Intel’s introduction positions its new and advanced Optane 3D Xpoint memory technology as the solution to today’s computer slow-down issues.
By Jon Gabay, contributing writer
It seems that
with its Optane technology Intel is trying to introduce a new
paradigm in the way in which we understand, see, and use computers. It’s not
uncommon for vendors to exaggerate the importance of claims and marketing
messages to stir up controversy, or, to make people believe that sliced bread
is really something worth getting excited about. That may be what’s happening
basic intent of the Optane technology and architecture is to create a smart
cache that lives between a CPU and a hard disk. By learning and anticipating
the processors needs, readily available data can be presented without latency. Intel claims 28% increase in overall
performance with a 14% increase with hard drive access performance. But numbers
can say anything.
introduction positions its new and advanced Optane 3D Xpoint memory technology as the solution
to today’s computer slow-down issues. The bottleneck they cite is memory. The
image they convey is that the common computer slowdowns we experience are due
to inefficiently designed computers with old and outdated memory architectures.
promote the technology, Intel has populated its website with video overviews. Basically, though,
Intel’s videos are dumbed down for easy consumption by the masses; they’re not
particularly useful for the engineer. Worded carefully, the architectural
enhancement Intel claims as the solution to this slowdown are for the times
when a CPU and hard disk need to access and work with large data sets. They
also claim modern memory is just too slow and inefficient in this application.
is that so? Modern memory is fast and dense, especially compared to
state-of-the-art memory from just a little while ago. Synchronous DRAM, for
example, can perform double and quad rate transfers and densities are
continuing to increase, albeit it not as quickly as in the past. Static cache
RAM, too, is plenty fast to keep a processor well fed — even those running at GHz rates using
multi-cores. In the end, it’s more a throughput issue, not a speed issue.
missing ingredient appears to be non-volatility at higher speeds, which can
bring something to the party if it’s fast enough that it doesn’t need to be
cached using higher speed SRAM or even DRAM. It’s true that denser memory pools
and non-volatility allow more complex operations and let programs do things
that were previously not possible. Take GPS and mapping for example. Without
the modern densities on non-volatile memory technologies, these common and
everyday accepted devices would not be feasible, especially with the feature
packed versions we’ve gotten accustomed to.
While row and column matrix
architecture and 3-D semiconductor fabrication techniques are not new, the fast
and dense no- volatile transistor less memory cells may propel Optane
technology as a mainstay between hard disks and CPUs going forward. Image source: Intel.
brings up another point about Intel’s Optane announcements. The new memory
technology Intel references is perhaps the most interesting part of the
announcement. The “revolutionary architecture approach” about using a row and
column topology is perhaps targeted to high school students who are unaware
that every memory device, keyboard layout, display technology, and so on rely
on a matrix such as this. Simply put, there’s nothing new here.
3D semiconductor architecture is not a new or revolutionary concept either.
Although requiring more masks and a costlier fabrication process, 3D
fabrication can and has been taken to many layers. The 3D stackup of the Optane
matrix does carry with it some improvements in density, though, since row or
column wires can now be re-used in a 3 dimensional array instead of just two.
fast and dense non-volatile transistor-less memory cells that do not need to be
refreshed is perhaps the announcement most worthy of mentioning. Still, ferroelectric
memory had this benefit years ago, but never really flourished in any
high-speed and high-density way.
times, however, it is the combination of existing as well as new technologies
that makes something news-worthy. That may be the case here. By combining the
modified memory architecture with the 3D and transistor-less, non-volatile
memory devices, Optane does seem to provide a performance boost in the use of
reality, though, for many users the most common causes of computer slowdown are
communications latencies end to end. The network is the computer, and with a
world wide web, data from multiple sources needs to feed our processors, and often
times we have to wait. Another major source of slowdown is the malware,
ad-ware, and the dozens of agencies, corporations, and individuals that have
their fingers on our cookies to monitor, scan, or halt our computer’s activities.
Also, many modern programmers write sloppy software which do not use or access
memory efficiently compared to coding techniques of years ago. This may reduce
resources and performance, even at an accumulating level.
“Intel used to be
a leader in memory technology early on,” said communications expert and former
editor in chief at Integrated Circuit
Magazine, Alex Mendelson. “The introduction
of Optane represents Intel’s steps to become a value added memory supplier.”
But does it merit the kind of excitement Intel is promoting?
Perhaps a higher
Optane memory fueling our next- generation devices and infrastructure backbones
will help streamline performance for what’s to come. More emerging memory
intensive applications like 3D holographic imaging and higher speed
communications may benefit, although they might also simply flesh out other
bottlenecks. Optane will thus at least benefit some applications, but likely
not all. In any case, though, to Intel: thanks for the memory.
Since studying electrical engineering, Jon Gabay has worked with defense, commercial, industrial, consumer, energy, and medical companies as a design engineer, firmware coder, system designer, research scientist, and product developer. As an alternative energy researcher and inventor, he has been involved with automation technology since he founded and ran Dedicated Devices Corp up until 2004. Since then, he has been doing research and development, writing articles, and developing technologies for next-generation engineers and students.