This site may earn chapter commissions from the links on this folio. Terms of employ.

E'er since Intel announced its 3D Xpoint memory (branded equally Optane), the company has claimed that the new class of memory would represent a fundamental jump frontward for the entire industry. Proof of these claims has been relatively boring to appear — it'southward difficult to replace existing retention technologies, and the existing memory stack (spinning difficult drives, NAND, and DRAM) covers a wide range of toll points, power consumptions, reliability, and capacity.

Intel is launching a new class of Optane it calls Optane DC Persistent Retention, with capabilities that it believes volition bridge the gap betwixt DRAM and non-volatile storage while expanding the amount of retention available per CPU socket to equally much as 3TB.

filling-the-gaps-between-memory-and-storage-after

Optane DC Persistent Memory is pin-compatible with DDR4, according to Anandtech, and will exist offered in packages of up to 512GB per stick (vi RAM slots = 3TB of addressable RAM per socket). Systems will be capable of fielding large Optane caches alongside smaller DRAM pools; 1 of Intel'south demos showcased a Cassandra database running on 256GB of DDR4 RAM + 1TB Optane DC PM, as opposed to 1TB of DRAM. Intel'due south major focus with Optane DC PM is on storage performance consistency. Using DRAM + NVMe-connected storage tin be limiting in certain scenarios, with performance bottlenecked past storage write-backs. An Optane cache avoids this trouble.

Intel is also working on a adequacy it calls Persistent Retentiveness over Fabric (PMoF), a low-latency data replication method with direct load/store access that can maintain operation of 287,000 ops/second, compared with initial performance of three,164 ops/2nd for a conventional DRAM + storage arrangement. (Eventual consistency for the DRAM + storage system, Intel notes, is comparable to the Optane rig). But Optane'southward depression latencies are also useful in other contexts, like when restarting a database. Intel reports a restart time of 2,100 seconds for a conventional configuration compared with 17 seconds for Optane.

Much of the piece of work to accomplish this has been done on the software side. Optimizations and file arrangement abstractions are necessary for putting the "Persistent Memory" in the Optane DC drives that Intel is launching. To help enterprises tune their databases and software for Optane DC PM, Intel has built a new Persistent Memory Development Kit (PMDK) with a collection of libraries, APIs, and other software tools. PM will be supported on both Windows and Linux and Intel has added support for the capability in its performance analysis software kit, Vtune.

Details on the SDK and how it interacts with software are a little scarce at the moment. Only the implication hither is that the heavy lifting to make Persistent Memory piece of work is handled exterior of the applications themselves. In other words, if you're running Application 10, X doesn't necessarily have to exist optimized by the original developer to take advantage of Persistent Memory; that's the job of the PMDK to handle. The implication hither is that app developer optimizations are useful and productive, but aren't a literal requirement — and that's an of import feature when talking nigh a adequacy with such potential to reshape retentivity hierarchies.

Other tidbits from the Q&A: These new capabilities will be tied to new, upcoming CPUs from Intel, with acquirement shipments to select customers in 2022 and broad availability in 2022. Intel as well expects to send QLC NAND drives, like Micron, in the back half of the year and into 2022. Intel isn't disclosing figures on DIMM Optane endurance, power consumption, or pricing at the moment, but DDR4-standard clock speeds are expected for Optane DIMMs.

Overall, this shift towards Optane, and an emphasis on cloth performance, are a critical component of Intel's ain transformation from a personal CPU-focused visitor to a house with stronger ties to data center and deject processing. Intel clearly sees new storage engineering science equally central to its long-term addressable marketplace spaces, and it's targeting its initiatives at these areas, possibly in a bid to tie companies more than closely to its CPU products. After all, if Optane provides pregnant functioning advantages and Works Best With Intel (or merely with Intel), then Intel has a neat method for keeping itself even more than relevant in the data center market place.

Then again, this may as well reflect the demand to optimize other facets of computing across pure CPU performance, perchance every bit role of pursuing exascale calculating, where DRAM ability consumption is a major limiting factor. With Qualcomm looking to exit the server business organisation, Intel isn't facing major rut from ARM in information centers, and AMD's Epyc ramp, while a competitive threat long-term, isn't expected to shatter Intel's server dominance (AMD hopes to take 4-6 percent of the server market this yr).