Intel Optane - now with a capacity of one and a half terabytes


    Intel Optane SSDs with 3D Xpoint memory are becoming more and more capacious. This is a natural evolutionary process, and here it is fixed a new milestone. Optane has once again almost doubled and now holds up to 1.5 TB of data. Under the cut, we will discuss the current state of affairs in the top part of the family, as well as some news relating to other segments.

    The new wave of increased volume drives includes consumer class models 905Р and professional Р4800Х. Both there and there drives are available in PCIe and U.2 form factors. In view of the novelties, the picture is as follows. Bold highlighted new products, in italics - submitted, but not yet released.
    900P 905P P4800X P4801X
    100 GBU.2
    280 GBPCIe, U.2
    380 GBM.2 PCIe, U.2M.2
    480 GB PCIe U.2
    750 GB PCIe, U.2
    960 GBPCIe, U.2
    1.5 TBPCIe, U.2PCIe, U.2
    U.2 half-terabyte models are the first drives that use BGA packages for packaging 3D XPoint chips - 8 chips per pack, 128 GB in total. The same construct can be used when creating M.2 of version 905P with a capacity of 750 GB; The problem is the sufficiently large power consumption of the 3D XPoint - even with a volume of half the total consumption is almost close to the limit (about 15 W), which M.2 can provide at maximum load per pin. So the 750 GB M.2 line will be technologically difficult: you will have to resort to throttling or other saving functions.

    In the column Р4801Х a model with a capacity of 100 GB attracts attention. It may seem that this is an analogue of the 800P entry-level model, but appearance is deceptive: the P4801X uses a full-fledged 7-channel controller with a PCIe x4 interface, so the performance is incomparable. On the other hand, this sample, of course, is inferior to more capacious models due to insufficient parallelism, except that the delays will be the same, and the consumption will be quite acceptable for M.2 8W.

    So, the Intel Optane lineup is expanding, becoming more versatile, suitable for solving various kinds of tasks. However, the main problem remains the same: Optane is still expensive, both in absolute values ​​and in terms of gigabytes (this figure still does not reach $ 1 / Gb). Breakthrough here is not observed.

    Also popular now: