPU cards.

Below is a table that shows which options and commands in Huygens can use single GPU acceleration or multi-GPU acceleration in batch mode.

	Accelerated with GPU (CUDA)
Deconvolution algorithms
CMLE deconvolution	+ Multi GPU support: Operation Window, Workflow Processor & Scripting
QMLE deconvolution	+ Multi GPU support: Operation Window, Workflow Processor & Scripting
GMLE deconvolution	+ Multi GPU support: Operation Window, Workflow Processor & Scripting
Tikhonov-Miller deconvolution
Workflow Processor	+ Multi GPU support
Deconvolution Wizard
Deconvolution Express
Batch Feeder
Localization algorithms for Huygens Localizer
MLE particle fitting
LSQ particle fitting
Center of mass particle fitting
Registration & Deconvolution
Object Stabilizer	+ Multi GPU support: Workflow Processor
Drift correction	+ Multi GPU support: Workflow Processor
Stitching & Decon Wizard	+ Multi GPU support: HuCore/HRM
Light-Sheet Decon & Fusion Wizard
Image processing commands
Image processing fitlers (min, max, variance ppu, avg, gauss)
Image properites functions (hist, stat, range)
Image processing commands 1 (estbg, phaseCorr)
Image processing commands 2 (getpix, setpix, cp, slice, multiMip, miniMip, sum)
Image processing commands 3 (shift, resample, mirror, hist2ch, label, perc, optrep)
Visualization & Analysis
2D slice renderers
3D SFP renderer	Since v.19.04
3D MIP renderer	Since v.19.04
3D Surface renderer	Since v.19.04
Colocalization Analysis	+ Multi GPU support: HuCore/HRM
Object Analyzer	GPU support for labeling and rendering