Slashdot Mirror
Will PCIe Flash Become Common In Laptops, Desktops?
Lucas123 writes "With Apple announcing that it is now using PCIe flash in its MacBook Air and it has plans to offer it in its Mac Pro later this year, some are speculating that the high-speed peripheral interface may become the standard for higher-end consumer laptops and workplace systems. 'It's coming,' said Joseph Unsworth, research vice president for NAND Flash & SSD at Gartner. The Mac Pro with PCIe flash is expected to exceed 1GB/sec throughput, twice the speed of SATA III SSDs. Apple claims the new MacBook Mini got a 45% performance boost from its PCIe flash. AnandTech has the Air clocked in at 800MB/s. Next year, Intel and Plextor are expected to begin shipping PCIe cards based on the new NGFF specification. Plextor's NGFF SSD measures just 22mm by 44mm in size and connects to a computer's motherboard through a PCIe 2.0 x2 interface. Those cards are smaller than today's half-height expansion cards and offer 770MB/s read and 550MB/s write speeds."
I don't yet own any flash drives either. I have about 40 magnetic drives. One reason I didn't buy flash drives was write endurance.
I recently found out that the newer Flash drives have the same or better life expectancy as magnetics, though. They have enough write cycles for like 40 years of hard use now, so that's basically a solved problem. Also, when they fail they normally become read-only, so you can copy everything over to a replacement drive. 18 months ago I wouldn't have purchased flash drives, but now that they have improved I will. To reinforce what I read, I have watched Flash drives perform reliably in busy database and web servers. Not that the eight or so flash drives in those servers are statistically significant, but it's nice when your own anecdotal experience is consistent with the studies.
Yes, of course one particular drive might last a long time or a short time. I've had magnetic disks that lasted a long time and magnetic disks that died quickly. On average, an SSD will last just as long as a spinning platter .
Even high IOPS is starting to become meaningless. Here's an Anandtech comparison of top SSDs from two years ago of typical tasks which stressed IOPS. He played it straight for this one page and showed benchmarks in units that matter to people's perception of speed - seconds to complete a task. The result is utterly uninteresting. The HDD is substantially slower. The SSDs are for all practical purposes identical.
But boring graphs are bad for review sites. If the reviews are boring, people won't read them, and the sites lose out on ad revenue. So they invert the metric to make smaller differences appear bigger. Instead of the practical sec/MB, they use the more ephemeral MB/sec. That makes the graphs more interesting and gets people coming back to the sites before buying, instead of just buying some random cheap SSD without really caring about the max speed.
"But sec/MB and MB/sec are the same number! Why should inverting it make a difference?" Because when you invert a metric, the big numbers become small numbers, and the small numbers become big numbers. e.g. Say you have a HDD which can read 100 MB/s, a cheap SSD which can read 200 MB/s, and an expensive SSD which can read 500 MB/s. So in 1 second, the HDD reads 100 MB, the cSSD 200 MB, and eSSD 500 MB. Expressed in MB/s you gain 100 MB/s switching from HDD->cSSD, and a whopping 300 MB/s switching from cSSD->eSSD. Switching from cSSD->eSSD gives you 3x the benefit of switching from HDD->cSSD! So the extra money for the expensive SSD is definitely worth it! Right?
Hold on. Invert to s/MB and say you need to read 1 GB. The HDD takes 10 sec, the cSSD 4 sec, and the eSSD 2 sec. Switching from HDD->cSSD saves you 6 seconds. Switching from cSSD->eSSD only saves you 2 sec. So in terms of time you spend waiting, the HDD->cSSD switch saves you 3x as much time as the cSSD->eSSD switch. The vast majority of your time saved can actually be obtained from the switch to the cheaper SSD. The next step switching to the expensive SSD only gives you a marginal improvement. (Even if you insist on using relative measures of time, the cheap SSD still wins. 10 sec to 4 sec is a 60% reduction in time. 4 sec to 2 sec is only a 50% reduction in time. Or if you want to be a purist, of the 8 sec saved going from 10 sec to 2 sec, the cheap SSD gets you 75% of that speedup, the expensive SSD gives only the remaining 25%)
Unless you're regularly doing tasks where you find yourself twiddling your thumbs for several seconds or minutes waiting for the SSD to finish reading/writing several GB of data, the difference between 600 MB/s and 1.25 GB/s is imperceptible despite being a 2x speedup. Twice as fast as the blink of an eye is still as fast as a blink of an eye to our perception.
The new Mac Pro isn't that great -- and I've been waiting for it. Really had my hopes up.
Flash drives seem to be characterized by very high failure rates. Changing the drive? Unclear this is a user operation. All real drives -- the ones you use for your data -- would have to be external bricks. Whereas standard HD's for the current design go in and out trivially. It's wonderful. Four of 'em.
External drives? External graphics? (3 display max it would seem unless you have external boxes.... yech) Nah.
Best thing right now seems to be the last generation of the big box. 12 cores, 12 more semi-competent hyperthreads, holds four drives, can push six monitors, RAM is (user!) upgradable...
And they finally fixed OSX so it handles multiple monitors correctly, fixed the broken menu paradigm, fixed how full screen apps work... perfect.
The mac pro.... unless there are some real differences between what they say they're making and what they actually make, I think it's the big box for me. My older 8-core can live in the ham shack doing SDR and digi-mode duty. :)
This way I know I can do the big jobs, and without littering my workspace, which I am quite particular about, with bricks and cables. I *really* don't understand what they were thinking.
I've fallen off your lawn, and I can't get up.