Jul 30th 2015
Know How... 154
3D XPoint, Bacteria Results, & PCIe SSD's
Crazy fast memory, identify our bacteria colonies, and PCIe-attached SSD's.
New, CRAZY-FAST Memory!
* This past Tuesday, Intel and Micron unveiled a new memory technology they created called "3D XPoint"
-- It's a non-volatile (persistent) memory part that they say is 1,000x faster that current NAND (flash) memory, more dense than DRAM, and FAR more robust than SSDs.
Let's talk about the technology:
1. Cross-point array structure
* A grid of parallel wires, perpendicular to each other, connect columns that are individual memory cells.
* You can address any single cell by using the two wires that connect the top and bottom of that column
* You can also STACK these columns on top of each other, creating a 3D grid of memory cells
2. Transistor-Less addressing
* In a NAND or DRAM device, each cell requires a transistor. That transistor is responsible for setting the cell or on or off
* In 3D X-Point, the "state" of a memory cell is not set by a transistor, but by the voltage sent down the two wires that connect that column.
** This DRAMATICALLY decreases the complexity (size) of any given cell AND increases the speed
* The speed of 3D-Xpoint comes from a few advances
-1. The lack of transistors: This decreases the latency of a state-switch
-2. The Cross-Point Array Structure: Allows each cell to be addressed more quickly,
-3. A fast-writing algorithm
* Together, these two advances allow state switches 1000x faster than a NAND cell.
* In a NAND cell, a transistor pushes a charge into an insulated cell (a capacitor). Every time it does that, the insulator is slightly weakened. Over time, that cell will lose the ability to hold a sufficient charge to be read as a state.
* In 3D XPoint, there is no insulator. The material of the column changes in a way that can be read. No capacitor = no real limit to the number of writes.
So what does this mean for us?
1. Bigger, Faster, more robust Storage!
* It'll be initially connected through PCIe, because even conventional NAND has outstripped SATA
* However, PCIe isn't even CLOSE to being able to stress the bandwidth limits of 3D XPoint"
-- a Revision 4 PCIe x16 can transfer 31.52 GBps, or almost 2GBps/per lane. (and we're JUST starting to get PCIe Revision 3... which can do almost 1GBps per lane)
2. New CPU/Chipsets/Motherboards
* We'll need a new class of computer that can accomodate this new technology
* Because it's WAY faster than a SSD, but still slower than high-speed DRAM, it might occupy an interesting position in future computer construction:
-- We could build systems that only have one type of memory (3D XPoint) that serves both as storage and system memory. (These would be systems that are built for tasks that don't require super-high-performance memory transfer speeds.
-- We could also build hybrid systems that still have ultra-high-speed DRAM for system memory, that also uses non-volatile 3D XPoint for "NearCPU storage" and SSDs for "System Connected" storage
-- We could have a new class of ultra-efficient machines that can go in and out of low power states without having to save to/load from disk.
3. The tech should hit the market in 2016
Check our Bacteria!
It's time to interpret our results
1. Check the Sterility control: If there are colonies in the SC plate, then we should throw out the experiment as the growth medium is assumed to be contaminated.
2. Compare the control and experiment plates.
* Our hypothesis was: "Exposing a cell phone to a UV light bath can significantly reduce the number of bacteria living on said phone's surface"
- If we see significantly fewer, smaller colonies and fewer types of colonies, then we've verified our hypothesis.
- If we don't see a significant change between the two sets, then we've disproven our hypothesis
- If we see change is one set, but not in the other, then we're inconclusive
** Remember that a REAL experiment should be repeatable.
* Also look for the paterns that you drew with the swab. If the colonies don't follow your patterns, there's a chance that you had airborne contamination of your growth medium.
Let's Identify our Bacteria Colonies!
1. Anytime we're handling Bacteria Colonies, we observe safety procedures
-- Gloves, glasses, face masks, lab coats, ets.
2. Look for different forms
- Go ahead and circle them
3. Turn the plate on it's side and determine the elevation of the colonies
4. Examine the surface of the bacteria. Is it smooth, dull, rough, glistening? What color is it.
5. Is the colony transparent, translucent, opaque?
6. Look at the colony under a Microscope -- Look at the edges
Thanks to Cachefly for the bandwidth for this show.
Connect with us!
- Don't forget to check out our large library of projects at http://twit.tv/kh.
- Join our Google+ Community at http://gplus.to/twitkh
- Tweet at us at @PadreSJ, @Cranky_Hippo, and @Anelf3