Some time ago Objective Analysis ran nearly 300 standard benchmarks on a PC with varying amounts of flash and DRAM and found that a dollar’s worth of flash provided a greater performance boost than a dollar’s worth of DRAM once the DRAM size grew above a certain minimum (1-2GB) depending on the benchmark.
You might wonder how this could possibly be true. Everyone knows that best way to improve any computing system’s performance is to add DRAM main memory. How could flash, which is orders of magnitude slower than DRAM, provide a bigger performance boost than DRAM?
It all makes sense if you think of the DRAM of something that is there only to make the HDD look faster. More is better, but if you can use a little less DRAM and add a large flash memory layer then disk accesses appear to speed up even more.
The benchmark data and the price/performance findings that are Continue reading
A couple of specifications for SSD endurance are in common use today: Terabytes Written (TBW) and Drive Writes Per Day (DWPD). Both are different ways to express the same thing. It seems that one vendor will specify endurance using TBW, while another will specify DWPD. How do you compare the two?
First, some definitions. “Terabytes Written” is the total amount of data that can be written into an SSD before it is likely to fail. “Drive Writes Per Day” tells how many times you can overwrite the entire capacity of the SSD every single day of its usable life without failure during the warranty period. Since both of these are guaranteed specifications, then your drive is most likely to last a lot longer than the number given by the SSD’s maker.
To convert between the two you must know the disk’s capacity and the warranty period. If drive maker gives you TBW but you want to know DWPD you would approach it Continue reading
From time to time IT managers ask The SSD Guy if there’s an easy way to compare SSDs made with MLC flash against those made using eMLC flash. Most folks understand that eMLC flash is a less costly alternative to SLC flash, both of which provide longer wear than standard MLC flash, but not everyone realizes that eMLC’s superior endurance comes at the cost of slower write speed. By writing to the flash more gently the technology can be made to last considerably longer.
So how do you compare the two? OCZ introduced MLC and eMLC versions of the same SSD this week, and this provides a beautiful opportunity to explore the difference.
As you would expect, the read parameters are all identical. This stands to reason, since Continue reading
This replacement for the company’s Z-Drive 4000 series is a complete redesign with an obsession for performance. OCZ tells me that they moved from a 2-hop design to a 1-hop by using the PMC Princeton PCIe SSD controller, and have passed the University of New Hampshire Interoperability Labs’ compliance tests to NVMe 1.1B compliance.
But how does it perform? Well the 1-hop design helps reduce latency (which is just starting to overshadow IOPS in users’ minds) and the latency of this SSD is significantly lower than competing NVMe SSDs: between 25-30μs, figures that OCZ tells me are very consistent, a big plus for enterprise applications. As for IOPS, the device can perform under a 70/30 Read/Write load at 330K.
The 6000 series is provided in both standard MLC and eMLC for those who want the security of eMLC and are willing to sacrifice a little performance to sleep better at night.
This product is a good fit for the market needs, and shows how devoted OCZ and its parent Toshiba are to providing high performance in the SSD marketplace.
I have to admit that it’s embarrassing when The SSD Guy misses something important in the world of flash storage, but I only recently learned of a paper that Baidu, China’s leading search engine, presented at the ASPLOS conference a year ago. The paper details how Baidu changed the way they use flash to gain significant benefits over their original SSD-based systems.
After having deployed 300,000 standard SSDs over the preceding seven years, Baidu engineers looked for ways to achieve higher performance and more efficient use of the flash they were buying. Their approach was to strip the SSD of all functions that could be better performed by the host server, and to reconfigure the application software and operating system to make the best of flash’s idiosyncrasies.
You can only do this if you have control of both the system hardware and software.
The result was SDF, or “Software-Defined Flash”, a card that Continue reading
So why would this appear in the pages of The SSD Guy blog?
In a nutshell, coffee demand is falling thanks to increased use of SSDs. It’s not that people were giving their computers coffee to speed them up, nor were managers ladling coffee into their employees to get more out of them when the real problem was slow PC performance. Instead it’s about boot-up time.
For the past few decades the average worker comes into the office, turns on the PC, then goes to the coffee room to get the morning’s first cup of brew while the PC slowly finds its way to full operation. While in the coffee room that worker may encounter workmates, and delve into a heady conversation about last night’s TV programs, or a sporting event, or even politics. This might turn that one-cup coffee-room visit into a 2-cup session.
Now that savvy bosses are Continue reading
Someone recently asked The SSD Guy if there is a way to determine whether an SSD is SLC, MLC, eMLC or TLC.
I found it a little odd to be asked this, since most vendors tell what kind of flash they use in an SSD’s specifications, especially if it’s SLC.
Not finding it there then the next thing I would look at is the price. Raw SLC NAND flash now sells for about 6-10 times as much as its MLC counterpart, so an SSD with a price of around $1/GB is likely to be MLC and one that sells for around $10/GB is probably SLC.
TLC SSDs are really rare. There is the Continue reading
Samsung recently introduced its 3D V-NAND-based 850 SSD which, according to The Tech Report, uses the same MEX controller as the company’s 3-bit planar SSD, the 840, introduced last year.
Samsung said in its keynote speech at the 2013 Flash Memory Summit that V-NAND consumes an average of 27% less power and runs at least 20% faster than its planar counterpart in an SSD application, all while providing ten times the endurance. It’s only natural to assume that this would allow designers to produce a V-NAND SSD that would significantly outperform its planar NAND counterpart.
Tom Coughlin and I have just released a new report that helps shed a lot of light on a pretty challenging subject: We asked nearly 200 IT managers to tell us how much storage performance their systems require. They provided candid replies about their IOPS, latency, and capacity needs for a number of leading applications.
I have just added a new white paper onto the Objective Analysis website: Matching Flash to the Processor – Why Multithreading Needs Parallelized Flash.
This document examines the evolution of today’s CPUs, whose clock frequencies have stopped increasing, but now exploit parallelism to scale performance. Multiple DRAM channels have also been added to performance computing to add parallelism to the memory channel.
Storage hasn’t kept pace with this move to parallelism and that is limiting today’s systems.
New NAND flash DIMMs recently introduced by Diablo, SanDisk, and IBM, provide a reasonable approach to adding parallel flash to a system on the its fastest bus – the memory channel. This white paper shows that storage can be scaled to match the processor’s growing performance by adding flash DIMMs to each of the many DRAM buses in a performance server.
The white paper is downloadable for free from the Objective Analysis home page. Have a look.