Alafia Ai, a US-based provider of high-performance precision medicine platforms, has rolled out AIVAS, an all-in-one interactive high-performance personal supercomputer.
AIVAS is an interactive supercomputer designed to advance the development and deployment of critical artificial intelligence (AI) software applications in healthcare.
It advances in-silico experiments and bioinformatics pipeline development by multiple orders of magnitude as compared to existing systems.
Alafia said that its supercomputer will provide bioinformaticians and healthcare professionals with advanced tools to get insights across various medical applications.
AIVAS helps surgeons and medical physicists conduct real-time reconstructions using standard DICOM images and videos, including CT and MRI scans.
It leverages the parallel processing power of both CPU and GPU with a unified memory architecture, to handle large datasets and boosts medical reconstruction performance.
Stanford University, Department of Neurology, research staff and Acacia Clinics vice-president of research Danielle DeSouza said: “Performing cortical reconstruction with AIVAS was significantly faster than traditional computers.
“We reduced individual subject reconstructions from over 24 hours to under 2 hours per subject and could parallelize to perform hundreds of subjects at a time. Clinical trial cohort pre-processing that used to take weeks to months was completed in under a week.”
According to the company, the rising complexity of applications in digital pathology, neurology, genetics research, and cardiac imaging demands more computational power and efficiency.
Its AIVAS supercomputer will address the market demand by providing the highest compute density and energy efficiency for low latency, mission-critical workflows, and secured workloads.
AIVAS enables next-generation high-throughput sequencing algorithms, computational mass spectrometry, and bioinformatics automation in clinical environments.
It also facilitates real-time data visualisation and processing of CT and MRI scans for use cases ranging from cortical reconstructions to image reconstruction of PET scans.
The key features of AIVAS include high performance, future-proof architecture, precision and efficiency, extensive software support, privacy, and security and compliance.
AIVAS leverages a cloud-native CPU and four NVIDIA GPUs, equivalent compute capacity of two full cloud-native AI servers, to process 1,000 trillion operations per second (TOPS).
It features 128 CPU Cores with 256 SIMD vector engines, 28,416 NVIDIA CUDA GPU Cores with 92GB of GPU unified high-bandwidth memory and 2TB of unified system memory.
