Not so far
Flash drives have been rocketing up the hype cycle for the last year, helped along by credulous gadget sites and misleading performance numbers. But the free ride is coming to an end.
Over on Storage Bits (see Flash gets a fight from disks and Hybrid drives: not so fast) I’ve been delving into the differences between flash and disks as well as the flash hype and what flash delivers.
The flash guys deserve some slack
Flash technology is young and rapidly improving with impressive Moore’s Law cost gains. And they’ve been working on the translation layer that mediates between the the weird world of flash and disk drivers. The engineers are figuring out the wrinkles and I have no doubt they will continue to improve.
But we also have to talk about where we are today.
20/20 vision: 20x the price and 20% better performance
Over on AnandTech they posted review last week that tested 3 drives on a 4 GB laptop running Vista Home Premium. The drives were:
- A 32 GB Mtron SSD MSD-SATA6025
- A Seagate Momentus 7200.2 160GB
- A Samsung MH80 FlashON 160GB
They tested maximum data rates – the Mtron smoking at 90 MB/s vs 49 and 38 for Seagate and Samsung – and Vista standby and hibernate modes. You’ll save a few seconds with the flash drives assuming Vista memory leaks don’t make the sleep modes as unreliable as they’ve been on XP and W2k.
But the real surprise were the pathetic application numbers. Here’s a chart I derived from AnandTech’s numbers:
I’d have thought that with double the bandwidth and a fraction of the access time that the flash drive would be a lot better than a 7200 RPM drive. Your average user couldn’t tell the difference between flash and disk based on these numbers.
Update: I’ve added the black line at 0% to indicate Seagate’s performance. The performance percentages are normalized against the Seagate drive. A positive percent means faster; a negative percent means slower. End update.
The StorageMojo take
One set of benchmarks on one product don’t decide the issue, especially this early in the game. But it gives one pause when a high-performance flash drive can’t do any better.
Maybe the translation layer isn’t the problem. Perhaps our OS’s do too many small random writes that hose flash performance. Whatever the case it is too early to assume that flash drives will even take a majority of the notebook market without better performance given their price and capacity limitations.
Comments welcome, of course. Maybe you can explain it.
try zfs?
I’ve found in my informal testing that Vista has massive I/O problems. A computer setup that can exceed 20 MB / sec on a stock SATA drive in standard large file copies under XP SP2, and 25 MB / sec in Linux, is only able to push around 9.5 MB / sec on Vista. I’ve disabled all of Vista’s “smart” features, such as their load-ahead stuff and their indexing, but it doesn’t change the performance numbers. As well, my primary test system sometimes locks I/O, in that open applications stay running, but it fails to do any disk I/O for around 2-4 minutes, before recovering in a flurry of queued I/O requests that lasts almost as long as the lock period.
What I suspect in reference to the Anandtech testing is that there is larger performance gains to be had from flash drives, but massive bugs in Vista’s I/O are failing to show them. I’d be interested in seeing performance comparisons from Linux, ZFS on Solaris, or even XP SP2, as I think they’d be a lot more valid.
It would be interesting to add energy efficiency to the equation as well. How well do the Flash drives hold up in that respect? I assume very well…but I don’t really know how much power it takes to write data to falsh…
Am I wrong to note that the “flash vs. disk” comparison graph you posted is comparing the pure-flash SSD to a flash-hybrid Samsung drive, not the conventional Seagate drive? (I believe the macroscopic numbers about marginally better normalized overall performance, and the article as a whole matches your analysis, but the key on the plot you included seems odd given that you’re drawing a flash vs. conventional disk comparison here, and not comparing flash to a conventional disk.)
Jonathan, thanks for the comment – it made me realize the chart wasn’t as clear as it should be. I hope the update helps.
Yeroc, flash is very energy efficient. The numbers from AnandTech were on the order of almost zero at rest and 0.55 W when writing. The Seagate was about 5x that when writing.
In the context of a laptop though, where disks account for about 15-20% of the power load, a flash drive will turn a 3 hour laptop into a 3.5 hour laptop. Good, but how many people will pay $500-$1000 for 30 minutes more battery life? As part of an integrated design using several power-saving techniques an SSD makes sense. As an upgrade not so much.
Graeme, interesting numbers. The AnandTech tests I referred to were done on Vista and they did better, so it isn’t clear what the issue might be. I haven’t played with Vista myself, but my big concern is memory leaks – a long time Windows problem – making the faster startups from standby and hibernate modes moot since the system will crash as it attempts to exit standby.
Robin
I know you mention memory leaks twice in reference to XP suspend issues but are you sure it is memory leaks that are to blame? I’ve heard (and only have limited proof by watching machines with the same image but differing hardware) that repeated suspension failure is often due to driver bugs and hardware/BIOS faults…
Anon,
I don’t think you can be sure of *anything* with Windows. However, memory leaks are a known problem. Driver and BIOS problems wouldn’t surprise me either. Thanks for relaying your experience.
Robin
Amdahl’s law. If disk access accounts for 40% of application execution time, a 100% improvement in disk access time will only lead to a 20% improvement in overall application performance. There is a law of diminishing returns and it is always important to build a balanced system. SSDs make the most sense for specific hot spots like Oracle redo logs.
Of course, treating flash like disk and having the potential I/O throughput bottlenecked by slow and latency-inducing SATA or SAS interfaces does not help.