3D XPoint

How 3D XPoint memory works

In their 2015 announcement of the technology, Intel and Micron claimed 3D XPoint would be up to 1,000 times faster and have up to 1,000 times more endurance than NAND flash, and have 10 times the storage density of conventional memory. Early products are faster and more durable than NAND and denser than conventional memory, but they haven't lived up to the full extent of the vendors' claims.

3D XPoint has a different architecture from other flash products. It's reputed to be based on phase-change memory technology, with a transistor-less, cross-point architecture that positions selectors and memory cells at the intersection of perpendicular wires. Those cells, made of an unspecified material, can be accessed individually by a current sent through the top and bottom wires touching each cell. To improve storage density, the 3D XPoint cells can be stacked in three dimensions.

Each cell stores a single piece of data, making a cell represent either a 1 or a 0 through a bulk property change in the cell material, which modifies the cell's resistance level. The cell can occupy either a high- or low-resistance state, and changing the resistance level of the cell changes whether the cell is read as a 1 or a 0. Because the cells are persistent, they hold their values indefinitely, even when there is a power loss.

Read and write operations occur by varying the amount of voltage sent to each selector. For write operations, a specific voltage is sent through the wires around a cell and selector. This activates the selector and enables voltage through to the cell to initiate the bulk property change. For read operations, a different voltage is sent through to determine whether the cell is in a high- or low-resistance state.

3D XPoint has the ability to write data at a bit level, an advantage over NAND. All the bits in a NAND flash block must be erased before data can be written. In theory, this capability enables 3D XPoint to have higher performance and lower power consumption than NAND flash.

Major products and vendors

Intel began shipping its first 3D XPoint products in the spring of 2017. Its 375 gigabyte (GB) Optane SSD DC P4800X Series was sent to select customers in March. Broad availability is expected later in 2017.

Intel Optane memory for consumer PCs shipped later in the spring of 2017. It's a cache drive that comes in 16 GB or 32 GB capacities. Optane memory only works on PCs with seventh-generation Intel Core processors, plugging into an M.2 slot on Intel 200 series chipset motherboards.

Micron plans to have 3D XPoint-based memory and storage products available under the QuantX brand in 2017. Both 3D XPoint Optane and QuantX products use the same core die for storage being produced in the Intel-Micron joint venture facility in Lehi, Utah.

3D XPoint speed and performance

With the 3D XPoint architecture, data no longer has to be stored in 4 KB blocks using a slow, file I/O stack. The new technology enables small amounts of data to be written and read, making the read/write process faster and more efficient than NAND. Initial products using the 3D XPoint technology bear this out, though not at the speed and performance levels Intel and Micron promised when they rolled out the technology.

While not as fast as DRAM, 3D XPoint has the advantage of being nonvolatile memory. From a performance and price standpoint, 3D XPoint technology falls between fast, but costly DRAM and slower, cheaper NAND flash.

According to Intel, the P4800X drive performed five to eight times faster than the company's NAND flash-based DC P3700 in internal tests at low queue depths using a mixed workload. The P4800X can reach as much as 500,000 IOPS -- or approximately 2 GBps -- at a queue depth of 11, Intel claimed.

Observers have speculated that the PCI Express (PCIe) bus used by the P4800X is holding it back from the promised speed of 1,000 times faster than NAND. Other system changes thought to be needed for the 3D XPoint technology to meet higher performance goals include segregating persistent from nonpersistent memory when handling machine check errors and using a compiler that enables persistent memory to be declared, along with using link editors that can build that memory into an application. The applications themselves must be rewritten to eliminate file I/O and to use single instruction and vector operations.

Nonvolatile 3D XPoint dual in-line memory modules (DIMMs) that fit into DRAM slots and use the double data rate bus also may help 3D XPoint reach its full performance potential.

Cost

As of August 2017, the 375 GB Optane P4800X add-in card is priced at $1,520, or $4.05 per gigabyte. By comparison, Intel's 400 GB flash-based NVMe PCIe P3700 SSD is $879, or approximately $2.20 per gigabyte.

Intel Optane memory for PCs is $44 for a 16 GB module and $79 for a 32 GB module.

3D XPoint use cases

3D XPoint is used as an additional layer of storage between flash and DRAM. It's a relatively common practice to tier storage between hard disk drives (HDDs) and flash. High-intensity data and applications that benefit more from high speeds are stored on the flash layer, while data and applications that are accessed less frequently are put on disk. 3D XPoint is another layer of storage above flash for data and applications that need even greater speeds.

Intel expects the 3D XPoint Optane SSD will be used for high-performance storage and caching, as well as to extend and replace memory. According to the company's projections, users will be able to increase server memory by as much as eight times and displace DRAM by as much as a 10:1 ratio for select workloads.

Intel has provided three ways to extend memory with 3D XPoint Optane SSDs:

• via an operating system paging mechanism that moves data out to the PCIe-attached SSD when DRAM fills for a workload;
• via optimized applications; or
• via Intel's Memory Drive Technology supported on its Xeon processors.

In the future, it will be possible to extend memory with the 3D XPoint DIMMs that Intel plans to release. Observers speculate that 3D XPoint Optane, and particularly Optane NVDIMMs, will be used to:

• expand the apparent size of DRAM;
• enable bigger, more-effective databases;
• help overcome big data network bottlenecks;
• facilitate high-performance computing applications;
• extend memory and boost instance storage performance in the cloud;
• provide the storage capacity and speed that hybrid clouds need; and
• possibly serve as primary memory tiers in hyper-converged systems.