Data center GPU usage goes beyond hardware. Admins must work with vendors and developers to have the right software architecture and code to benefit from GPUs
"Graphics processing unit hardware is straightforward to install on most general-purpose servers, but its benefits are not automatic; they require the combined support of a software framework, drivers, hypervisor extensions and a GPU-enabled application to unlock any benefits.
A graphics processing unit chip can contain one or more separate GPU cores -- similar to CPU chips. However, the GPU chip itself is almost never enough to fulfill GPU roles, so the latest GPU chip sockets are rarely directly integrated onto server motherboards..."
Hyper-convergence and up-and-coming HCI 2.0 and composable infrastructure can help you achieve your IT and business goals. See why, how and where their strengths and weaknesses lie
"Demands placed on infrastructure fluctuate due to changing business requirements and ever-evolving technology challenges. Two of these challenges include digital transformation, an overarching term covering a range of issues and technologies, and a more technology-specific scenario known as edge computing..."
Review notable hyper-converged, HCI 2.0 and composable disaggregated product news to gain insight into data center convergence technology as you make infrastructure decisions
"The hyper-converged data center has been around for years now, going back to at least when pioneer Nutanix helped jump-start the market with its first appliances in 2013. A lot has changed since then. Continued product innovation has advanced hyper-convergence.
Today, hyper-converged technology has gone mainstream, and big and small organizations have adopted hyper-converged infrastructure and its descendants to meet the storage and processing requirements of their data and mission-critical workloads..."
See all Archived IT - Server articles
See all articles from this issue