Are high-end desktop platforms still relevant?
Revolutionary tech is coming. That’s revolutionary with an R.
“Probably sick of being off the front page, Intel has been taking steps in recent months to tell the world that it isn’t going anywhere”.
2019 has been the year of AMD. Desktop Ryzen and enterprise Ryzen Pro, along with Navi GPUs have been the big tech highlights of the year. Probably sick of being off the front page, Intel has been taking steps in recent months to tell the world (that is, the tech media) that it isn’t going anywhere and to keep an eye on it in 2020 and beyond. We attended a recent tech day in South Korea where we were given a look at some of the things Intel is working on with regards to its storage products.
With 10nm products yet to really see the light of day, there’s not a huge amount to talk about with regards to new CPU architectures and process technology. We want to see something new, something cool, something awesome, and maybe we have it with the revelation of new Optane technologies.
I’ve personally been a fan of Optane for some time. The problem has been the messaging behind it. Optane SSDs offer blazing fast response times and I/O and they can have a measurable effect when used as a cache drive with a HDD. The problem has been is that Intel doesn’t include Optane support on its cheapest chipsets. To really succeed as a cache drive, even the lowly H310 chipset needs Optane support. These users are the ones most likely to be using a HDD.
What’s really exciting is Optane persistent memory, which is memory that retains data when it is powered down, unlike traditional DRAM. This kind of tech has some really amazing promise, though for now it is limited to the data centre. Think of something like rebooting a database. There’s a reason your internet banking goes down for several hours most weekends. It’s not necessarily to fix something that’s broken; a lot of it is the time it takes to reboot something that large from storage!
It essentially combines the best properties of Optane storage, with closer-to-the-CPU DRAM. It elevates storage one step up the data pyramid. On-die caches will always be the fastest, followed by DRAM and then storage with mechanical drives at the bottom. The end result is faster computing, with better response times, faster application loading and boot times. Sounds great!
The problem is, for now, this technology is very expensive. We might see it trickle down to workstation class products in the following years, but for now, Intel remains tight lipped about whether we will see persistent memory make it to the desktop.
Intel is also gearing up to launch new SSDs including the 665p, a new 96 layer QLC NAND drive that should theoretically bring down the dollars per gigabyte ratio even further, all but eliminating hard drives from notebook class devices.
We were also teased with second generation Optane products, though there is some time to go until they launch. Codenamed Alder Pass, these products could end up making existing Optane drives look positively sluggish. Will we see an enthusiast class 512GB or 1TB 2nd generation Optane drive? Though it would never be called cheap, and NAND drives would have nothing to worry about, such a device would be fast. Very fast. Yes I would take two.