Forget dual-core or quad-core chips. Octa-core? Meh. Who wants a docosa-core computer?
Intel announced on Thursday (PDF) the launch of the Xeon processor E5-2600 v4 product family designed for two-socket servers. The company says this new series provides 20 percent more cores and cache than the previous generation, and supports faster memory: DDR4 2,400 MHz. On a whole, the v4 family includes processors with up to 22 cores and up to 44 threads per socket, up to 55MB of last-level cache (LLC), and support for up to 24 DIMMs per two-socket server.
That all makes the eight-core, 16-thread Intel Core i7-5960X — the fastest consumer processor currently available — seem a bit anemic, right?
In a fact sheet provided by Intel, the company has unleashed 16 processors for two-socket servers, three for “frequently optimized” two-socket servers, two chips for low power two-socket servers, one chip for two-socket workstations, and five processors that fall under the “storage and communications” category. The company says this family will be offered with 27 different server SKUs that range in price from $213 to $4,115.
The king of the lineup is Intel’s Xeon Processor E5-2699 v4. It has 22 cores, a frequency of 2.2 GHz, 55MB of LLC cache, and a power envelope of 145-watts. It’s crazy expensive, costing the previously mentioned $4,115. The Intel Xeon Processor E5-2698 v4 chip is the next step down price-wise, costing $3,226. With that chip you get 20 cores, a frequency of 2.2 GHz, 50MB of LLC cache, and a thermal envelope of 135-watts.
The new processor family includes improved virtualization performance, improved Turbo Boost Technology, Intel TSX instructions, Intel AVX2 instructions, and support for Intel’s new Scalable System Framework. There’s also cache monitoring and allocation technologies, memory bandwidth monitoring, and the Intel Node Manager, monitors and controls “server power, thermals, and utilization.”
Intel says the new family enhances security by including an integrated random number generator for creating security keys. There’s also a random bit generator offered in the new processor family for “seeding’ software. Futher, new instructions accelerate secure session initiation protocols that are based on RSA, ECC, and SHA, or Secure Hash Algorithm. Intel says it has also improved its Data Protection Technology with the new family.
The new server chips are based on the “Broadwell” architecture, meaning they’re built on the 14nm process, as reported by Ars Technica. The complete price list is located in a PDF format here if you’re curious, and additional details regarding this new processor family can be found here. The announcement was actually part of Intel’s Cloud Day 2016 event held on Thursday, which you can learn more about in its dedicated Cloud Day 2016 web portal here.
- Intel Alder Lake destroys AMD in new gaming performance leaks
- Alder Lake CPUs: Everything we know about Intel’s upcoming 12th-gen hybrid chips
- Intel Alder Lake may be up to 45% faster than the previous generation
- AMD’s powerful Threadripper 5000 chips may have been delayed again
- AMD vs. Intel