Yes, improved threading is needed...
...But why do the performance enhancements not scale down? Where are the four, eight, and sixteen core systems? I smell a Cash Cow Patty in the making. Hold your nose lest you get a whiff.
Intel now says its "Sierra Forest" Xeons will actually offer 288 cores, twice as many as previously disclosed, when it launches in the first half of 2024. When Intel first revealed the efficiency-geared Xeon earlier this year, the company said the chip would be the first to use its Intel 3 (3nm) process node and feature a …
Microsoft and Oracle are probably already rubbing their hands with glee at the news of the cow patty's core count since their softwares are licensed by the core...
if one would try to run MS Windows on it, a single 288-core CPU will need a whopping total of 18 full 16-core licenses for Windows Server.
18 Windows licenses for a single CPU. ?!!!! That will be a pretty hefty cash cow.
From what I understand, the efficiency core is not a performance enhancement but a setback that scales up to support more VMs while still meeting data center thermal and power requirements. As likely as not, eight to 32-core CPUs are chasing single thread performance and anyway don't have much trouble meeting the power limits of even the oldest racks.
The performance gap between Intel’s P-Cores and E-Cores isn’t as big as the difference in die-size and power requirements. In Alder Lake, the E-Cores provided about 50~80% of the performance of the P-Cores, depending on workload. On the die, the P-Cores are about 4x the size of the E.
E-Cores are perfectly capable of running any HTTP-borne service at a couple of thousand connections per second, so from a data-centre’s point of view, it’s far more desirable to have 250 of those in a rack than having P-Cores and dividing them using software... after all, there's not much use for the P-Cores’ fancy matrix maths or hyperthreading when all your applications are sitting inside single-threaded runtimes like Node.
In a datacentre, it makes perfect sense to specialise hardware like this. On a desktop, where the one system has to handle all workloads, then mixed-core designs are the way to go.