Slashdot Mirror
Are SSDs Finally Worth the Money?
Lucas123 writes "The price of 2.5-in solid state drives have dropped by 3X in three years, making many of the most popular models less than $1 per gigabyte or about 74 cents per gig. Hybrid drives, which include a small amount of NAND flash cache alongside spinning disk, in contrast have reached near price parity with hard drives that hover around the .23 cents per gig. While HDDs cannot compare to SSDs in terms of IOPS generated when used in a storage array or server, it's debatable whether they offer performance increases in a laptop significant enough that justify paying three times as much compared with a high-end a hard drive or a hybrid drive. For example, an Intel 520 Series SSD has a max sequential read speed of 456MB/sec compared to a WD Black's 122MB/sec. The SSD boots up in 9 seconds compared to the HDD's 21 seconds and the hybrid drive's 12-second time. So the question becomes, should you pay three times as much for an SSD for twice the performance, or almost the same speeds when compared to a hybrid drive?"
Great, first post but only because I noticed a hole in the latest revision of the system.
And I'm not sure it's a hole. Would it have still allowed me to post without the post anonymously box?
And the actual question still stands- is the memory/storage paradigm just traditional at this point, or is it still useful?
SJW: a person who perceives an injustice, and while correcting it, commits a greater injustice.
The summary mentions hybrid drives, but I can't seem to find any for desktops - am I looking wrong, or do hardware makers assume a desktop user like me doesn't want one?
Joy! Beautiful spark of the gods!
Remember, no spinning platter means you don't have to worry about bumping a gyroscope - an SSD is inherently more shock resistant. I'm under the belief an SSD uses less power than a HDD.
I have one SSD. It's in my netbook, I removed my perfectly functional factory HDD and replaced it with a smaller SSD since I really don't need my storage space, 90% of what I do with my netbook is on the web browser, and a netbook with Kubuntu and the netbook/tablet desktop is way cheaper than a Chrome book. I wish those were cheaper, I would practically be a marketing exec for those without the outrageous pricetag, but never mind that.
There's advantages other than performance to an SSD.
The preceding post was not a Slashvertisement.
You could just get a smaller drive, which isn't too much of a hindrance in most cases. All of my media is streamed from a server anyway.
I bought my first SSD-equipped laptop back in 2007. It was a Dell XPS. The laptop still works great today and flies in comparison to this brand new, work-issued HP laptop -- even with it's 7200rpm drive.
There isn't any comparison whatsoever. And throughput is almost moot, it's the IOPS that matter.
or about .74 cents per gig
Wow, $0.0074 per GB! That's cheap!
I am willing to pay a large premium for storage device that won't break if I drop it a smallish distance.
if Lucas123 can answer that question, he'll find his answer.
If you're pushing the cloud so much is storage much of an issue at all? Seriously I can put Chromium OS on a 4GB thumb drive and boot up a laptop and do web stuff all day long.
Sometimes people don't have access to the internet but still need a computer. Remember the old days before the cloud existed? Yeah - you can't get on the Internet everywhere. Some of these rural areas people still think dial up is not only an option but they still think it's normal.
The preceding post was not a Slashvertisement.
What's worth it is what you're willing to pay. For my part SSDs were worth it when they still cost three times more than they do today. To someone else not so much. So what's the value in the question?
There appear to be a couple of extraneous decimal points in the post. If there's someplace that I can buy hard disks for 0.23 cents per gigabyte (a bit over $1.00 for a 500-gig drive), I haven't seen it.
Dropped by 3X? Dropped by three times what?
Is that the same sort of thing as "todays temperature is twice as cold"?
I think you meant "the price has dropped by 2/3rds" or "prices today are 1/3rd what they were 3 years ago".
"Would it have still allowed me to post without the post anonymously box?"
Sure, that's why we pay the big bucks here. :-)
Even two years ago, I configured my then new laptop with a 160 gig SSD for $150 more and I felt it was worth it given the speed gains. That same SSD now boots Windows 8 in 7 seconds, Photoshop CS6 in 5 seconds (first boot), Word 2010 (first boot) in a fraction of a second. I use an external drive for media. After that first SSD, I now always configure my laptops and desktops now with a SSD on the primary partition for the OS install and application installs.
The biggest performance boost of an SSD compared to a traditional harddisk is random access times, this is what matters a lot more than sequential read performance.
That and a computer without any moving parts is just so nice and quiet.
Since bandwidth is not unlimited, nor is it always connected, I would say the paradigm is as valid is it ever has been.
Cloud Storage is just a re-branded version of what people have been already doing for decades, and thus factors in the same basic manner. There are what, about half a dozen levels of memory between a remote server and your CPU? Each one is a trade off between speed, size, and cost.
Putting a SSD as my OS/game drive has made by far the largest difference I've ever seen in a single upgrade.
In the past it was: "More ram..ooh yeah bit smoother...Faster CPU, bit peppier..." Etc, helped but not blow your socks off.
You put an SSD for your main apps, OS, and games, and it will astonish you how quickly things go. Firefox and other apps load instantly. When I had a macbook pro I swapped to SSD and normally the icons for my startup stuff would bounce for a bit as they loaded etc. After SSD like 5 icons would do a half bounce and bam all 5 loaded done.
So for a desktop, do what I do. Throw a big spinner in there as a drive for games you don't need a fast HDD on, media, etc. Then you will have the best of both worlds. It is by far the least buyers remorse I've ever felt on a PC upgrade.
It's easier to fight for one's principles than to live up to them.
The result for opening the word document which shows the SSD performing worse than the others (57/10 sec, 48/9 sec, 58/10 sec.) is odd. I didn't notice the author mention how many times he performed his tests, so I am going to assume he just performed them once.
I would like to see this result repeated several times to verify whether it is an outlier, or whether an HDD will have such a large impact on MS Word performance (which TBH I would expect was mainly CPU bound).
SURELY NOT!!!!!
If you compare sequential reads it's obvious HDs seem to have a chance against SSDs. It's in non-sequential reads where SSDs completely outclass any HD.
The post makes it sound like Hybrid is close to SSD ... it is not ...
Max. read speed (4K blocks)
SSD: 456MB/sec.
Standard: 122MB/sec.
Hybrid: 106MB/sec.
Max. write speed
SSD: 241MB/sec.
Standard: 119MB/sec.
Hybrid: 114MB/sec.
1.19GB file transfer
SSD: 15 sec.
Standard: 34 sec.
Hybrid: 29 sec.
For a serious computer user, an SSD has been worth the money for a while now.
* If you need to do serious disk I/O with a mid-size or smaller notebook, RAID isn't even an option for increasing speed.
* Running multiple virtual machines? Want them to boot quickly? An SSD makes them feel native.
* Running Windows as a native operating system, and have more than one or two programs that you legitimately want to launch at boot, and can't/won't disable? An SSD makes your computer usable within tens of seconds as opposed to multiple minutes.
* Doing compilation? Syncing of filesystems with a system such as Unison? Doing anything filesystem heavy? The speedup is insanely awesome.
If all you care about is running Your Web Browser and editing Word documents, or storing a few photos, obviously an SSD is a more questionable upgrade, and probably will be for the foreseeable future.
Insert self-referential sig here.
The factors have changed a little, but the basic equation remains the same. There is a trade off between cost and size, so how much space you need is important.
For instance, one of my computers has a 16GB SSD. I am not even using up that much space, so any larger a drive is just wasted, and at 16GB the cost differnce between SSD and a good platter drive are not that huge, so it make sense. I do my photo editing on a computer with a 120GB SSD, with a large platter drive attached externally for storage. Again I do not need huge amounts of space in the box itself, so the advantages of the SSD outweigh the minor advantage of 'it could be bigger for the same cost' of another drive.
One answer: http://en.wikipedia.org/wiki/Betteridge's_Law_of_Headlines
In reality it depenfs on the intended use.
If space isnt an issue then go SSD, if it is then a hybrid solution is best for the average joe who doesn't know or want to bother with splitting apps and os onto different drives. If you are power user you don't mind that as much and build your own combo solution.
In other words, until ssd_price == hdd_price, solutions will vary based on use case, and Bettridges law holds true once again.
Silence is a state of mime.
It's still useful. The random access latency on an SSD is still about 1000x slower than RAM, but SSDs can store data without consuming power.
Keeping a terabyte or two of current RAM technology active requires substantial power supply and cooling, whereas these amounts of SSD or more can be kept and used in mobile or residential situations.
I took the plunge a year back and bought an SSD for my laptop, it was the best decision I have made I'd say. Performance has been stellar, and my laptop is so quiet without the click-pop-buzz of the HDD... truly a worthwhile investment for me personally.
Prices dropping means I'll just be buying more of them to put in other machines I own. Quiet, fast and (comparatively speaking vs a few years ago) inexpensive makes SSD a winning combo.
SSDs are now at a price where it's a no-brainer for a media-PC hooked up to a NAS. They're pretty much cheaper than the cheapest normal hard drive you can buy, and far quieter.
For the desktop, the cache drives still make the best sense. Most users don't have the technical ability to be able to force installs of software to secondary drives and keep their boot drive clear of clutter enough to be able to warrant the cost of an SSD.
I went for a cache drive myself about 4 months ago and it's been one of the best purchases I've made for my home system, but for joe public I'd still say a RAM upgrade should come way before anything else if they want things to just work faster.
I travel a lot, including a lot of hitchhiking, not like your average business trip. Consequently I've had to deal with several broken hard drives. Since I switched to SSD this hasn't happened anymore. That means I've likely saved some money by not having to buy a new hard drive and a lot of time from having to deal with a broken hard drive.
I have a laptop that has two drive bays. SSD's won't be replacing the 1TB main hard drive any time soon, the prices for those are more than the laptop is worth.
But a 256Gb is surprisingly affordable and given that my "primary" partition is that size, it would be a cinch to install one, move the data over and even mirror the partitions to a traditional HDD if I needed to.
And the speed difference *would* make a huge difference. It always has done.
The problem is not the speed increase relative to anything else because it blows things out of the water. The problem is cost per Gb, as always in storage. 256Gb now costs about one-third the price of my laptop, or a half-decent graphics card for a PC. That's within the realms of possibility for an upgrade.
$0.74/gig is very different from $0.0074/gig...
I'm pretty sure that OP used to work for Verizon...
coding is life
This goes for just about everything unless you are talking about generational technology gaps. Honestly, double the performance for 3x the price is pretty much a bargain, comparatively speaking.
An item yet unmentioned at the time I post this, is SSD lifetime. The are finite, you know, and probably a lot more finite than a well-protected HDD. The manufacturer states the number of write operations the storage cells can take on average before going kaput, and its up to the controller & OS to "age" them all equally to ensure maximum longevity (thanks, TRIM). This and speed are the main determinants of the cost of the devices and the differentiator between user and server-grade SSDs.
Nowadays with shady outfits jumping onto the SSD bandwagon, we'll see really crappy devices made from rejected storage chips hitting the markets, which will fail prematurely and give the technology a bad rep.
The cloud isn't nearly fast enough or cheap enough to replace any sort of local storage. That's not even getting into the obvious question of reliability and availability that so many people like to just gloss over.
A Pirate and a Puritan look the same on a balance sheet.
Anyone else tired of seeing comparisons between a massive capacity magnestic drive and SSDs?
Indeed. The whole question of " should you pay three times as much for an SSD for twice the performance" is mis-formulated. The question really is would I pay 20% more for a laptop that boots twice as fast, never has data access lag from a sleeping drive, performs noticeably better on frequent persistent storage access scenarios, and has substantially better battery life?
Starships were meant to fly, Hands up and touch the sky - Nicky Minaj
Some of these modernized areas internet access is not fast enough, even for the home user.
When it matters, I still can't depend on my wi-fi connection via my cellphone - which, to my mind, means until someone tells me the entangled particles in said device are good anywhere in the universe or my money back, then "the cloud" is not something I want to rely on having.
Or secure enough.
I bought my first 160GB SSD for $600, and would do it again in a heartbeat. If you use your machine for any kind of productivity the speed difference is night and day, moving to a SSD is the single most noticeable improvement in overall speed of my computer that I have EVER DONE. About the only thing I can relate it to is 20 years ago when I upgraded to a 3dfx graphics card for the first time, and seeing a software 3d engine vs the new hardware one. Now if you only use your computer for word processing or internet usage as most people, there is little reason to upgrade except for perhaps data security as SSD failure rates are much lower, but since you should be backing up anyways that shouldn't be much of a consideration. I have 3 SSDs totaling 700GB and will never buy another computer again without one.
After 25 years or so of slapping upgrades in computers of various sorts, I'd have to say an SSD made the most immediate, noticeable difference of any upgrade I've done. Better CPU? Yeah, the new one's a bit snappier...I think; or maybe I want to think that because I spent money. More RAM? Seems like it's not swapping as much, sure. Replace spinning platters with SSD? Did someone just secretly swap out my old computer for a new one? Everything seems faster (okay, not ripping DVDs in Handbrake).
Forget boot times, who reboots enough to even notice? App loading, compiles, anything involving disk access is nearly instant. I'll sacrifice capacity for what an SSD buys me.
Now I'll admit that I wasn't as impressed as I thought I should have been. Two years ago when I bought my first one, bloggers were wetting themselves a bit much over the extra snappiness of an SSD. But SSDs are still a damned impressive upgrade. I really noticed the difference when I went back and forth between my SSD-equipped MacBook Pro and my iMac with a better CPU but 7200 RPM hard drive. When the iMac hits disk, it's annoyingly noticeable.
In summary, SSDs have been worth the money to me for over two years now. The only spinny hard drive I'll be buying from now on will either be a secondary drive, or will go in the NAS.
Cameras and camcorders dude. In case you haven't noticed cameras and camcorders are dirt cheap, and between that and the 5MP+ ones being built into the smartphones I find customers just chewing through space. An SSD is fine for a mobile as long as you don't have any cams involved, but if they want to take and edit photos? Get a hybrid or plane Jane HDD, they'll need the space.
ACs don't waste your time replying, your posts are never seen by me.
The question misses my key factor: Reliability.
Yes, SSDs have a limited lifespan, but it is relatively predictable.
HDs on the other hand, especially with as much of a commodity (meaning nearly non-existent quality controls) as they have become, are completely UNpredictable on reliability.
The same HD from a different batch might fail nearly immediately whereas the very next production run might produce a drive that will last for many years.
I got VERY tired of it.
I run SSD for the majority of my apps. My data I stick on a separate large mirrored array.
The hybrid drives may be fairly cheap, but they are inherently as unpredictable as HDs (they use the HD less, which is a bonus, but they add a second layer of complexity, which is a detractor, so I end up considering them equivalent).
I had some problem with my first SSD due to firmware issues ... but once cured all of my SSDs are still running solidly while I've had multiple HD failures of newer HDs.
It is more productive to voice thoughtful opinions (reply) than to judge (moderate) others.
Having done a number of HDD->SSD upgrades for friends and family, I can tell you this quite simply. Anyone asking the question has never used an SSD, because if they had they wouldn't be asking it.
How a desktop "feels" to the user isn't about raw throughput, but it is very often about IOPS and more importantly latency. It may not seem like waiting 5-8ms for the rotational latency of a drive is a big deal, but spread that out over a pile of IOPS and it is a huge deal. The original post even shows how much, boot time with an SSD was 9 seconds, HDD 21. That's 50% faster. Now probably most people don't care if the boot time is 9 or 21 seconds, but I bet most folks would like their system a lot better if every application load time was 50% faster!
SSD is the single biggest no-brainer upgrade to me, it's even surpassed the "add ram" no brainer. The only time SSD's get questioned is for bulk storage. If the users needs include large music, photo, or video archives then it is worth asking questions about the cost of storage. Even in those cases, going with a hybrid drive or two drives is always the right answer.
If you aren't just a digital couch potato, you will find value in a storage device larger than 64G. This pretty much eliminates using SSD exclusively unless you see your PC as a glorified iPad.
This group of users that are more than just consumers also includes people like grandma.
A Pirate and a Puritan look the same on a balance sheet.
My smart phone came with PhotoBucket turned on. I do store pictures on my Terrabyte NAS, just to have a local backup, but when I share, it's the Photobucket or some other cloud storage solution.
SJW: a person who perceives an injustice, and while correcting it, commits a greater injustice.
If you spend most of your computer time loading and closing applications, booting or rebooting, or sifting through large directories of files - then an SSD is probably a good buy.
Personally, I don't spend much of my time doing any of those things. In the morning my computer resumes from sleep in 3 seconds. My web browser is still open, and I can continue from my last session. There is hardly any disk access. My work involves a lot of writing, a lot of programming, and a lot of MATLAB. None of these things involve disk accesses, because I already started these applications a month ago when I last booted.
Occasionally I play a game. Except between levels, there are no disk accesses.
For me, given that my hard drive is hardly accessed as it is, I don't really see the point. The only time I've ever wished for an SSD is when I run out of RAM. Make sure you have enough RAM, and - at least in my experience - disk speed is rarely an issue. I have 1.5 TB in my laptop, and couldn't imagine sacrificing that space just to speed up my once-a-month boot.
"The price of 2.5-in solid state drives have dropped by 3X " Hmmmm.... where I come from, if something is reduced by 1X, it's at zero. Perhaps he meant it's dropped by 1/3, or 2/3".
There is no God, and Dirac is his prophet.
I would not build any new system that didn't have a SSD. This applies to both desktops and laptops, but the advantages on a laptop are even greater. Mechanical HDDs for laptops are slower than on desktops (due to the requirements of form factor, power consumption, noise, heat, etc.) so there is a bigger relative jump in performance than there would be on a desktop. And there's also the durability issue! SSDs are immune to vibration and shock; laptop mechanical HDDs, though they are better than they used to be, can still be blown out of commission by enough rough handling. For people who routinely carry their laptops from place to place, this is perhaps the biggest improvement of all.
Many users will be fine with a 128GB SSD and no secondary drive. You only really need more than that if you're storing a ton of videos or music, or a SHIT-ton of high-resolution photos, or are installing a bunch of A-list games or high-end CAD software (but people like that are already buying or building multi-thousand dollar desktop workstations).
Intel's "Smart Response Technology" operates under RAID drivers, but does NOT support RAID. If the SSD fails, you lose data.
Unless you live in Kansas City.
drool....
I have no problem with your religion until you decide it's reason to deprive others of the truth.
I bought an SSD for my netbook about two years ago. Last I looked, the 'wear indicator' in the SMART data had reached 1%, so in theory it has a couple more centuries to go. That's running Linux with no swap and /tmp and the Firefox cache in RAM.
At the other end of the scale, people who bought SSDs to improve compilation speed have reported burning through them in less than a year. But the programmer time saved easily pays for buying a new SSD every year.
Take the only moving part of your PC subsystem and swap it with something that has none. Fans don't count. It is the only upgrade that you will instantly see a significant difference in response times. Seriously, ask your neighbourhood PC gamer. Since OP has issues with fractions, I'll put this succinctly, get an SSD. Whichever one you can afford. Even if you have to move software storage to a seperate disk from OS. I've never tried hybrids, but with 128 GB SSDs under $100 at Newegg, there's never been a better time to lose your HDD as your boot drive.
I smell FUD, and that's kinda bad in view of the power consumption figures being explicitly stated in *easily* publicly available datasheets
Let's see how much supply current is needed to self-refresh a 1 terabyte of DDR3L SDRAM.
Let's look at 8 gigabits MT41K1G4 chips from Micron. The chip takes 28mA max at 1.35V. That is 37.8mW per 8 gigabits. A terabyte has 8000 gigabits, or 1000x as much -- that's 38W or about as much cooling as a CPU found in someone's desktop PC might dissipate.
If powering and cooling one CPU is "substantial power supply and cooling", then, well, obviously we've got different points of view on this stuff.
Do notice that those chips dissipate more power only if you access them, so 38W is the idle state but even if you *do* access them, you don't dissipate all that much more -- you'll be probably only accessing a couple of chips at a time. The worst case all banks interleaved read current on those chips is 320mA, so if you access 4 chips at a time, that's still only 1.8W of extra power on top of refresh power.
Of course the logic used to piece together all the chips into a storage device will also use up power, but that logic is in idle low-power state when the chips are not being accessed, so it's a big deal.
A successful API design takes a mixture of software design and pedagogy.
s/it's a big deal/it's not a big deal/. Slashdot, seriously, make a time-limited edit button, will you?
A successful API design takes a mixture of software design and pedagogy.
Slashdot, seriously, make a time-limited edit button, will you?
+1 Hell Yea
An enigma, wrapped in a riddle, shrouded in bacon and cheese
Not to mention the same reliability and security issues we've been dealing with since it was called client/server.
I've found the best strategy is a multi-level backup solution. My customers have their important data on their systems, backed up to USB hard drives they carry off site (not large enough for tape drives to be viable) and in the cloud. This way even the worst events won't wipe out the data. Last year I had a customer who came back from vacation to find his business burned to the ground, he simply picked up the USB hard drive he had left with a relative before leaving and I slapped the data on one of my spares which i let him hang onto until he could get the insurance mess straightened out, he was back up and running before the end of the day.
So I don't think anything is replacing anything, at least not if you're smart. they all have their pluses and minuses, best to use them in concert for best results.
ACs don't waste your time replying, your posts are never seen by me.
...then the answer is no.
Honestly, this question seems to be coming up a lot lately. As with any tech, it either meets your needs or it doesn't, and if it doesn't then its not worth it at any price.
Truth is like the sun. You can shut it out for a time, but it ain't goin' away. - Elvis Presley (source: imdb.com)
While HDDs cannot compare to SSDs in terms of IOPS generated when used in a storage array or server, it's debatable whether they offer performance increases in a laptop significant enough that justify paying three times as much compared with a high-end a hard drive or a hybrid drive.
It's only debatable if you are severely limited in budget or have SSDs have pretty much every advantage except price. Even if the price is 3x as high, the cost of the hard drive is only a smallish percentage of the whole cost of the device - maybe 20-30% total. While price is an important consideration, if my budget can accommodate an SSD I'll go with it every time. Sure, if/when I need a few terabytes of storage space then a spinning platter is the way to go (for now) but that's not true of most devices anymore. I have a server for mass storage needs but 128GB-256GB is usually more than enough for any day to day needs and a SSD in that range is affordable already and dropping fast. My phone and laptop and my primary desktop all have solid state drives. I have two spinning platters in my house - one in an older desktop that sees limited use and the other on my file server. The new laptops we're buying for work have are solid state as well. I don't see myself ever buying a laptop without a solid state drive ever again.
If speed were the only benefit of SSD, that would be a valid question to use when evaluating a purchase. SSDs additionally offer substantially reduced power usage and - the deciding factor in my case - the substantially increased durability that comes with no moving parts.
The correct answer is it depends.
If your main use of the computer is serving the web, checking emails, typing some documents, then the extra cost of the SSD isn't going to make a major difference. OTOH, if what you are doing is somewhat disk intensive, then it very well could.
As with most things in life, technology or otherwise, wants usually carry more weight with one's decisions than do needs. Marketers count on that.
Each one is a trade off between speed, size, and cost.
Well, and like you said, availability. And features, I suppose. The "cloud" storage buys you a little safety, but (excluding local cacheing) is pretty darn slow and expensive. Working the MRC on most storage services against a real drive, the outside service is going to lose on $/gb alone. It's still handy though.
True SSD's, on the other hand, are speedy and are a good addition in laptops for their lower power use and lack of moving parts. They're just more expensive. But I figure, it's not like swapping a six cell to a nine cell is cheap, either.
I actually ran the numbers on this for my company. Based on average usage on our standard laptop image and typical employee salary:
$1.82 saved in salary time per bootup (assume one bootup per day)
$2.23 saved in salary time per day due to files opened/programs launched
That's $4.05/day saved due to time I'm not waiting for my hard disk.
ROI for a $300 aftermarket SSD is 75 working days, after that they're effectively earning back ~$1000/year. Considering that our replacement cycle is 3 years, that pays back the purchase cost of the hardware. My boss now buys SSD upgrades for all of our new laptops.
On a personal note, I happily payed $1.00/GB for a hard drive several years ago, and thought it was a pretty good deal. I retired that drive only last month (too small for even my kids' computer these days). Now that SSDs are $1.00/GB it's an easy sell to my wife, and she sees every day the difference in boot times between her desktop and the kids' one (which she used to use until a year ago). I don't think I'll ever run a spinning platter HDD as a boot drive again.
"Space Exploration is not endless circles in low earth orbit." -Buzz Aldrin
SSDs are, in case of a typical desktop system, a hardware solution to a software problem. The software problem lies squarely with braindead legacy APIs that last made sense in the 80s. Those are exposed by seemingly every operating system out there. The issue is as follows: when an application accesses storage, it has no way of telling the operating system what are its plans besides the very next access*. The OS can't plan any hard drive access patterns nor do any sort of large-scale elevator access coalescing because, for the most part, it only knows about the very next access a thread wants to do. Using threading as a workaround to this issue is just silly, you don't need multiple threads, just better async and queuing APIs, and programming languages that can actually deal with them.
Say you know that you want to read the entirety of, say, a dozen configuration files, and also want to read some known byte ranges on other files. It's not simple or even possible, as things currently stand, to tell the OS: here's all that I want to do, wake me up when it's all done. There are asynchronous APIs, but those are not in widespread use because widely deployed C-like programming languages are a very poorly suited to dealing with such problems. As in: the code becomes a royal mess. That's why many GUIs get blocked by every file access and whatnot: it's messy to code an event driven application in a C-like language. Clean, linear-flow code becomes fragmented across functions/methods or case sections. Ugh.
On top of that, all higher-level APIs: those that encapsulate file- and network access, almost universally hide the low level operations and do not allow any sort of asynchronous operation from the caller. Just look at every single damn database library: it's all blocking access! Compression libraries: blocking access! File format libraries (scientific, GIS, office, XML, make your pick) -- same thing. There's no way to use such a library to essentially queue a bunch of requests with the OS, that the OS can then elevator sort on, etc.
Same goes for the runtime linkers/loaders: there's no provision, usually, for any sort of parallelism in queueing the file access requests to the OS. The linker/loader will deal with one file at a time: open it, read some of it, process, rinse and repeat, in spite of knowing a priori a large number of such requests that could all be optimally accessed.
Sure, a realtime database system that needs to have lots of random *read* transactions probably must have an SSD, there's no way around it. A realtime system with mostly random writes can use a log, though, data from the log can be fed back to the database pages after being elevator sorted and coalesced as appropriate, trading off battery-backed RAM for HD performance.
*Let's discount the file access hints as those don't make much of a dent in typical use.
A successful API design takes a mixture of software design and pedagogy.
No, it's not debatable. They offer a huge performance increase in both laptops and desktops.
Not only do programs open much faster, files open instantly, hibernation faster etc, but there's no moving parts (in case the laptop is dropped, at least the data is safe), and also SSDs use much less power (improved battery life).
Yes, they are pricey. But it's the best investment to speed up a laptop.
If you think you can get by on 60Gb or less? You are one of the less than 10% that don't run Windows. In another one of Ballmer's boneheaded moves all Windows since Vista has "anytime upgrade" which means it has ALL the files and ALL the patches of Windows Ultimate, even if its Basic or Home.
Because of this a fully patched Win 7 SP1 can easily get up into the 70s when it comes to Gbs and it sure as hell ain't easy to strip all that anytime upgrade shit out. Just one more way the marketing drones fuck up what should be a simple idea.
ACs don't waste your time replying, your posts are never seen by me.
If an attacker is determined to get to your data, Most cloud storage vendors have better security than most consumers' local network/computer. That said, an attacker would more likely attack cloud storage than some random consumer's computer since the payoff is significantly bigger..
$740 for terabyte of solid state disk; $23 for a terabyte of magnetic disk. Perhaps it should have been $0.023 per gig.
Due to the cache mechanism employed by web browsers, loading times for already visited web sites is significantly reduced when reading from ssd. Web browsing will feel significantly faster with a ssd then with additional ram.
On a laptop, for sure, the SSD is a good buy. HDDs are ridiculously huge these days and very few people actually *need* a terabyte of storage on their portable machine. What they do need is the best possible battery time, and hybrids aren't any better than traditional drives for that.
Modern computers were light years faster than their storage systems until SSDs came along. I've heard so many people complain that their hot, new, $2000 computer is still slow as molasses in January, and it's because of the hard disk.
You can have a 50,000RPM hard disk with a 100Gbps interface and you still won't move the heads more than 100 times per second in practice.
Hard drive transfer rate was never the issue. The issue was the speed at which random data could be retrieved. SSDs solve that problem completely.
The cloud isn't nearly fast enough or cheap enough to replace any sort of local storage.
it is if all you're ever doing is spreadsheet and word documents that are small.
Lately they way I've been rolling out IT to employee desktops is a 120 Gig SSD (operating system + software), or 60 gig depending, but 60 gig is pushing your luck if they ever want even one or two big files.
And then all of the employee data is hosted on servers.
I agree, cloud storage isn't anywhere near viable for heavy hitting computing tasks, but conceptually it's viable for some problems (depending on what exactly you want to count as a 'cloud', hosting your own servers is the same as the 'cloud' but it's done in house rather than with a third party, if you're small you need a third party running your IT anyway, and if you're big enough your IT department looks a lot like rackspace already so calling rackspace the cloud and you 'centralized servers' isn't helpful).
Disclosure: I'm by all means not an Apple fan boy nor do I own a single Apple product.
My gf was looking to buy a Mac Book pro and sought my advice. I lobbied hard for the Mac Book air with the SSD. The price was the same for the two 13 " models. Mac Book Pro had 500 GB vs the 128GB SSD of the Air. The pro had a faster CPU and couple other differences which I won't get into.
Considering she doesn't need massive hard disk storage, my selling point was the start-up speed as well as the speed of opening files (this includes application start-up). Sure enough the Air started-up a around 10 sec vs the 45-60 sec of the Mac Book Pro. Elements application started up quicker on the Air as well. Needless to say she was sold.
The time she'll save on each start-up and file location will far outweigh the the time saved on say doing CPU-heavy CAD operations she'll perform 0.000001% of the time. I think far too often people will place disproportionally heavy emphasis on outlier user-cases. You don't need a sword if you are slicing apples 99.9% of the time. Of course, ultimately the decision lies in the needs of the user.
It is disingenuous to call it a "software problem". The underlying problem is a hardware one, ie that seeks on spinning media are fundamentally expensive. You could write software better to mitigate exposure to that problem but that would only be attempting a 'software solution to a hardware problem". You add software complexity and can't solve the problem, only (attempt to) minimise it.
SSDs are a hardware solution to a hardware problem.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
I'd really like to see someone hack my home server. Not just because it runs OpenBSD with some rather paranoid pf rules (I use OS fingerprinting to refuse connections from any known-insecure or evil OS).
Mainly because since Verizon is taking literal *weeks* to even acknowledge my purchase order, I'm forced to leech off a neighbor's Wifi, and my home server has no wireless capabilities.
Second, what about NCQ? The consolidation of hundreds of server processes onto several VMs running on a powerful multi-core computers all accessing a shared underlying storage mechanism does offer quite a bit of opportunity for re-ordering. How much extra benefit would come from pushing this down to the lowest level - individual processes?
Still too expensive even for 64 to128Gigabytes but for linux it's not bad to get 64 GB ssd, for windows 7 probably 128gb would do and just use a small hd for swap file. I would suggest to buy large HD's for storing your movies, software application ISO's, data files, etc... use NAS, or use usb external enclosure(with caution), or those removable drive bays. I use to use usb external cases but those got too hot and failed. Now using removable drive bays for storing my data, ISO's, movies. I had a whole bunch of seagates and western digitals go sour within 6 month's, a year or even 3 years. My hitachi still running for the past 8 month's. As demand increases for SSD's and manufacturing becomes cheaper the SSD will probably end up same price or even lower than HD's.
My brand spanking new i7 work laptop with a 7200 rpm drive feels almost unusable compared to my crappy circa 2006 laptop (amd turion 2) with a no-name (microcenter) SSD. I can start the 2006 laptop, start visual studio, shut it down, start it, open visual studio in the time it takes the i7 laptop to boot and load visual studio once. I don't have benchmarks for this but I suspect that the crappy 2006 laptop would build code faster. It generally feels *much* snappier than the i7.
I would never, ever, ever use a normal HD again unless I had to (see: work laptop).
s/it's a big deal/it's not a big deal/. Slashdot, seriously, make a time-limited edit button, will you?
Proofread your comments.
It doesn't mean much now, it's built for the future.
Clearly the reader is expected to solve for X.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
Answer: No.
Why? You ask. Because SSD capacity is still useless - they are in the low hundreds of Gb, if you are lucky.
Fallout New Vegas ultra edition: 25 Gb.
Old Republic: 25 Gb
Battlefield 3: 14 Gb
Any total war game: 15+ gb
Most MMOs: over 16 gb
You can't fit everything, you'd waste at least half your hard drive capability with solid states for your important games and OS. Assuming you're a gamer or you do 3D work. One day...one day my friends.
Actually, I was seriously considering SSD for my database server. I have a large, ill-organized database that needs high performance. The rest of my server system can get by just fine on traditional hard drives, but the database tables just happen to be about the right size for recent-vintage SSDs, and if they'd save me enough on latency, I'd gladly pay the premium.
If I'd been a little braver I would have gone that route, but I didn't have time to discover tuning tricks for SSDs and I managed to get the system tweaked enough to meet deadline. One or 2 more price cuts, though, and It's Xmas, I think.
I find the whole article stupid.
ssd's have been well worth the money for two years now. it's just that much faster.
the blurb sounds like a hybrid drive advertisement. smartdrv only gets you that far before you'll need to hit the disc. sure, a hybrid with 100gb of nand would probably compare favorably in the long run, but a regular hdd vs. sdd... then it's not really a debatable which one is faster, except in the sense that you can also have a debate about if hitler lost or not(revodrive which is mentioned in the article has 100gb and goes in pcie - actually even mentioning it in the same article with the seagate is stupid, like mentioning a ferrari hybrid that has a power boost from the electric when someone is trying to sell you a hybrid yaris which makes no sense, even if it technically does the same).
the current momentus hybrids have 8 gigs of ssd in them(this information is not thanks to computer world or seagate! the older smaller model has 4gb btw). sure, it makes for faster boots if you do three boots in a row. but consider this: it's not now unusual for games to take over 4 gigs, sometimes over 10 gigs(hell, max payne 3 is 30 gigs installed from steam) and many other things as well. so the optimizing algorithm is going to have fun time figuring out what to keep on the ssd portion - it's pretty much a benchmark cheat more than anything else.
in short, computer world sucks as usual and the article is a hybrid hdd advertisement. "save a few bucks and get one of these! it's excellent if you're budget oriented!".
(disclaimer, my laptop has both ssd and hd. and yes both a car analogy and a hitler reference)
world was created 5 seconds before this post as it is.
I got into SSD's when they just barely hit $1/gig on special as they came up on Techbargains.
I would do it again at that price. The speed up was ridiculous. 3x on my old laptop, 6-8x on my
gaming computer.
Needless to say... I had a sniiif when I started seeing them @ .50c/gig on Techbargains.... ;(
But I've been doing this a while, I've payed plenty of early-adopter tax.
I was buying $600 smartphones while 98% of people were carrying stick/flip/feature phones.
It'd be nice to have all that money back now... but meh. Yolo.
-AI
For me, it is far better to grasp the Universe as it really is than to persist in delusion
Provided you're using them for your OS and software, and not storage of large quantities of data.
And games (yes, I know... "software") but I'm just sayin... Skyrim load screens take like less than 10s now.
-AI
For me, it is far better to grasp the Universe as it really is than to persist in delusion
Flying back home recently, there was a spare seat between myself and another passenger. The other passenger left his MacBook on the seat accidently as we were landing.
Upon the thrust reversers coming on, his MacBook flew off the seat and slammed hard into the floor.
Well, it was a new one with SSDs, and it was OK!
I'm sold.
No reason you can't have a second, platter based drive to hold your extra data. Or, if you're using a laptop, an external drive.
It's not a big deal compared to say a space heater, or microwave, but compared to the power usage of an SDD? Not to mention any additional active cooling that needs to be done. Personally, I'd rather have that other CPU than a bunch of RAM that mostly won't be used.
Well, if you feel the need to keep a dozen or more games installed on your system at any time, and choose not to have a larger capacity secondary disk to hold the stuff you're not heavily using right now, then that's your problem.
You've obviously never used a SSD -- otherwise your answer to this question would be very different. My first SSD was only 80GB, and Windows took up almost 20 of that. Even with that limitation, and having to juggle which games I had installed on the SSD at any given time, it was still worth every penny of what I paid ($300). The difference between platter-based drives and SSDs is so great that many people feel that it's worth the hassle of juggling installed apps in exchange for better performance.
Because I really want to have to lug an external drive everywhere I take my laptop just so it will boot a few seconds faster the one time a day I boot it up.
Bingo - drop survivability and heat generation. These are two of the best reasons to use SSD in a laptop, and not HDD. Nothing to do with performance.
Solid state drives are pretty much better in every meaningful way except price per GB. Speed, shock resistance, noise, heat, latency, and power consumption are all better in solid state drives. If you need a lot of storage space (terabyte+) a spinning platter remains the way to go for now just due to cost but otherwise there really is no other advantage to them. Price is an important consideration sometimes but unless you are on an extremely tight budget or need huge amounts of space, I can't really see any reason to pick a spinning platter drive.
best option for Macbook Pro owners is to swap out the DVD drive for an SSD.
I kept my 500GB HD in my 2010 Macbook Pro and dropped a 256GB ($170) SSD into the connection that my DVD drive used to inhabit (with $12 adapter from Amazon.com).
OS and all apps go on my SSD. My home directory (movies, music, photos, etc) go on my HD.
My God can beat up your God. Just kidding...don't take offense. I know there's no God.
SSDs have been worth it for years on desktops and portables, for the speed, low noise and low heat/power. But not for mass storage, only because of the cost. IMHO it will be decades before rotating media is completely displaced in the latter, if ever. But SSD is moving steadily down the food chain, and certainly is the preferred solution for personal use unless budget is really tight.
My solution: I run my workstation root filesystem on a modest sized SSD including my home directory and keep the big stuff like photos on rotating media, spun down. This is really easy on Linux, just install noflushd. You do not want to swap heavily in this situation, it will quickly eat the SSD, so have plenty of RAM. Not that that's the least bit unusual these days.
When all you have is a hammer, every problem starts to look like a thumb.
An item yet unmentioned at the time I post this, is SSD lifetime. The are finite, you know, and probably a lot more finite than a well-protected HDD.
The evidence that HDDs have a longer lifetime than SSDs remains rather inconclusive. Most of the data I've seen is either manufacturers data that should be taken with a huge grain of NaCl or anecdotal evidence with tiny data sets. Even if they do actually have a shorter life, I'd argue that the difference is relatively small basically meaningless. You really shouldn't trust either type of drive to be reliable. Data should be backed up and you should basically assume that your drive is going to fail at any moment because it might. SSDs don't actually have to last longer than HDDs, they just need to last the useful life of the computer. Anything longer is basically pointless.
Some of these rural areas people still think dial up is not only an option but they still think it's normal.
Have you seen the other options in rural areas, satellite $60 for 1Mbps down and 0.2Mbps up, DSL if the phone lines are newer but typically they are not, LMDS systems are also available but spectrum costs money so it's fairly expensive too. There is not a really good option for getting the internet out in rural areas.
Knowledge = Power
P= W/t
t=Money
Money = Work/Knowledge so the less you know the more you make
It wasn't a question of "5 times as much an SSD" (and of course much more if the SSD is idle or off) it was a question of "too much power and heat to be used in a residential situation." It's foolish to have 1TB of ram in a desktop and then never be able to turn it off without losing everything, but not because the power consumption is so high that you couldn't use residential wiring ;)
ASCII stupid question, get a stupid ANSI
At $740 per terabyte (that is, $.74 per gigabyte; over 10 times the price for rotary storage), this is hardly a "good deal." To my mind, the only reason to invest in solid state drives is if you need to do something that mechanical drives might not be good at--for example, portable hardware that might get dropped.
Since my terabyte data drive regularly hangs out at around 80% full, it's simply not acceptable to pay solid-state rates for the storage I need. Also, this doesn't take into consideration the fact that SSDs aren't yet able to make drives of this size very readily (last I checked). A TB drive is therefore likely to wind up being a lot more than even $740.
I'm sure other people have uses for the things, though.
I said most consumers. People who setup a home server with routing and firewall capabilities would probably not be in the "most" category I was talking about..
For me, no. Too many people I know have experienced catastrophic failures of their SSDs. I can't remember the last time I had a rust drive fail catastrophically with no warning like SSDs do. When Western Digital starts making SSDs, I'll consider them. Until then, a fast enough spinning rust disk and gobs of ram is good enough for me.
Our database server at work is going to be getting SSDs in a RAID because we need the speed. We will see how well they actually work.
I'd pay 3X for a 1TB SSD for my MacBook Pro... now shut up and take my money.
Ooops, too bad they aren't $300-ish, but over $1,100
Ah. You are right of course it's not too much to use residential power. It is enough to be a noticeable power bill (potentially more than the cost of the SDD over the life of the computer). I got the impression from the ggp that he was referring to reasonable, not absolute max possible. 38+ Watts in a mobile device of any kind is certainly not reasonable. In a desktop that doesn't absolutely need it, seems overkill as well even if will run just fine.
Highly compressible data only applies to Sandforce SSD controllers. I'm not aware of other controllers currently using compression. If you are planning to buy an SSD with a Sandforce controller then you should read the reviews on reliable sites that test using data of various realistic types. I like Anandtech myself.
SSD IOPS and RAID IOPS are quite different. A RAID made of hard disks may have a very high IOPS rating when it is using a high queue depth, but when using a queue depth of 1, its IOPS will be the same as a single drive. An SSD on the other hand, will have a very high IOPS rating even when using queue depth of 1. And most desktop software only issues requests at QD 1.
If you are planning to read 2 GB of images a single SSD will still beat a single HD. If you are using 6 Gbps SATA for the SSD you'd need at least 5 10K hard drives to match the transfer speeds.
And for compiling large sources, again the SSD wins because of how silly build tools are. As another commenter pointed out, current build software reads one include file, then the next one, then the next, etc. It reads one library file at a time. It reads or outputs one object file at a time. So it makes inefficient use of NCQ or RAID. Plus, some build tools seem to like to sync the build results to disk for no real reason (Visual Studio I am looking at you!).
Horse shit. My current laptop has Win 7 Pro installed on a 60 Gb hdd. And it's only about half full.
I've not got to play with SSD drive yet - what's the weight comparison to an HDD? Do they weigh more than the extra three cells?
I'm mostly going by what dell tells me I need to provision in a power supply (roughly 500 more watts needed by adding 1TB of LV RDIMM to an R910) and Google searches for wall-power consumption, which seem to be in the ballpark of 5-10W (average, not peak) added per DIMM. We're talking a few hundred more watts to power and cool.
I'm not sure how to square that with the tech doc you posted, is that actually the sort of chip you could build into LRDIMMs and attach 1,000 of to a system?
If the system can sleep most but not all of the RAM without sleeping the computer this would draw a lot less power but it does not look like this is a configuration that current computers actually do.
Ive not seen another upgrade that has affected the performance of a computer so dramatically. Maybe upgrading a computer from a woefully low amount of RAM would compare to the performance gain of a SSD. For me, it's almost like getting an entirely different laptop. Everything is more responsive, my laptop boots much quicker, and the laptop "feels" faster overall. It was so much of a difference that I couldn't go back to the spindle drive that the laptop came with. I felt like a step backwards even though I had ten times the space of the SSD. I eventually put the SSD back in and ditched the internal optical drive.
Windows is pretty aggressive about tracking the reads executables always perform during process launch and prefetching them. It works pretty well. It also tries to preload data into ram with a bunch of weird user-prediction heuristics that sometimes work well and sometimes just make your system flush it's read cache for no reason to read strange things off your disk.
Agreed about the database libraries though, synchronous-only is no way to perform anything dominated by latency like that.
Yes... Windows can do some pretty quick boots these days. They effectively log off the user and hibernate the kernel. When you reboot, there is a big sequential read of the hibernate file and the rest is CPU and non-drive IO. A full reboot only needs to happen when something in the kernel is changed, like patches and certain drivers.
It's amazing that the parent comment has been moderated a troll. It only repeats what an Intel engineer said in a meeting.
There are lots of reasons to choose an SSD over an HDD. I really jostle my laptop quite a bit. Performance would degrade with an HDD when I would shake my leg but that never happens with an SSD. My SSD runs cooler, requires less battery, doesn't vibrate, is thinner and lighter and all around better than HDDs in every way. Speed and size are not the only factors here.
WTF planet are they shopping on? Unless they're comparing a 500GB spinning drive to a 512GB SSD, that's just utterly wrong. Regardless, it comes down to if the customer needs the space or not. Out of my 400 or so past customers, around 10 of them have filled over 70GB on their hard drive. That means the other 390 are good to go with an $80 OCZ Vertex 4 (they were on sale when I bought them, lol). So this 3x the price bullshit is ridiculous. I can't find a price on a 120GB spinning SATA drive since they basically don't exist anymore but if you don't need the space, a 320GB HDD is $80+ so there you go. All of the last 6 PCs I built for customers had SSDs in them. They're all around $500-675 retail and with mostly pentium sandy bridge chips, they feel so much faster than my $1000 gaming rig at home, it's not even fair. For almost all the builds, it was slightly more expensive to pick up a decent 500GB spinning disk than it was to get a blazing SSD.
If youre just looking to upgrade your gaming rig or even your own personal machine, why not? I mean does it really have to be that thought-out for these instances? It's obvious SSD has been created by hopefully intelligent engineers in fortune 500 tech companies, and like everything else tech, their job is to progress to bigger and better things (mostly while reducing/controlling heat). Fuck yeah I'm gonna get one if I have the money. For work that is just a different matter. Research and show the numbers to whoever, but for my own personal box, there are better things to worry about. I'm not building a huge infrastructure. If you are, by all means research, it's your money.
Agreed. A thousand times agreed.
...or by connecting a USB drive, or loading a network share, or waiting for a CD to spin up, or for a friggin' javascript function to execute...
I can't tell you how many times a day I have to stop myself from swearing up a blue streak because the entire UI of an application (or the ENTIRE BLOODY OPERATING SYSTEM) has been locked up by a file read/write.
How is it that we've had the personal computer for decades and are just beginning to realize that the UI needs to be threaded completely separately from background processes?
It is a software problem and I fear it will only continue to escalate: hardware gets faster, programmers get proportionately lazier. How long is going to be before software becomes such an incredible kludge that even SSDs get bogged down?
Who boots their computer everyday? Most of the people who I know boot once per month - if that.
People whose work computer is a laptop and who take it home every night. On these rigs sleep & suspend tend to take as long as a full shutdown and restart, so there's no benefit to leaving it on during transport. Almost everyone I work with does a full shutdown when they leave at night. Argue all you want that it's a lousy setup, and I'll agree with you; on the other hand, corporate computing environments are like that.
I could have made a comparison between various power modes and HDD vs SSD, but honestly it wouldn't have been worth it. I don't work in an IT shop, and my managers barely even realize that there's a difference between shutdown and sleep mode. I made the best business case I could, we all got the hardware I wanted, and everyone has more time to do their work now. Mileage varies on how that translates to extra productivity =)
"Space Exploration is not endless circles in low earth orbit." -Buzz Aldrin
I'm awaiting delivery of a new ThinkPad, and I expect that the second upgrade (after a 4 GB SODIMM) is going to be a mSATA SSD. It's a trade-off in that I can't do internal WWAN, but I wasn't going to be doing that anyway.
With a small one I could just configure it as a cache drive, but with a larger one I'll likely end up using that for the KUbuntu install and just leave Windows and storage on the HDD.
fencepost
just a little off
They weigh less than a mechanical drive, whereas swapping from a six to nine cell battery is a weight addition. To be fair, 3 additional cells would probably buy you more time than an SSD change, depending on model of drive and how you use the machine. I don't have hard stats on that though.
Personally, I'd always do an SSD change before upgrading a working 6 cell battery to a 9, though. Batteries are expensive. With an SSD you might lose some storage space over the stock mechanical drive (per $), but you gain speed, run time, don't have to think about jostling the laptop around so much, you'll probably spend less out-of-pocket over a new battery, and you shave off a bit of weight. I've done it for workplace machines and everyone seems really pleased.
They make a lot of sense in laptops as the price comes down, which is why we're seeing them in the slimmer, lighter, faster, longer running laptops. Of course it makes even more sense if the machine ships that way and you don't have to replace a working part... though a laptop drive has some utility of its own after it has been replaced.
I've not got to play with SSD drive yet - what's the weight comparison to an HDD? Do they weigh more than the extra three cells?
2.5" SSDs and HDDs both weigh around 100 grams, with the SSD probably being 10-20g lighter. A desktop 3.5" HDD however weighs around 700 grams, but there the weight is usually not a concern.
Hey, it wouldn't be Slashdot without using a some extreme counterexample to try to refute the argument.
Nope, the problem is that software is and was designed as if the hard drives were SSDs, but they weren't. So SSDs do solve a software architecture problem, and a prevalent one. That problem isn't limited to accessing storage, though, and once you solve it, SSDs are not a must have on the run-of-the-mill desktop anymore.
A successful API design takes a mixture of software design and pedagogy.
The programming model doesn't have to be very complex; besides, the basic GUI paradigms call for this very event driven paradigm anyway, it's just that storage is not normally deal with that way. It's not even about what can execute in parallel -- hard drives are serial random access devices, you're not parallelizing anything, just reordering and coalescing serial accesses. Just look at what any run-of-the-mill application does when you fire it up. First it opens and then reads/mmaps a whole bunch of files. That's just to get the executable ready to run. Then it opens and reads even more files -- that's the configuration, plugins, extensions, etc. It'll also perhaps access some mapped out registry pages, perhaps wake a few partially paged out other processes it may interact with, etc. The deal is: a lot of those things do not have to happen in a fixed order, they can be reordered (not parallelized!) by the OS, there's a fixed set of join points where results of the previous operations have to be ready to proceed. Such join points can be made more granular if the developer so wishes (before diminishing returns kick into gear). Parallelizing things is hard, because suddenly there's a whole class of problems that don't exist in serial execution. What I talk about is merely reordering serially executed requests, that's all -- well, first you need to have some requests instead of just one.
A successful API design takes a mixture of software design and pedagogy.
The problem is that with Windows, the buck stops at the executable stuff, and even that doesn't encompass plugins and extensions (just look at startup of any modular software package like Acrobat, Eclipse, LibreOffice, Qt Creator).
A successful API design takes a mixture of software design and pedagogy.
Async IO wouldn't add complexity if the programming languages were not stuck 4 decades in the past when it comes to dealing with event-response style of processing. That's precisely why I sneer, it's the inverse network effect. Nobody uses them because they are a pain, thus we all suffer spinning rings/beachballs where the OS is loading stuff page-by-page simply because it doesn't know upfront what pages a memmapped chunk of data or executable will need next.
I do agree about seeks having a minimum duration, but the most glaring problem is when seeks are done to read single pages scattered around an area that could be read in the entirety in as long as it takes to do one or two of those seeks. It's my experience that such seek storms happen when a long disused application has been slowly paged out to disk and now there's hardly anything left of it save for popular shared library pages. The problem is so bad that on both Windows 7 and OS X when this happens, killing and restarting the application is actually faster than waiting for it to be paged back in when there's a seek overhead interspersed between each pagein. In a bad case you end up waiting for 30 seconds or more as the OS laboriously pulls in maybe 100MB of data at a rate that was state of the art on PCs back in the 90s. There's no mechanism AFAIK (not in Windows 7, not in OS X) to mark them up at runtime as to which page usually brings up what other page -- here we need both code and data pages, and that information is transient, there's no way to prelink anything. An application could provide self-generated hints both for how it does explicit file access but also access to data, if that was shown to be somehow better than the OS collecting the data at runtime.
As for the buffer cache: duh, I was merely saying in what high transaction rate scenario a database wouldn't necessarily benefit all that much from SSDs.
A successful API design takes a mixture of software design and pedagogy.
The R910's memory access architecture is not designed for hard-drive-like storage. It's hugely wasteful, power-wise, because it interfaces all that memory with a CPU that's reading the memory in cacheline chunks. When you treat RAM as a hard drive, you're reading it in larger chunks, and you can design things to isolate individual chips so that large buses that span dozens of chips don't have to dissipate power due to switching. Not only that, but wiring on regular DIMMs might not be power optimal for that use either, because the bus is shared among so many chips. Low power memory architecture likely wouldn't work with off-the-shelf DIMMs, you need an isolation register in front of every couple SDRAM chips -- certainly less than what fits on a DIMM stick.
The provisioning figures must be conservative and worst case. Feel free to compare the power consumption between sequential read of terabyte of RAM vs. random accesses where each cacheline read hits different row or page. The precharge energy and burst penalties alone will eat into your power budget like no tomorrow, from what I can figure out from the datasheet. I do use DDR3 in a design but it's vastly underutilized and only there because it's cheaper and easier to get than DDR2, so I'm far away from peak power anyway.
A successful API design takes a mixture of software design and pedagogy.
Waitaminute, you scoff at a mere factor of 8 difference for a memory system that gives you two orders of magnitude higher bandwidth and 4+ orders of magnitude lower latency for that 5W delta from idle to active power? If you need performance, throwing hard drives into a RAID system to get higher IOPS and bandwidth will overshoot the DDR3L SDRAM's power consumption when you've got 6 hard drives, and you're still nowhere near the performance you get from SDRAM, even if we're talking only about the baseline performance you get when you allocate 5W to the active use (over refresh). If you want to burst as much power into those RAM chips as they'll take, you can be extracting performance that will match tens of kilowatts worth of hard drive power consumption.
A successful API design takes a mixture of software design and pedagogy.
Of course it's not meant for a laptop duh ;)
A successful API design takes a mixture of software design and pedagogy.
I wouldn't call 23 cents per GB near price parity with 4 cents per GB...
> twice the performance
This comparison misses the point about SSDs. Yes, SSDs may have somewhat better bandwidth, and may improve startup times slightly, but that is not their advantage. They have awesomely better seek times, which makes some operations hundreds of times faster. Putting Visual Studio's .sdf files on an SSD avoids lots of VS 2010 hangs.
This blog post I wrote discusses the random I/Os to the Windows Live Photo Gallery SQL database at startup. On my photo collection I see 5,000 random disk I/Os, which are painful on a laptop HDD but would be a non-issue on an SSD:
http://randomascii.wordpress.com/2012/08/19/fixing-another-photo-gallery-performance-bug/
In situations like this an SSD is probably a *hundred* times faster than an HDD. Database accesses seem to be a common scenario where an SSD is worth its weight in gold.
In short, if an SSD is only twice as fast then it's not worthwhile. If it's ten to a hundred times faster, then hell ya.
And you cannot RAID the SSD's, as I said, not using SRT mode.
OK, I get it now. I was assuming we were talking about using /dev/shm to store bulk data in system ram, not constructing an SSD out of SDRAM instead of flash.
What do you use for an interface on something like that? Seems like SATA/SAS like most of the PCIe flash devices I can find would be a bottleneck.
Well, the interface would be on an FPGA, so probably PCI express would be the simplest things to implement. Do note that power scales with bandwidth, so you want only as much bandwidth as you need and no more. The win is that I/O transaction latency is negligible compared even to an SSD -- you don't waste power waiting for things to happen.
A successful API design takes a mixture of software design and pedagogy.
My relatives have wireless broadband at 3mb/2mb down/up for $65/month. Satellite was far more expensive, had monthly caps, and super high latency. There really aren't any great options out there, but wireless isn't bad. Good enough to switch to VoIP and pay for itself at least :)
SWM seeks new sig for a brief fling
The whole point of SRT is to get fast read times, and optionally fast write times if you want to risk your data. It also eliminates the need to actively manage which drives your data is on (as opposed to you putting certain programs on an SSD manually) as it will actively change the cache data depending on your usage pattern. It's works incredibly well for it's intended purpose. In my own testing I could not tell the difference between day to day use on pure SSD vs SRT. It's easy to see if you benchmark it, but boot times and app launch times are essentially the same.
Look at it this way, if you are putting your windows OS on an SSD, why do you care if your KB items uninstalls are accessed quickly? Do you really care that some DLL that never gets accessed in your system32 directory is speedy? All that garbage can sit nicely untouched on your spinning disk while the stuff that you use all the time is fast.
It's the same kind of theory behind why defrag programs for spinning disks like Ultimate Defrag work well since it keeps the stuff you want accessed quickly, and the stuff you don't care about at normal speed.
Rather the question is, do you trust the Hybrid drives not to lose any data in power failure situations?
es? ... most be the poltergeist showing his discontent or something. SSD's dont have heads, right? i'm not really supertechnic man. For that alone i'd get one or two, just to put the os on. I could care less about startup speeds really.
i had like three drives crash on me in two months
Free speech was meant to be free for all... how can anyone grow up in a nanny state ?
SRT does NOT support RAID. That's all I said. I was repeating what an Intel engineer told me personally.
If SRT supported RAID 1, it could be both faster and more secure.
RAID issues at Intel are badly managed, in my opinion, so badly managed that it demonstrates that the Intel CEO has little understanding of technology. (Intel makes RAID adapter cards, also.)
Wow! I am sure your "solution" is so simple you can finish implementing it tonight. Congratulations, Mr trillionaire.
Bingo Dictionary - Pragmatist, n. A myopic idealist.
The "solution" is not very simple. Your assumption that just because I notice something I need to be the one implementing it is just, um, silly. You need a paradigm shift in both programming language design and OS design. That doesn't mean that there isn't a problem. Your shout is a rather typical fallacy.
A successful API design takes a mixture of software design and pedagogy.
Your fallacy is that people who can implement it and could have implemented it have not noticed "it".
Bingo Dictionary - Pragmatist, n. A myopic idealist.
I'm not saying what you say I do. Some people have noticed it, some didn't. There's a lot of variability in how good software developers are at noticing stuff that's not really taught or common industry practice or whatever. Heck, most desktop software people probably never learned the IEC 61131 programming languages used in PLCs, and even those have some nice concepts that make it easier to produce reliable software that reacts to events. Most desktop software people probably never learned about functional safety aspects of programmable electronic devices, subject of IEC 61508, even though said standard lays out the engineering in 'software engineering' (as opposed to hacking stuff together). And so on I could go.
A successful API design takes a mixture of software design and pedagogy.
You were more ranting against programming languages. Most software developers do not create programming languages. So your above post is irrelevant to your original point against programming languages.
Idea is worth shit if implementing it is not practically possible. Which describes your "idea" very well. The fact that it is not widely used doesn't imply no one else had that idea before, like you are suggesting. It means that they also understood that implementing this "idea" in a widely used manner will be impossible. Exactly because of points in your above rant post against software developers.
Thinking about asynchronous constructs is hard and error prone. Large proportion of today's software is still not free of basic bugs. Your suggestions of complicating programming languages further is total shit at this point.
Bingo Dictionary - Pragmatist, n. A myopic idealist.