Research and development of an innovative tenzor-engine for the automatic optimisation of quantum mechanics, quantum chemistry and quantum information technology used in industry for graphics cards (GPU) supported by translation technology

Operation co-funded by EU cohesion policy (European Regional Development Fund (ERDF)) under the programme Competitive Central-Hungary - ERDF/ESF (2014HU16M2OP002), managing authority: Nemzetgazdasági Minisztérium. Beneficiary: StreamNovation Szoftverfejlesztő Korlátolt Felelősségű Társaság. Location: Budapest (NUTS HU110); Hungary. Budget: total eligible expenditure EUR 234,364.72; EU co-financing EUR 105,932.85 (co-financing rate 45.2%), as stated by the source. Runs 2018-06-01 to 2020-05-31. Source: Kohesio, EU cohesion policy projects managed by national and regional authorities (not CORDIS). Summary from Kohesio (English machine translation by the source): A) During the project, a translation program will be developed that provides the opportunity to optimise the generation of most simulation algorithms used in quantum mechanics, quantum chemistry and quantum computing for nearly automatic graphics cards (GPUs). The software to be implemented is capable of generating quantum multiparticle integrals and based on these tensor conversion and optimisation methods optimised for GPU. This would allow nearly automatically optimised production of software used by a very wide area of industry. As a result of automation, the development time of any new method could be reduced by order of magnitude (thus increasing the cost-effectiveness of software development companies) and by optimising for GPU, running time could be reduced to a minimum of ten or even a century, which increases cost-effectiveness in the simulation phase by ten or even a hundred times the same proportion in the application industries (pharmaceuticals, plastics, nano-electronics, etc.). Quantum simulation software is stratified, where the next layer is based on the calculations of one layer. We already have significant practice in porting the time determinant step of quantum chemistry software (calculation of two-particle integrals) to GPU and optimising algorithms. A program ported to such a GPU will serve as the lowest layer of software to be developed in this application. The effective solution of the task is based on the serious international success of our previous software development experiences and achievements. This is shown by a patent submitted (P1500504) and a protected patent (PCT/HU2013/000051), international journal publications (Rák, Ádám and György Cserey. “The BRUSH algorithm for two-electron integrals on GPU.” Chemical Physics Letters 622 (2015): 92-98) and an important cooperation agreement with a market leader in material science software (qchem Inc.. USA). The software to be developed in the tender, using the Know-how and patents [shortened; the full text is on the Kohesio page]

Original source.

Source: Kohesio (EU cohesion policy projects), European Commission, Directorate-General for Regional and Urban Policy (DG REGIO). Retrieved 2026-10-21. Licence: CC0 1.0 (Cohesion Open Data Platform, dataset 2014-2020 Kohesio projects). Selected metadata, adapted presentation; the publisher's record is authoritative. No endorsement by the source is implied. Licence: public domain dedication (CC0 1.0).

AI-assisted summary: The project focuses on developing a tensor-engine for optimizing quantum mechanics, chemistry, and information technology for GPUs, co-funded by the EU's ERDF. It runs from June 2018 to May 2020, with a total budget of EUR 234,364.72. Suggested technology pillar: simulation. Generated by AI from the public source text; it may contain errors and the original source is authoritative.