Sorry been busy so haven't gotten back to this thread till now. My problem is still the same. I mean sure you can get an 60gb SSD for $110 and the one you listed looks nice but... it's got like what? a 250mb sustained sequental read? and 170mb write?
Why bother with that when I could simply do a raid 0 or raid 1 setup with HDD's? I mean hell 4 Samsung spinpoint f4's would give me over 1 TB of storage and give me around 640mb read and transfer rates.
I mean hell the $175 107gb one you linked only gives you 280mb read and 270mb write. If your worried about the Raid 0 setup go with a raid 1 and you still end up with 320mb read and still over 600gb of storage and your datas safe for that same $175 lol.
The value isn't there yet for SSD's lol. It wasn't an opinion. For the same money or less you can outperfom the SSD's WHILE keeping nice storage capacity. On a budget of $1500 that money that could be spent on an SSD is better spent on other more worth while components.
I mean sure SSD's are the future, but as I said before they simply aren't the present. The value just isn't there especially when I can get better results for less. When a single SSD costs more than a raid o or raid 1 setup using 4 HDD's and still can't out perfom them well... then you have a problem.
Here's what you're not catching. You're focusing on sequential transfer speeds. Sequential transfer speeds don't matter. That's a slight exaggeration, but only a slight one. Everything is fast at sequential transfer speeds, except for external USB 2.0 hard drives, USB flash drives, and some other things that don't get used much.
The hard drive bottleneck is small random read and write speeds, or equivalently, IOPS. As modern SSDs go, the Samsung 470 is all right. It's not blazing fast, but it's not terribly slow, either, and it does perform like a good SSD, as opposed to the st-st-stuttering JMicron junk or some of the other early SSDs. In a 4 KB random read test with queue depth 1 (e.g., Crystal Disk Mark does this), the Samsung 470 will offer about 20 MB/s. Unimpressive, right?
Consider that a relatively fast 7200 RPM consumer hard drive will struggle to each 0.5 MB/s in the same test. Note the decimal point. That's half of a MB/s.
So you say, RAID 0 will fix that? Fine then, put eight 7200 RPM hard drives in a RAID 0 array. And then in the same test, you'll still get about 0.5 MB/s. Now, at higher queue depths, you'll see some benefit from the RAID array. But the SSD will also go a lot faster at higher queue depths, too.
And it's not just a problem that two hard drives aren't enough. Make it eight hard drives, and make them 15K RPM enterprise hard drives, and small file IOPS performance still won't be in the same league as the better SSDs, even at very high queue depths.
Look at the random read and write tests that I linked above, for example. Those graphs show a WD VelociRaptor, in addition to a bunch of SSDs. And the VelociRaptor gets completely destroyed. In one of the charts, they don't even bother to list it, because its bar length would round to 0 pixels long. Make it a bunch of VelociRaptors in RAID 0 and maybe you could get the bar to at least register. But it still wouldn't be competitive.
So why does this matter? When does your hard drive make you wait? For loading one large file, it's fast. For loading hundreds of small files, it's slow. There are a number of situations where you have to read or write large numbers of small files at once, such as when loading programs, zoning in many games, or opening several web pages at once in a browser.
But it's not just the "now you only have to wait 1 second instead of 3 seconds" effect. Any little operation where the computer needs to read or write a file and then do something else will benefit from an SSD, as it will get to the something else" a little faster. For some programs, it's basically an offline equivalent of taking 100 ms off of your ping time in an online game. Or perhaps switching from a software cursor to a hardware cursor. You can't time it with a stopwatch, but the feel is clearly there.
When I use my old desktop or my office computer at work, I'm constantly like, hey, what's wrong with this, it's slow. I don't have that same problem with my netbook. And my netbook has a much slower processor (Zacate E-350 versus desktop Core 2 Duo in the two desktops). The difference? My netbook has an SSD. The difference that makes is pretty glaring.
You keep forgetting that we are talking actual value here. As of now I can play SC2 on max settings, while watching a movie, while opening a browser and multiple tabs and surfing and every last browsers/tab is poping up instantaniously or close enough that I can't tell the difference.
The noticable performance gain I could get from an SSD I could get for the same price (If not less) with a Raid setup of HDD's on top of that I could do so with a great deal more storage space. Next my HDD's don't have as short of a life span. That SSD has a set number of writes and the lower the st
As the poster that replied after my initial reply stated, if your noticing that much of a performance increase by swaping to an SSD your rig has other issues.
SSD tech is still being fleshed out, many of it's problems are being worked on. I don't argue that SSDs are the future, but they aren't worth it at this moment in time. It's like the Ipod, Ipad and Iphone. Some people will swear by it even though it adds nothing meaningful or actually useful that can't be found cheaper elsewhere.
Except that you can't get the same performance boost from a RAID 0 setup. You can't get much of a performance advantage at all from a RAID 0 setup. If the original poster can tell the difference between a WD Caviar Blue and a WD Caviar Black, and it's highly plausible that he can, then there was a definite hard drive bottleneck with the Caviar Blue. Do you think that replacing it by something 30% faster is enough to completely eliminate the bottleneck, or only ease up on it somewhat?
As for SSD lifespan, it's likely that they'll tend to last longer than hard drives, simply because there aren't mechanical parts that can wear out. Silicon tends to be pretty resilient, and SSDs don't put out enough heat to have meaningful problems with overheating or thermal cycling.
That said, we don't really know how long SSDs will tend to last. The only real way to find out how long they'll last under real world circumstances is to make a bunch, use them, and see. The first good SSDs didn't arrive until September 2008 (Intel X25-M first generation), and TRIM support didn't show up until October 2009 (Indilinx Barefoot), which really marks the start of the modern era for SSDs. If there's some weird defect that will make SSDs tend to mysteriously die sometime in the 3-5 year time frame, we wouldn't know about it yet. I don't think that's likely, but we don't really know for certain.
You bring up the write cycle endurance stuff, but that's irrelevant to typical consumer use. If you write to an SSD really aggressively, you can wear it out more quickly. This can be a problem for servers that want to run a database on SSDs and write to them aggressively all day every day. But for typical consumer use, you'd be looking at centuries before you use up all of the write cycles. Most likely, it will be dead of other causes before then. But we know pretty well that hard drives aren't going to last centuries, either.
-----
If you've ever played League of Legends, when you sit at the loading screen, it's glaringly obvious who has an SSD and who doesn't. The results there aren't atypical. Rather, I bring up that example because it lets you directly compare loading performance on ten machines at a time, and see how quickly everyone else is loading.
Except that you can't get the same performance boost from a RAID 0 setup. You can't get much of a performance advantage at all from a RAID 0 setup. If the original poster can tell the difference between a WD Caviar Blue and a WD Caviar Black, and it's highly plausible that he can, then there was a definite hard drive bottleneck with the Caviar Blue. Do you think that replacing it by something 30% faster is enough to completely eliminate the bottleneck, or only ease up on it somewhat?
As for SSD lifespan, it's likely that they'll tend to last longer than hard drives, simply because there aren't mechanical parts that can wear out. Silicon tends to be pretty resilient, and SSDs don't put out enough heat to have meaningful problems with overheating or thermal cycling.
That said, we don't really know how long SSDs will tend to last. The only real way to find out how long they'll last under real world circumstances is to make a bunch, use them, and see. The first good SSDs didn't arrive until September 2008 (Intel X25-M first generation), and TRIM support didn't show up until October 2009 (Indilinx Barefoot), which really marks the start of the modern era for SSDs. If there's some weird defect that will make SSDs tend to mysteriously die sometime in the 3-5 year time frame, we wouldn't know about it yet. I don't think that's likely, but we don't really know for certain.
You bring up the write cycle endurance stuff, but that's irrelevant to typical consumer use. If you write to an SSD really aggressively, you can wear it out more quickly. This can be a problem for servers that want to run a database on SSDs and write to them aggressively all day every day. But for typical consumer use, you'd be looking at centuries before you use up all of the write cycles. Most likely, it will be dead of other causes before then. But we know pretty well that hard drives aren't going to last centuries, either.
-----
If you've ever played League of Legends, when you sit at the loading screen, it's glaringly obvious who has an SSD and who doesn't. The results there aren't atypical. Rather, I bring up that example because it lets you directly compare loading performance on ten machines at a time, and see how quickly everyone else is loading.
In the end this is what it boils down to. SSD + HDD or a Raid 0 setup gives you faster load times but nothing else of any note worthyness. It doesn't effect game play or game performance, and no with my wife playing leauge of legends with me and having a Vertex SSD in her comp there is no glaringly obvious difference. Even with 3D modeling, rendering, etc, there simply is nothing that even remotely screams or even wispers "Hey I'm worth the money".
I have to say this though, random writes SSD is better, Sequential writes a raid setup is better. Random writes are used more often, but... the performance boost in random writes isn't as noticable as the performance boost in the sequential writes. Some one speeds up something that already doesn't take much time it's not a big accomplishment where as speeding up something that actually warrants being sped up is something that adds some value to me.
In the end we will simply have to agree to disagree. You find loading windows up a few seconds faster and not gaining much more than that worth another $100-$200 bucks, I don't especially when I can get close to the loading speed you gain for less and on top of that gain substantial storage space all the while I can get far greater sequential write/read speed which is something that is very noticable. I mean recording videos, copying large files and moving them around, rendering tends to fall under sequential and not random and these are things that are important to me. Having a gain there would be far more valuable rather than shaving 10ms off of my random writes and reads lol.
Again I mean were talking a marginal increase, SSDs are good for random access which isn't slow to begin with while a decent raid setup will actually improve a problem area the sequential read and writes. Something actually worth speeding up.
Until SSDs see a performance increase (especially on sequential) and or a significant drop in price no one with any sense could honestly say they were worth it. It's a meaningless fluff item considering how little return on your investment there is.
But in any case, I'm out. Good discussion. After this discussion though I'd be willing to bet you have an iphone or an ipad lol.
Originally posted by Eerazer Originally posted by Quizzical You'll probably have to consult your motherboard manual. BIOSes can be weird, and every vendor does theirs differently. You've found the boot priority list in the BIOS, and the WD Caviar Black isn't an option? What exactly does it list? Mine lists my SSD, my optical drive, and a USB flash drive, and gives me three options as to what order I want the computer to try to boot from them. Have you copied everything onto the new hard drive yet? If so, I'd try unplugging the old hard drive and seeing what happens. That won't damage anything; at worst, the system won't boot until you plug it back in.
It just isnt in the list, like i said the only hdd that shows up is my c blue.
even though my black now has windows on it as a bootable drive its not even in the list, but i check my connected sata drives and its right there. along with the blue.
so thats what i dont get. the drive works fine, but its not bootable, 'apparently' - possibly there is a conflict between the 2 drives, im not sure, or maybe by blue has become priority in the boot menu because thats what my bios is used to.*
*something tells me my last sentence is nonsense though, i used to work as a computer technitian for 2 years, but still i think you know more than me about this issue quiz, back when i worked i was used to working with ide drives and windows 95/98/mil/server/xp/vista. dont have much experience with raid or sata, but i know how it works. roughly. what else.. erm there is my black using the 6gb option in the bios, *run all hd's at 6gb's if available*
not sure if that would cause any problems as my blue isnt 6gb/s. - even that too is not the issue i think. hmm. with that said i am lost really, i suppose i could try unplugging the other sata drive, to see what happens. in relation to that though - after i installed windows on the new drive i did take the blue off from the boot list to see if it would auto boot from the new drive, and it said invalid system disk
which makes sense. hmm. the only way to find out i think is to actually unplug the blue and see what happens.
You made the mistake of installing Windows with the Caviar Blue still plugged in
The fix can get fairly complicated. What I would try doing: Unplug the Caviar Blue. Your computer probably won't boot up once you do that (which is ok). With the Caviar Blue unplugged, reboot off the Windows installation DVD, and go to the Repair options. It should give you the option to Repair and Windows Installation. Do this, with the Caviar Blue unplugged. Once that is done, that should allow you to boot up without the Caviar Blue plugged in.
Now the crap shoot is, when you plug the Caviar Blue back in: Hopefully it will keep right on trucking properly from the Repair fix you just did, but there is a chance it could go right back to the dual-boot menu option; in which case you have to do some MBR magic that isn't easily explained to really fix it. It's also possible that you can go into the BIOS and just change the boot order: set the Caviar Black to boot over the Caviar Blue as the primary drive. But you need to do the WIndows Repair from the install DVD first before this will do anything.
Originally posted by Quizzical Except that you can't get the same performance boost from a RAID 0 setup. You can't get much of a performance advantage at all from a RAID 0 setup. If the original poster can tell the difference between a WD Caviar Blue and a WD Caviar Black, and it's highly plausible that he can, then there was a definite hard drive bottleneck with the Caviar Blue. Do you think that replacing it by something 30% faster is enough to completely eliminate the bottleneck, or only ease up on it somewhat? As for SSD lifespan, it's likely that they'll tend to last longer than hard drives, simply because there aren't mechanical parts that can wear out. Silicon tends to be pretty resilient, and SSDs don't put out enough heat to have meaningful problems with overheating or thermal cycling. That said, we don't really know how long SSDs will tend to last. The only real way to find out how long they'll last under real world circumstances is to make a bunch, use them, and see. The first good SSDs didn't arrive until September 2008 (Intel X25-M first generation), and TRIM support didn't show up until October 2009 (Indilinx Barefoot), which really marks the start of the modern era for SSDs. If there's some weird defect that will make SSDs tend to mysteriously die sometime in the 3-5 year time frame, we wouldn't know about it yet. I don't think that's likely, but we don't really know for certain. You bring up the write cycle endurance stuff, but that's irrelevant to typical consumer use. If you write to an SSD really aggressively, you can wear it out more quickly. This can be a problem for servers that want to run a database on SSDs and write to them aggressively all day every day. But for typical consumer use, you'd be looking at centuries before you use up all of the write cycles. Most likely, it will be dead of other causes before then. But we know pretty well that hard drives aren't going to last centuries, either. ----- If you've ever played League of Legends, when you sit at the loading screen, it's glaringly obvious who has an SSD and who doesn't. The results there aren't atypical. Rather, I bring up that example because it lets you directly compare loading performance on ten machines at a time, and see how quickly everyone else is loading.
In the end this is what it boils down to. SSD + HDD or a Raid 0 setup gives you faster load times but nothing else of any note worthyness. It doesn't effect game play or game performance, and no with my wife playing leauge of legends with me and having a Vertex SSD in her comp there is no glaringly obvious difference. Even with 3D modeling, rendering, etc, there simply is nothing that even remotely screams or even wispers "Hey I'm worth the money". I have to say this though, random writes SSD is better, Sequential writes a raid setup is better. Random writes are used more often, but... the performance boost in random writes isn't as noticable as the performance boost in the sequential writes. Some one speeds up something that already doesn't take much time it's not a big accomplishment where as speeding up something that actually warrants being sped up is something that adds some value to me.
In the end we will simply have to agree to disagree. You find loading windows up a few seconds faster and not gaining much more than that worth another $100-$200 bucks, I don't especially when I can get close to the loading speed you gain for less and on top of that gain substantial storage space all the while I can get far greater sequential write/read speed which is something that is very noticable. I mean recording videos, copying large files and moving them around, rendering tends to fall under sequential and not random and these are things that are important to me. Having a gain there would be far more valuable rather than shaving 10ms off of my random writes and reads lol.
Again I mean were talking a marginal increase, SSDs are good for random access which isn't slow to begin with while a decent raid setup will actually improve a problem area the sequential read and writes. Something actually worth speeding up. Until SSDs see a performance increase (especially on sequential) and or a significant drop in price no one with any sense could honestly say they were worth it. It's a meaningless fluff item considering how little return on your investment there is.
But in any case, I'm out. Good discussion. After this discussion though I'd be willing to bet you have an iphone or an ipad lol.
Since your out, it's time for me to put my 2c in.
If you haven't actually used a computer with a good SSD, you have no idea. The difference is night and day. It's much more than just "faster boot speed".
In the end of the day, faster writes only matter when you are doing something really file intensive (such as FRAPS). It isn't very often that we need to write large files and are restricted by the media speed: crap only streams off the internet so fast, optical media installs are limited by the DVD drive and not the hard drive, and not many people run large online relational databases that need to update large datasets frequently. And saving your Word documents, well, most people don't try to save 500Mb Word documents that often.
Most people, on their computers, read small files frequently more than any other operation. And this is what most of the bottleneck is on everyday computing use: small file random reads. ~This~ is your game load times. Take a look at your typical game: it's made up of thousands of very small files. It's usually not one big 5Gb file that has to get loaded (where superior sequential read transfer speeds would be at an advantage). Each time you request a new file, you incur the seek time penalty. Then you get the read transfer speed for the duration of the file. If you have many, many small files, that's a lot of seek times to add in there. (And even Blizzard's big .mpq files really are just archives of smaller files, the game requests them individually inside the archive, and those get hit with seek penalties as well).
A RAID array setup does absolutely nothing for random access. Your still limited by the seek time of the drive: RAID does nothing for seek times, it just provides parallel throughput. Sure, you get better read and write times, because your able to do that across multiple platters at the same time. But guess what, an SSD is in effect a massive RAID 0 device in and of itself! The SSD firmware treats each RAM chip as a parallel data path (they call them "channels"). Plus you get the benefit of near-instantaneous seek times (there is no platter to spin, and no head to reposition, all sectors are available with equal fast speeds)
Sure, you can argue TRIM and performance degredation and blah blah blah, but truth of the matter is, even in a traditional hard drive you still have fragmentation to content with that degrades performance, and even degraded, SSDs still perform very well.
You can look at numbers all day long and wonder why some people preach about SSDs... but until you've used a system with one for any period of time, there is much more to it than just a few numbers. As far as I'm concerned, platter drives are a dying technology to be delegated to cheap bulk storage in the near term, and a dying breed doomed to the same obsolence as the floppy drive in the long term.
"It doesn't effect game play or game performance,"
Usually an SSD doesn't help your frame rates in games. There are some occasional examples of games where it will get rid of stuttering. Making loading screens go faster is quite nice in some games, and unimportant in others.
"I have to say this though, random writes SSD is better, Sequential writes a raid setup is better."
Reads are more commonly the bottleneck, not writes, simply because you do a lot more reading than writing. But you could substitute read for write and the statement is still more or less true. In some cases, it would take more than two hard drives in RAID 0, but you can add enough drives and make it true.
"Random writes are used more often, but... the performance boost in random writes isn't as noticable as the performance boost in the sequential writes."
You must have an extremely unusual workload in order for a hard drive to be meaningfully bottlenecked in sequential transfer speeds, but not in random transfer speeds. In particular, that would mean that you're not using an awful lot of real programs at all. One can easily come up with synthetic applications that are like that (e.g., I'm going to copy this 4 GB file from this drive to that one), but those aren't the sort of thing that a typical user would commonly do. I guess there are probably industrial applications where you have to record some incoming stream of purely sequential data very quickly, and then it will be analyzed elsewhere. But that's the best I can come up with to fit what you describe.
The other way isn't hard. There are plenty of programs that are meaningfully constrained on a hard drive by small file IOPS, but not sequential transfer speeds. For example, operating systems, web browsers, e-mail clients, and most games. That's the sort of thing that a lot of everyday consumers use.
Now, if you're comparing one good SSD to another, then there's more of a case for that, because they're all so fast in IOPS, too. The reason why I say random reads and writes matter so much is that that's the bottleneck for hard drives.
Suppose that you're looking to buy a video card, and have your choice of card A or card B. They both have the same price tag, and equivalent feature sets, power consumption, cooling systems, warranties, and so forth. The only difference is performance. There are only two games that you'll ever want to play on the cards. In one game, card A offers 300 frames per second at the settings you want, and card B offers 200 frames per second. In the other game, card A offers 10 frames per second at the settings you want, and card B offers 60 frames per second. Which card do you pick?
This isn't a trick question. Card B lets you play both games well, while card A only lets you play one of the games. The better choice is card B. That card A wins in average frame rates, 155 to 130, doesn't matter. That card A is faster at the first game doesn't matter. That difference isn't meaningful. The difference that is meaningful is the second game, and that's where card B wins.
Now, that sort of situation is unusual for video cards. Usually if one card is clearly faster than another card in one game, it will at worst be kind of competitive in other games, with only a handful of buggy outliers that will probably be fixed in future drivers. But in storage, a situation like that is ubiquitous. A RAID 0 setup might well beat an SSD in sequential transfer speeds, but they're both so fast that the difference will never particularly matter to most consumers. How often do you really write 1 GB of sequential data at a time? The only times that I can think of that I've ever done that is installing an OS and making an image backup of my SSD, both of which are bottlenecked by other things (optical drive and USB 2.0 connection, respectively), so even then, the sequential transfer speed of an SSD or hard drive still doesn't matter.
"Having a gain there would be far more valuable rather than shaving 10ms off of my random writes and reads lol. "
It's not just 10 ms, once. If that's all it were, then it would be no big deal. It's 10 ms per access. Multiply that by hundreds of things that you have to load at once and it adds up to several seconds, easily, when loading a program. On my old computer, it took about a minute to launch Europa Universalis II because the game had to load thousands of small files. Or even multiply it by dozens of things for loading a typical web page and it's a considerable fraction of a second.
It's kind of like how 10 ns (that's nanoseconds, not ms) wouldn't be a big deal if it were just once, but if you increased the L1 and L2 cache access times by 10 ns for a typical processor, you'd completely cripple it and it would be painfully slow. It doesn't just access them once, but does so many times, and that adds up.
-----
"After this discussion though I'd be willing to bet you have an iphone or an ipad lol. "
All right, so that's kind of a random response. But you get the point.
Would you believe that I don't have a cell phone or a tablet of any sort, and got my first laptop earlier this year?
I'm not one to jump on every little tech gadget that comes along. Among the things that I'm unimpressed by are Blu-Ray, "widescreen" monitors, too many buttons on a mouse, headphones, vent or equivalent programs as used in games, power supplies of ridiculously high wattages, tablets, "smart" phones, and just about any hardware with "gaming" in the name except for a gamepad. (Video cards don't put "gaming" in the name because it's implied.)
But SSDs are a big deal. So, incidentally, are multiple monitors, uninterruptible power supplies, and KVM switches. But I digress.
FRAPS is the only thing I can think of that most gamers would see a big benefit in sequential speeds over an SSD.
Anything else is like... active relational databasing, large scale video rendering/editing, realtime data acquisition.. not stuff most people do with their home PC.
And I totally agree: SSDs are probably the biggest deal to hit computers since the jump from punch paper to magnetic tape. Certainly a bigger deal than even multi-core CPUs or anything else we've seen in the last 20 years.
And UPSes and multiple monitors are all great too... KVM switches, I don't know about. I prefer VNC.
"It doesn't effect game play or game performance,"
Usually an SSD doesn't help your frame rates in games. There are some occasional examples of games where it will get rid of stuttering. Making loading screens go faster is quite nice in some games, and unimportant in others.
Funny thing: It did actually matter in EQ2, particularly in Neriak and a few other zones. I got about 20% bost in FPS in some areas, I never noticed any difference in another game... But I do know that Neriak is poretty poorly coded and that it loads a lot while you are in there, and the lag is worst in the game...
Zooning in Vanguard didn't get that "hickup" either when you moved to a new area but that is about it.
Everything loading is fast but that rarely affects FPS. It do makes games like AoC and EQ2 nicer to play since it cuts down the loading screen time a lot.
Comments
You keep forgetting that we are talking actual value here. As of now I can play SC2 on max settings, while watching a movie, while opening a browser and multiple tabs and surfing and every last browsers/tab is poping up instantaniously or close enough that I can't tell the difference.
The noticable performance gain I could get from an SSD I could get for the same price (If not less) with a Raid setup of HDD's on top of that I could do so with a great deal more storage space. Next my HDD's don't have as short of a life span. That SSD has a set number of writes and the lower the st
As the poster that replied after my initial reply stated, if your noticing that much of a performance increase by swaping to an SSD your rig has other issues.
SSD tech is still being fleshed out, many of it's problems are being worked on. I don't argue that SSDs are the future, but they aren't worth it at this moment in time. It's like the Ipod, Ipad and Iphone. Some people will swear by it even though it adds nothing meaningful or actually useful that can't be found cheaper elsewhere.
Except that you can't get the same performance boost from a RAID 0 setup. You can't get much of a performance advantage at all from a RAID 0 setup. If the original poster can tell the difference between a WD Caviar Blue and a WD Caviar Black, and it's highly plausible that he can, then there was a definite hard drive bottleneck with the Caviar Blue. Do you think that replacing it by something 30% faster is enough to completely eliminate the bottleneck, or only ease up on it somewhat?
As for SSD lifespan, it's likely that they'll tend to last longer than hard drives, simply because there aren't mechanical parts that can wear out. Silicon tends to be pretty resilient, and SSDs don't put out enough heat to have meaningful problems with overheating or thermal cycling.
That said, we don't really know how long SSDs will tend to last. The only real way to find out how long they'll last under real world circumstances is to make a bunch, use them, and see. The first good SSDs didn't arrive until September 2008 (Intel X25-M first generation), and TRIM support didn't show up until October 2009 (Indilinx Barefoot), which really marks the start of the modern era for SSDs. If there's some weird defect that will make SSDs tend to mysteriously die sometime in the 3-5 year time frame, we wouldn't know about it yet. I don't think that's likely, but we don't really know for certain.
You bring up the write cycle endurance stuff, but that's irrelevant to typical consumer use. If you write to an SSD really aggressively, you can wear it out more quickly. This can be a problem for servers that want to run a database on SSDs and write to them aggressively all day every day. But for typical consumer use, you'd be looking at centuries before you use up all of the write cycles. Most likely, it will be dead of other causes before then. But we know pretty well that hard drives aren't going to last centuries, either.
-----
If you've ever played League of Legends, when you sit at the loading screen, it's glaringly obvious who has an SSD and who doesn't. The results there aren't atypical. Rather, I bring up that example because it lets you directly compare loading performance on ten machines at a time, and see how quickly everyone else is loading.
In the end this is what it boils down to. SSD + HDD or a Raid 0 setup gives you faster load times but nothing else of any note worthyness. It doesn't effect game play or game performance, and no with my wife playing leauge of legends with me and having a Vertex SSD in her comp there is no glaringly obvious difference. Even with 3D modeling, rendering, etc, there simply is nothing that even remotely screams or even wispers "Hey I'm worth the money".
I have to say this though, random writes SSD is better, Sequential writes a raid setup is better. Random writes are used more often, but... the performance boost in random writes isn't as noticable as the performance boost in the sequential writes. Some one speeds up something that already doesn't take much time it's not a big accomplishment where as speeding up something that actually warrants being sped up is something that adds some value to me.
In the end we will simply have to agree to disagree. You find loading windows up a few seconds faster and not gaining much more than that worth another $100-$200 bucks, I don't especially when I can get close to the loading speed you gain for less and on top of that gain substantial storage space all the while I can get far greater sequential write/read speed which is something that is very noticable. I mean recording videos, copying large files and moving them around, rendering tends to fall under sequential and not random and these are things that are important to me. Having a gain there would be far more valuable rather than shaving 10ms off of my random writes and reads lol.
Again I mean were talking a marginal increase, SSDs are good for random access which isn't slow to begin with while a decent raid setup will actually improve a problem area the sequential read and writes. Something actually worth speeding up.
Until SSDs see a performance increase (especially on sequential) and or a significant drop in price no one with any sense could honestly say they were worth it. It's a meaningless fluff item considering how little return on your investment there is.
But in any case, I'm out. Good discussion. After this discussion though I'd be willing to bet you have an iphone or an ipad lol.
even though my black now has windows on it as a bootable drive its not even in the list, but i check my connected sata drives and its right there. along with the blue.
so thats what i dont get. the drive works fine, but its not bootable, 'apparently' - possibly there is a conflict between the 2 drives, im not sure, or maybe by blue has become priority in the boot menu because thats what my bios is used to.*
*something tells me my last sentence is nonsense though, i used to work as a computer technitian for 2 years, but still i think you know more than me about this issue quiz, back when i worked i was used to working with ide drives and windows 95/98/mil/server/xp/vista. dont have much experience with raid or sata, but i know how it works. roughly. what else.. erm there is my black using the 6gb option in the bios, *run all hd's at 6gb's if available*
not sure if that would cause any problems as my blue isnt 6gb/s. - even that too is not the issue i think. hmm.
with that said i am lost really, i suppose i could try unplugging the other sata drive, to see what happens. in relation to that though - after i installed windows on the new drive i did take the blue off from the boot list to see if it would auto boot from the new drive, and it said invalid system disk
which makes sense. hmm. the only way to find out i think is to actually unplug the blue and see what happens.
You made the mistake of installing Windows with the Caviar Blue still plugged in
The fix can get fairly complicated. What I would try doing: Unplug the Caviar Blue. Your computer probably won't boot up once you do that (which is ok). With the Caviar Blue unplugged, reboot off the Windows installation DVD, and go to the Repair options. It should give you the option to Repair and Windows Installation. Do this, with the Caviar Blue unplugged. Once that is done, that should allow you to boot up without the Caviar Blue plugged in.
Now the crap shoot is, when you plug the Caviar Blue back in: Hopefully it will keep right on trucking properly from the Repair fix you just did, but there is a chance it could go right back to the dual-boot menu option; in which case you have to do some MBR magic that isn't easily explained to really fix it. It's also possible that you can go into the BIOS and just change the boot order: set the Caviar Black to boot over the Caviar Blue as the primary drive. But you need to do the WIndows Repair from the install DVD first before this will do anything.
Since your out, it's time for me to put my 2c in.
If you haven't actually used a computer with a good SSD, you have no idea. The difference is night and day. It's much more than just "faster boot speed".
In the end of the day, faster writes only matter when you are doing something really file intensive (such as FRAPS). It isn't very often that we need to write large files and are restricted by the media speed: crap only streams off the internet so fast, optical media installs are limited by the DVD drive and not the hard drive, and not many people run large online relational databases that need to update large datasets frequently. And saving your Word documents, well, most people don't try to save 500Mb Word documents that often.
Most people, on their computers, read small files frequently more than any other operation. And this is what most of the bottleneck is on everyday computing use: small file random reads. ~This~ is your game load times. Take a look at your typical game: it's made up of thousands of very small files. It's usually not one big 5Gb file that has to get loaded (where superior sequential read transfer speeds would be at an advantage). Each time you request a new file, you incur the seek time penalty. Then you get the read transfer speed for the duration of the file. If you have many, many small files, that's a lot of seek times to add in there. (And even Blizzard's big .mpq files really are just archives of smaller files, the game requests them individually inside the archive, and those get hit with seek penalties as well).
A RAID array setup does absolutely nothing for random access. Your still limited by the seek time of the drive: RAID does nothing for seek times, it just provides parallel throughput. Sure, you get better read and write times, because your able to do that across multiple platters at the same time. But guess what, an SSD is in effect a massive RAID 0 device in and of itself! The SSD firmware treats each RAM chip as a parallel data path (they call them "channels"). Plus you get the benefit of near-instantaneous seek times (there is no platter to spin, and no head to reposition, all sectors are available with equal fast speeds)
Sure, you can argue TRIM and performance degredation and blah blah blah, but truth of the matter is, even in a traditional hard drive you still have fragmentation to content with that degrades performance, and even degraded, SSDs still perform very well.
You can look at numbers all day long and wonder why some people preach about SSDs... but until you've used a system with one for any period of time, there is much more to it than just a few numbers. As far as I'm concerned, platter drives are a dying technology to be delegated to cheap bulk storage in the near term, and a dying breed doomed to the same obsolence as the floppy drive in the long term.
"It doesn't effect game play or game performance,"
Usually an SSD doesn't help your frame rates in games. There are some occasional examples of games where it will get rid of stuttering. Making loading screens go faster is quite nice in some games, and unimportant in others.
"I have to say this though, random writes SSD is better, Sequential writes a raid setup is better."
Reads are more commonly the bottleneck, not writes, simply because you do a lot more reading than writing. But you could substitute read for write and the statement is still more or less true. In some cases, it would take more than two hard drives in RAID 0, but you can add enough drives and make it true.
"Random writes are used more often, but... the performance boost in random writes isn't as noticable as the performance boost in the sequential writes."
You must have an extremely unusual workload in order for a hard drive to be meaningfully bottlenecked in sequential transfer speeds, but not in random transfer speeds. In particular, that would mean that you're not using an awful lot of real programs at all. One can easily come up with synthetic applications that are like that (e.g., I'm going to copy this 4 GB file from this drive to that one), but those aren't the sort of thing that a typical user would commonly do. I guess there are probably industrial applications where you have to record some incoming stream of purely sequential data very quickly, and then it will be analyzed elsewhere. But that's the best I can come up with to fit what you describe.
The other way isn't hard. There are plenty of programs that are meaningfully constrained on a hard drive by small file IOPS, but not sequential transfer speeds. For example, operating systems, web browsers, e-mail clients, and most games. That's the sort of thing that a lot of everyday consumers use.
Now, if you're comparing one good SSD to another, then there's more of a case for that, because they're all so fast in IOPS, too. The reason why I say random reads and writes matter so much is that that's the bottleneck for hard drives.
Suppose that you're looking to buy a video card, and have your choice of card A or card B. They both have the same price tag, and equivalent feature sets, power consumption, cooling systems, warranties, and so forth. The only difference is performance. There are only two games that you'll ever want to play on the cards. In one game, card A offers 300 frames per second at the settings you want, and card B offers 200 frames per second. In the other game, card A offers 10 frames per second at the settings you want, and card B offers 60 frames per second. Which card do you pick?
This isn't a trick question. Card B lets you play both games well, while card A only lets you play one of the games. The better choice is card B. That card A wins in average frame rates, 155 to 130, doesn't matter. That card A is faster at the first game doesn't matter. That difference isn't meaningful. The difference that is meaningful is the second game, and that's where card B wins.
Now, that sort of situation is unusual for video cards. Usually if one card is clearly faster than another card in one game, it will at worst be kind of competitive in other games, with only a handful of buggy outliers that will probably be fixed in future drivers. But in storage, a situation like that is ubiquitous. A RAID 0 setup might well beat an SSD in sequential transfer speeds, but they're both so fast that the difference will never particularly matter to most consumers. How often do you really write 1 GB of sequential data at a time? The only times that I can think of that I've ever done that is installing an OS and making an image backup of my SSD, both of which are bottlenecked by other things (optical drive and USB 2.0 connection, respectively), so even then, the sequential transfer speed of an SSD or hard drive still doesn't matter.
"Having a gain there would be far more valuable rather than shaving 10ms off of my random writes and reads lol. "
It's not just 10 ms, once. If that's all it were, then it would be no big deal. It's 10 ms per access. Multiply that by hundreds of things that you have to load at once and it adds up to several seconds, easily, when loading a program. On my old computer, it took about a minute to launch Europa Universalis II because the game had to load thousands of small files. Or even multiply it by dozens of things for loading a typical web page and it's a considerable fraction of a second.
It's kind of like how 10 ns (that's nanoseconds, not ms) wouldn't be a big deal if it were just once, but if you increased the L1 and L2 cache access times by 10 ns for a typical processor, you'd completely cripple it and it would be painfully slow. It doesn't just access them once, but does so many times, and that adds up.
-----
"After this discussion though I'd be willing to bet you have an iphone or an ipad lol. "
http://www.lukesurl.com/archives/1118
All right, so that's kind of a random response. But you get the point.
Would you believe that I don't have a cell phone or a tablet of any sort, and got my first laptop earlier this year?
I'm not one to jump on every little tech gadget that comes along. Among the things that I'm unimpressed by are Blu-Ray, "widescreen" monitors, too many buttons on a mouse, headphones, vent or equivalent programs as used in games, power supplies of ridiculously high wattages, tablets, "smart" phones, and just about any hardware with "gaming" in the name except for a gamepad. (Video cards don't put "gaming" in the name because it's implied.)
But SSDs are a big deal. So, incidentally, are multiple monitors, uninterruptible power supplies, and KVM switches. But I digress.
FRAPS is the only thing I can think of that most gamers would see a big benefit in sequential speeds over an SSD.
Anything else is like... active relational databasing, large scale video rendering/editing, realtime data acquisition.. not stuff most people do with their home PC.
And I totally agree: SSDs are probably the biggest deal to hit computers since the jump from punch paper to magnetic tape. Certainly a bigger deal than even multi-core CPUs or anything else we've seen in the last 20 years.
And UPSes and multiple monitors are all great too... KVM switches, I don't know about. I prefer VNC.
Funny thing: It did actually matter in EQ2, particularly in Neriak and a few other zones. I got about 20% bost in FPS in some areas, I never noticed any difference in another game... But I do know that Neriak is poretty poorly coded and that it loads a lot while you are in there, and the lag is worst in the game...
Zooning in Vanguard didn't get that "hickup" either when you moved to a new area but that is about it.
Everything loading is fast but that rarely affects FPS. It do makes games like AoC and EQ2 nicer to play since it cuts down the loading screen time a lot.