Following the successful launch of AMD Ryzen amid, some hitches, AMD's Zen architecture just recently released AMD Naples, the next generation data center, and enterprise processor.
The two-socket server chip is directly pitted against Intel's Broadwell-EP-based Xeon E5 V4 range of server chips. Each Naples chip features 32 cores that can handle 64 simultaneous threads, with eight memory channels, capable of supporting up to 2TB RAM and 128 PCIe 3.0 lanes. Intel's offering pales in comparison to having 22-cores and 44 threads, four memory channels, and support up to 1.5TB of RAM, according to ArsTechnica.
Employing the CPU in a popular 2p server platform, this would result in having 64 cores with 128 threads of computing resources. Specs pitted against Intel's top-end 22-core Xeon E5-2669 v4, it would be clear, that AMD Naples has an advantage.
Additionally, AMD Naples can handle up to 16 DIMMs of memory that can be allocated per CPU, which would give it up 4 TB of DDR4 RAM. Intel has a quad-channel memory setup with an 1866MHz DDR4 interface speed, according to Forbes.
With its impressive spec, AMD's best quality is its strength in serial IO connectivity, would do well its architecture. AMD Naples is kept in constant communication links between CPUs within the company's Infinity Fabric. Naples has PCIe lanes that can connect directly with the CPU root complexities such as GPUs and other co-processors.
With all numbers stacked up against Intel, it seem it is capable of running second only to AMD Naples, with its architecture potentially more powerful in terms of raw CPU resources. Its CPU cores are primed for support with memory bandwidth and additionally given direct access to PCI Express.
During tech analyst day, AMD has shown different scenarios, the dual-socket Naples server was listed 2.5 faster than Intel's Xeon server, tested with specific data analytics workload. Having Naples direct attach PCIe connectivity adds another plus for its platform.
Driving an even wider wedge is AMD Naples is partnered with the recently announced AMD Radeon Instinct GPUs employed for AI machine learning applications. It would be interesting to follow developments of this slugfest for server dominance between AMD and Intel in the data center.
