PIMPMe: Processing-In-Memory Primitives for Data Management

Funding: DFG

Processing-In-Memory Primitives for Data Management (PIMPMe) is a joint project of the Computer Architecture and Embedded Systems Group and the Databases and Information Systems Group at Technische Universität Ilmenau. The project is part of the DFG priority programme on Disruptive Memory Technologies (SPP 2377).

Ongoing developments in the field of memory technology open up great potential for building highly efficient data-intensive systems addressing the challenges of modern applications. At the same time, it requires to rethink traditional assumptions to fully exploit this technology. In this project we plan to investigate the benefit of memory-centric computing paradigms: we will conduct research on offloading computation to memory in order to reduce the amount of data to be transferred between memory and CPU and in this way increase bandwidth, to reduce the CPU load for computation, and in the end, help to reduce the energy consumption of modern IT systems. We plan to particularly extend the results of the first project phase on offloading database computation to memory. Specifically, we plan the following two dimensions of extension: (1) PIM abstractions: While in the first phase we considered primitives tailored for the UPMEM memory technology and are still working on initial FPGA support, in the second phase we plan to design and develop higher-level abstractions that support different storage and execution models. (2) Distributed computing: With the advent of interconnects such as CXL.mem, but also feasible with networking technologies such as RDMA, shared memory infrastructures are becoming available where multiple compute nodes can access the same memory and offload computation. We plan to explore the challenges and opportunities of such infrastructures in the specific context of distributed database processing.

More infos...

PollenNet: Phänologie-basierte Pollenvorhersagen und EEG-basierte Bewertung allergischer Reaktionen mittels KI

Funding: Carl-Zeiss-Stiftung

The PollenNet project involves studying the plants that produce allergenic pollen and the properties of the pollen itself, relying on these knowledge to performe predictions or forecasting of the plants that produce these allergic pollen and when they are often released, hence helping individuals who are allergic to this pollen to be notified in advance, and promoting the need for being aware of the situation of pollen rise and fall in their environment. 
Our major goal for this project is to harness the pollen datasets into a single spatial database: PostGIS. PostGIS consists of grid cells, and in each cell, we can store related information of a particular environment (longitude, latitude) about pollen counts and other weather conditions. One of the major challenges is creating a homogenous pipeline between weather components and pollen counts, in addition to retrieving these vector datasets (pollen counts together with weather conditions) as fast as possible. In pollen transport, weather conditions play a huge role. For instance, wind components promote the transportation of these pollen in different directions, while the temperature helps in pollen release. An extensive data analysis of these components is being studied, as it will help to understand the properties of the pollen and their transportation patterns that invariably characterise the pollen plants. 

MemWerk - Memristive Materials for Neuromorphic Electronics

Funding: Carl-Zeiss-Stiftung

The MemWerk project is an interdisciplinary project in partnership with other research groups of TU Ilmenau. It aims at comprehensively investigating memristive materials for neuromorphic electronics, i.e., electronics inspired by biology, which are highly energy-efficient. Our goal is the management and analysis of complex material science data from the MemWerk project as graph data. We first develop a material atlas for collecting material and process parameters. Further, we model and manage the data as graph data and carry out graph analytics to gain insights into the material and process parameters concerning the characteristics and performance parameters of memristive devices and neuromorphic circuits. Thus, we can explore and extract information with regards to the material or process parameter(s) that will likely influence certain neuromorphic properties, the interplay that can be inferred between process components and measurements, the extent to which components are densely/loosely connected in a process, the parameters that are most central or important in certain processes, the parameters that play similar roles in the processes, the outlier measurements of parameters and/or neuromorphic properties etc. Ultimately, this would enable the realization of materials that are well-suited for neuromorphic systems.

More infos...