Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
DRAMless SSDs Hit A New Low
#1
http://techreport.com/review/32665/toshi...reviewed/7
Quote:Well, that's that. The TR200 is the pokiest SSD (aside from our baseline X25) we've yet tested on this rig. That distinction used to be held by the Trion 100, but the new guy slid under it by just a few points. Perhaps the one positive for the TR200 is that it's about as fast as the Trion 100 without even having a DRAM cache on board.

It's worth noting that our geometric mean conceals the drive's pokey write performance and fine read performance, and if your workload primarily involves reading data, the TR200 may prove tolerable. That's still not a ringing endorsement, but it may be the way the cookie crumbles for the budget systems in which this SSD might find a home.

In fairness, Toshiba made clear from the outset who the target audience is for this drive. The TR200 will feel mind-bogglingly quick to anyone who is only now upgrading from a traditional hard drive. The chief problem that I have is that Toshiba's suggested prices, while low, are well within striking distance of faster drives that frequently go on sale. But let's back that up with data, as is the Tech Report way. In the plots below, the most compelling position is toward the upper left corner, where the price per gigabyte is low and performance is high. Use the buttons to switch between views of all drives, only SATA drives, or only PCIe drives.

http://www.tomshardware.com/reviews/tosh...241-5.html
Quote:We'll let the performance charts do most of the talking about the TR200's performance. We can’t ignore the DRAMless market as a whole, but these products promised to deliver in two key areas; lower prices and reduced power consumption. It's obvious they don't deliver on those promises.

Removing the expense of one or two DRAM packages provides minimal cost savings. DRAM prices are up, just like NAND, but companies like Toshiba buy the components on a large scale, so they shouldn't bat an eye at the expense. The results do not justify the means, even with a $10 decrease in component expense. The only price that users care about is what they pay, and DRAMless SSDs are not cheaper than products already on the market.

Surprisingly, DRAMless SSDs do not reduce power consumption. A single picture doesn't tell a complicated story. You could argue that less power consumption in a snapshot is better, like during a test while the SSD is at idle or under heavy load, but SSDs are complex and dynamic instruments. Our battery life tests are a better measure of real-world power consumption, and the TR200 trailed the competing products.

For the most part, DRAMless SSDs have failed to deliver a single benefit to end users. Dollar for dollar we struggle to find a viable reason to even consider this product class. We have to question why companies would bring these SSDs to market given the current competition and pricing, especially the existing models that have dominated the landscape for years.

The next generation of DRAMless products promises to address many of these issues. DRAMless NVMe SSDs technically do not have DRAM, but the protocol supports using your system memory as a buffer for the physical-to-logical address map. It's a great theory, and it might even work in the mass market. Companies keep displaying the promising technology at trade shows, but they aren’t releasing samples or products. That leads us to believe it may have issues under the surface that we don’t see on the trade show floor.
Valve hater, Nintendo hater, Microsoft defender, AMD hater, Google Fiber hater, 4K lover, net neutrality lover.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)