> "Besides this, the overall load on the cluster fell below 50 percent thanks to Hyper-Threading technology, further contributing to the impressive results. In

> "Besides this, the overall load on the cluster fell below 50 percent thanks to Hyper-Threading technology, further contributing to the impressive results. In broad terms, when the load increases to over 50 percent, HT-engines, which aren’t as useful as physical engines, start working." How did "overall load on the cluster fell below 50 percent thanks to Hyper-Threading technology"? Isn't HT simply exposing a virtual core for every physical core on the CPU? When an HT CPU is monitored to be at 50%, it's actually at 100% or close to it, so it appears that what fell 50% was the reported values for CPU usage. I don't understand how there's a 50% decrease in load due to "when the load increases to over 50 percent, HT-engines, which aren’t as useful as physical engines, start working"