Intel Unveils SSDs With 6Gbit/Sec Throughput

CWmike writes "Intel announced a new line of solid-state drives (SSDs) on Monday that are based on the serial ATA (SATA) 3.0 specification, which doubles I/O throughput compared to previous generation SSDs. Using the SATA 3.0 specs, Intel's new 510 Series gets 6Gbit/sec. performance and thus can take full advantage of the company's transition to higher speed 'Thunderbolt' SATA bus interfaces on the recently introduced second generation Intel Core processor platforms. Supporting data transfers of up to 500MB/sec, the Intel SSD 510 doubles the sequential read speeds and more than triples the sequential write speeds of Intel's SATA 2.0 SSDs. The drives offer sequential write speeds of up to 315MB/sec."

Comments

that's smokin'

[Anonymous Coward]

I hope this blazing speed is used for good and not for evil purposes.

Re:that's smokin'

[ehrichweiss]

If it's twice as fast as all the other SATA devices I have that are rated 3Gb/s then that'll only average about 40MB/s with peaks around 80MB/s..

Re:that's smokin'

[Wovel]

Your devices are not rated at 3Gb/s, the sata connection was. This device is..Supporting data transfers of up to 500MB/sec,....

Maybe just read the summary :)

Re:that's smokin'

[bluemonq]

When transferring files between SATA hard drives on my desktop I usually get around 90 MB/s. I suspect it's your devices that are the problem.

Re:that's smokin'

[Anonymous Coward]

OP seems like too low numbers.

I get 100 MB/sec between two SATA 3 devices on separate connectors. Intel 58 mobo and core i7 and plain jane 1TB Seagate drives.

Re:that's smokin'

[ls671]

Guys, don't forget to take into account cached disk data/buffers, use free to see how much you've got. Just run hdparm -tT to see the difference between cached reads and non cached reads. If this sounds to technical just do a test with a 20 GB file. This should be enough to make sure cached disk data doesn't give you a false illusion of speed. Also read man hdparm under the -t section.

90 MB/s seems on the upper end to me while 40 MB/s is on the lower end but I have seen it with generic drivers.

Also, try to upgrade your SATA drivers to one that is designed specifically for the SATA controller you are using instead of using a generic one. Best bang for the buck if you are looking for speed.

Re:that's smokin'

[Anonymous Coward]

He probably uses Ubuntu.

I don't know what the hell they're doing with drive access, or exactly which hardware suffers from the problem, but I get numbers like that too, and I've heard the same story from lots of other ubuntu users, across Jackalope, Koala, and Lynx.

I'm too lazy to install another OS. I've been using different flavors of linux for a decade now, and the slow transfers and everything freezing up for 10-60 seconds every few hours is just not enough of an issue to go through all the effort of setting things up again, just in hope of not having a worse problem in Suse, Mint, Debian, or whatever.

Re:that's smokin'

[Anonymous Coward]

Which Ubuntu are you using? I think maybe you have a hardware problem or somehow DMA for disk transfers isn't on.

#this is on a 1GB RAM machine so disk cache is <1.3GB.
#disks are 4 x samsung 2TB 5400RPM drives.
cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=10.04
DISTRIB_CODENAME=lucid
DISTRIB_DESCRIPTION="Ubuntu 10.04.2 LTS"

cd $raid10mountpoint

time dd if=/dev/zero of=test.bin bs=131072 count=10000
10000+0 records in
10000+0 records out
1310720000 bytes (1.3 GB) copied, 6.3466 s, 207 MB/s

real 0m6.377s
user 0m0.008s
sys 0m2.220s

time dd if=test.bin of=/dev/null bs=131072
10000+0 records in
10000+0 records out
1310720000 bytes (1.3 GB) copied, 4.71161 s, 278 MB/s

real 0m4.749s
user 0m0.012s
sys 0m1.504s

cd $mirrormountpoint
time dd if=/dev/zero of=test.bin bs=131072 count=10000
10000+0 records in
10000+0 records out
1310720000 bytes (1.3 GB) copied, 11.8113 s, 111 MB/s

real 0m11.888s
user 0m0.000s
sys 0m2.360s

time dd if=test.bin of=/dev/null bs=131072
10000+0 records in
10000+0 records out
1310720000 bytes (1.3 GB) copied, 10.8265 s, 121 MB/s

real 0m10.875s
user 0m0.012s
sys 0m1.316s

Re:that's smokin'

[Rockoon]

If you are averaging a sustained 20MB/sec, then your drive, controller, partition alignment, or drivers suck.. and by suck I mean really really really suck.

It is hard to imagine the great lengths of time you would have to invest in finding a collection of SATA 2.0 hardware that bad, so its almost certainly your partition alignment or drivers.

You do know that modern drives require 4096 byte partition alignment, while most older OS's presume that 512 byte is good enough?

Re:that's smokin'

[omfgnosis]

But would you prefer it be used for good or awesome?

Re:that's smokin'

[Stalks]

By adding sync you are forcing all writes from cache to be flushed to its destination, making for a more accurate test..

time (dd if=test.bin of=/dev/null bs=131072; sync)

Re:that's smokin'

[TheRaven64]

Or, like most people these days, he's using a laptop. 30MB/s is about the fastest I've seen from a laptop drive, and that was when it was completely new so every write was a sequential write. That said, sequential speed is pretty much irrelevant these days. Even cheap hard drives are fast enough that something else (typically the network) is usually the bottleneck when doing sequential reads or writes. Random read / write speed is much more important - on a mechanical disk this can be as low as 100KB/s (seek after every read and you're closer to 50KB/s, even if your read / write operations are infinitely fast), which absolutely cripples performance in a lot of cases.

Re:that's smokin'

[Anonymous Coward]

Yeah, except 500 MB/sec = 4Gb/s

Re:that's smokin'

[Vegemeister]

8b/10b error correcting code.

Re:that's smokin'

[Vegemeister]

I agree on random read performance, but random write speed is pretty much irrelevant. You can always cache writes to ram and write out the cache sequentially when convenient.

Re:that's smokin'

[rwa2]

How to clear your disk cache:
echo 3 | sudo tee /proc/sys/vm/drop_caches

By default, bonnie++ will test using file sizes that are twice your RAM, to make sure that disk caches get overrun.

You'll really want to use bonnie++ or iozone instead of hdparm if you're comparing HDs and SSDs, since SSDs really only shine with lots of small files. The hdparm results would be rather meaningless.

For my part, I'd rather drop $100 on a RAM upgrade than $200 on an SSD. Once you have all your files read into disk cache in RAM, HDs and SSDs perform pretty much identically. If you use the readahead utility to proactively preload all your small files into disk cache, HDs (esp. with RAID10 or RAID5) can perform just as well as SSDs.

http://trumblings.blogspot.com/2010/11/using-readahead-to-speed-up-disk.html

Re:that's smokin'

[pointbeing]

...30MB/s is about the fastest I've seen from a laptop drive, and that was when it was completely new so every write was a sequential write.

Then you'll like this. This was just run off my Atom 270 netbook (HP Mini 110c) with a 500gb Samsung drive and Kubuntu 10.10 -

wizard@wizard-netbook:~$ sudo hdparm -tT /dev/sda
[sudo] password for wizard: /dev/sda:
  Timing cached reads: 1228 MB in 2.00 seconds = 613.74 MB/sec
  Timing buffered disk reads: 312 MB in 3.01 seconds = 103.70 MB/sec
wizard@wizard-netbook:~$ ;-)

Re:that's smokin'

[pointbeing]

write test on the same netbook:

wizard@wizard-netbook:~$ sudo dd if=/dev/zero of=/tmp/output.img bs=8k count=256k && sudo rm /tmp/output.img
262144+0 records in
262144+0 records out
2147483648 bytes (2.1 GB) copied, 19.916 s, 108 MB/s
wizard@wizard-netbook:~$

Re:that's smokin'

[ehrichweiss]

It was on Fedora and the drivers are fine, It *could* be that it was a 4096 vs 512 situation but everything I saw said that both drives were 512...I will investigate this. As for the hardware, it's a non-consumer grade server....not something shoddy.

Re:that's smokin'

[ehrichweiss]

Nope. It's a non-consumer grade server, not a laptop. The SCSI disks it boots from will transfer much more steadily and quickly than anything to do with the SATA by a factor of about 5-10.

Re:that's smokin'

[Skal Tura]

I get 110M/s write speed on magnetic SATA 2 drives :P
Infact, can't supply data fast enough for them to write X)
Theorethically capable of atleast 180M/s write speed, but can't test as i don't have source fast enough.

Now i'm looking to increase that to 240M/s write speed and about 400-440M/s read speed, with approximately 370-440 IOPS (I/O Operations per second). Yes, on magnetic drives.

How? Very easy: Striped RAID. And yes, that works on windows too.

How much?

[Anonymous Coward]

I'll just re-mortgage my house in order to buy one.

Finally, decent write speed from Intel ...

[haruchai]

on a (rich) consumer SSD. But, while I'm loving all the Marvell / Sandforce / Intel hypersonic speed-worthiness, how about a decently fast, really affordable solid state drive? How much longer will these be 20x the per GB cost of a HDD?

Re:Finally, decent write speed from Intel ...

[fucket]

The 80gb Intel X25-M is/was a decently fast, really affordable solid state drive, even at the $240 I paid for mine. Easily dollar-for-dollar the most effective computer upgrade I've ever made.

Re:Finally, decent write speed from Intel ...

[Rockoon]

We saw prices of the previous generation drop from $4/GB to under $2/GB while waiting for SATA 3.0 adoption. Now those same SATA 2.0 SSD's will drop much further... by years end it will be a great time to pick up older SATA 2.0 SSD's

Re:Finally, decent write speed from Intel ...

[haruchai]

Having been a somewhat early adopter of SSDs, I got bitten a couple of time by the JMicron bug - a pox on them. I've decided I won't buy another SSD until I can get a 128 GB drive for $100

Re:Finally, decent write speed from Intel ...

[Anonymous Coward]

How much longer will these be 20x the per GB cost of a HDD?

Probably for the next 10 to 20 years. Flash memory isn't the only thing that gets cheaper over time. Spinning "iron rust" has a pretty good record at that, too. (In 1977, you might have paid $5 each for floppies that stored less than 100 KB. Now $5 would get you 1/16th of a 500 GB portable notebook drive => 300,000 times as much storage for your money.)

The question is not so much when these become cheaper than hard drives, as when these become cheap enough to use. (And if you were satisfied with the capacities and prices that were common a years ago, that time might already be here.)

Re:Finally, decent write speed from Intel ...

[TheLink]

But this new Intel SSD uses a Marvell controller - the same as one used by Crucial's C300. Why should people pay a higher price for an SSD with the same controller? The uninformative article says nothing about that - no proper benchmarks or tests (random reads/writes, sequential/read writes, max latency of a random read/write transaction that interrupts those random/sequential read writes, read/writes of high/mid/low compressible data). Such an article should be pretty useless for real Slashdot nerd.

As for Sandforce, apparently they have problems when waking up from hibernation (google for that). This is not a problem for many desktop users but a showstopper for many laptop users (I won't be buying a sandforce drive for this reason alone).

Re:Finally, decent write speed from Intel ...

[Neil Boekend]

IMHO SSD's should be used as "something in between your HDD and your memory". As long as it's big enough to contain the OS of your choice and all the programs it's big enough. Your MP3's and movies do not require the high throughput. In a desktop this means there should be 2 disks: an ssd for speed and a hdd for capacity. In a laptop there should be an SSD on the mini-PCI-e and an HDD on sata OR, if you must (for space reasons), an SSD on the mobo and an HDD on sata. This would optimize both capacity and speed, while keeping costs (relatively) down.
The cost of an SSD is paid back by the speed, not the capacity. What I find strange is that shops list SSD's by EUR/GB instead of EUR * s/MB. The speed is the defining attribute, not the capacity.

Re:Finally, decent write speed from Intel ...

[mabhatter654]

HDD tech is established... it's like complaining that 1.5 TB disks cost $100 when a backup tape the same size costs $50. You're paying for speed and essentially today's device helps pay the companies to make tomorrow's devices as well.

The thing I really want to see is better syncing of devices. Really I don't NEED more than 64GB or 128GB on a laptop, certainly not for my day job... and I'm one of the biggest users of HDD storage at my company. If there was a good, solid way to sync to a 1TB drive easily and painlessly. I'm not opposed to syncing my laptop like I do my iPhone... but the media selection needs to be seamless. Done correctly, I could even sync music from my drive at home over the internet... movies would be a little harder but not impossible.

Re:Finally, decent write speed from Intel ...

[thue]

> Your MP3's and movies do not require the high throughput.

And more importantly, your MP3's and movies do not require the random reads and writes which is an SSD's greatest strengths.

Re:Finally, decent write speed from Intel ...

[Lennie]

That is what ZFS with L2ARC ('level 2'-cache) does, it uses the SSD as a cache for the slower but bigger disks.

On Linux a fairly new development called bcache does something similair

Re:Finally, decent write speed from Intel ...

[Anonymous Coward]

SynchML / Funambol?

Funambol Server

Re:Finally, decent write speed from Intel ...

[symbolset]

Blame Apple and Android. They'e consuming so much flash capacity that there's precious little left for this stuff.

Re:Finally, decent write speed from Intel ...

[TheRaven64]

For a server deployment, I'd agree, but on a laptop the fewer moving parts, the better. Now if only we could get rid of the fans...

Re:Finally, decent write speed from Intel ...

[Anonymous Coward]

Because performance is not a single parameter.

You have (reads, writes) * (sequential, random) * (various queue depths) + latency. So which one do you want?

Re:Finally, decent write speed from Intel ...

[ShakaUVM]

>>IMHO SSD's should be used as "something in between your HDD and your memory"

You're talking about what is essentially Vista/Win7's readyboost feature, using flash as a cache between RAM and HDD.

You could do this now. Just buy an external SSD and plug it in.

Re:Finally, decent write speed from Intel ...

[ShakaUVM]

>>HDD tech is established... it's like complaining that 1.5 TB disks cost $100 when a backup tape the same size costs $50.

Except two or three high speed HDDs in a RAID configuration will outperform a SSD on most tasks, and cost about 10x less.

I keep looking at SSDs, but their price and performance just aren't where they need to be to buy one other than... just 'cause they're cool. If they even had a 512GB model available for $200 (which is twice as much as I paid for my three 500GB drives 6 years ago), I'd probably buy one. But 512GB SSDs are still around a thousand bucks, which is just outrageous...

Re:Finally, decent write speed from Intel ...

[jbolden]

-- How much longer will these be 20x the per GB cost of a HDD?

They spread will come down but there will be a spread until long after no one is using HDD. http://en.wikipedia.org/wiki/File:Disruptivetechnology.gif

I'd agree with your 20x right now at say the 256 g - 500 g levels. But lets say when SSD is around 3t the spread might only be 10. At 20t the spread might only be 5. And maybe at 20t people start to switch in mass. This at first drives the spread back up close to 10 but within 2 years, the spread is like 2-3 with 35t SSD costing the same as 100t HDD. At that point only the most demanding users, those people who need huge amounts of cheap storage, exabytes worth still buy HDD. That's called "a retreat to quality". That small market kills their ability to get the innovations they need and so likely SSD becomes cheaper when it hits 200t or something.

Those are rough numbers but lets say around the time you are buying 1p SSD hard drive it will be cheaper than the HDD alternative.

Re:Finally, decent write speed from Intel ...

[Neil Boekend]

No. I am talking about splitting the stuff we put on the harddisk now manually. Not as a cache, but the only version of the OS and the most important software.
Having said that: Yes, I do think cache is a good use of SSD's. Just not the best use imaginable.

Re:Finally, decent write speed from Intel ...

[dc29A]

Never had any issues with my Vertex 2 waking up from hibernation.

Re:Finally, decent write speed from Intel ...

[chainsaw1]

The speed of these devices on SATA3 (500 MB/sec) is almost that of PC133 DRAM (570MB/sec). While you could make that argument for a modern system, anything spec'd to run on a P3/1st gen Athlon "should" be able to use the HD as raw memory with similar performance to those machines.

Re:Finally, decent write speed from Intel ...

[jbolden]

No a mediocre SSD will crush a very expensive RAID-5 in read tests. If you go up to something like RAID-15 across 30 drives or so, yeah the raid will be better. SSD really is much faster otherwise no one would be all that excited about paying 20x the cost.

Re:Finally, decent write speed from Intel ...

[dc29A]

Except two or three high speed HDDs in a RAID configuration will outperform a SSD on most tasks, and cost about 10x less.

Of course, however ...
(1) SSDs own platter drivers in random read/write. What people do on their machines? Mostly random reads/writes!
(2) SSDs consume less power, are dead silent, have no moving parts.
(3) You can't cram three 15k SCSI drives into a MacBook. Well ... maybe you could but it would look very funny and be way less portable.

Re:Finally, decent write speed from Intel ...

[timeOday]

Well, the *sequential* read/write speeds of this new drive also dominate spinning hard drives by a fantastic amount. Just a few years ago I was striping hard drives to hit 50 MB/s for a video capture device; ONE of these does over 300 MB/s, and they fit in a laptop! "SSD's: they're not just for IOPS any more."

Re:Finally, decent write speed from Intel ...

[drinkypoo]

IMHO SSD's should be used as "something in between your HDD and your memory"

Unfortunately, only ZFS and Windows Vista or later with Readyboost and Superfetch will actually DO this without application support. There is an experimental dm-cache module for Linux which does this (it's a block-level cache) but it does not work on contemporary kernels and appears to have been abandoned.

Your idea has a little merit, though, because hybrid drives already exist. Of course, they have even poorer driver support.

Re:Finally, decent write speed from Intel ...

[MarkRose]

Unlikely. If it's sitting on a plain jane PCi bus, the max throughput is only 133 MB/s (32 bits * 33 MHz). And the latency is likely 10 to 100 times higher.

Re:Finally, decent write speed from Intel ...

[Yvan256]

Except that even "spinning iron rust" drives themselves have slowed down, they don't go up in capacity every six-eight months like they used to. Maybe parallel recording has already hit its limits?

Re:Finally, decent write speed from Intel ...

[Yvan256]

I meant perpendicular, not parallel.

It would be nice to have a 15 seconds window to edit our posts.

Re:Finally, decent write speed from Intel ...

[Anonymous Coward]

No, they shouldn't. Flash caching mechanisms have been tried and they're lame. There just isn't any desktop (Notice I didn't say enterprise) performance to be had by throwing some flash memory and a memory or block based caching scheme at a computer.

Why? The system can't tell which blocks or pages or whatnot should be on flash and which shouldn't. Diminishing returns kicks in really fast. It's not worth the trouble
You need to bite the bullet and opt for flash storage. Lots needs to be purchased. The prices need to be driven down by volume and advances in manufacturing.

If you want split flash/traditional storage only application/OS level awareness will get you any sort of semi-automatic optimization. Consider a storage scheme where you have "Fast/small"(flash) and "Slow/big"(traditional magnetic) storage. Have an API where applications can store data and code that's best loaded from flash, and bulk content that can be loaded from magnetic (Which still has fine sequential performance) The API would also let the OS manage user data in a similar manner, offloading large bulk storage to magnetic.

I do a lame hacked up ad-hoc version of this with NTFS hard links and built in windows features. I've got an 80 gig intel SSD where windows is installed. I let programs install to their default locations in c:\Program Files or C:\Program Files (x86). Most programs are pretty small and I let them reside on flash. Makes them quick to load. (Thats why you get an SSD, right?) Games, however, are big and impractical for my smallish SSD. When a game installs I check it's default folder, then make a hard link to a different folder on a 1TB hard drive. This way the program stays in it's default folder but it's big hogging content doesn't occupy expensive flash blocks. Steam makes this trick particularly easy because all of steam's content, user data, everything ends up in one neat folder.. I pretty much buy games on steam exclusively, now (Except for WoW.. Which is the only game allowed to live on the SSD. Gotta have priorities, you see ;D - Even then I've linked out some of the folders where it's downloaded bulky update files too)

Windows lets you change the locations of your default user 'content' folders like pictures, documents, movies, music, etc. This is handy because application makers usually use these folders as default locations. Normally these folders reside on your profile, which is in the C:\Users folder. I just point these folders to places on the hard drive and now I don't have untold gigs of RAW files I snapped ending up on the SSD.

Policing your profile folder is a bit annoying under windows. Some programs have a bad habit of dropping gobs and gobs of mostly useless temp files and cached data there. This is a problem even on magnetic storage as a big profile folder is one of the main causes of windows 'slowing down' for users. Lots of routines and processes have to touch the profile folder and it's filled with 100's of thousands of useless files they all need to be parsed in some manner by FS drivers and whatnot.

Re:Finally, decent write speed from Intel ...

[Lennie]

And I also found flashcache which was developed by Facebook which does the same.

bcache or flashcache both are not in the mainline kernel (yet)

Re:Finally, decent write speed from Intel ...

[haruchai]

My suggestion to you is to ask a question or two before taking a flying fuck at a rolling assumption. Did you overlook the word "early"?
I bought an OCZ core in 2008 largely based on the (at the time) impressive cost-per-GB and this review: http://benchmarkreviews.com/index.php?option=com_content&task=view&id=200&Itemid=1&limit=1&limitstart=9

I later got a Kingston V-series which was supposed to have a fixed controller - at first, they didn't admit it was actually a JMicron but a Toshiba. After questioning them, it turned out Kingston, JMicron and Toshiba made some adjustments to the design to mitigate stuttering at the cost of a reduction in read/write speed. It was better but would still stutter under load, just not as often or for quite as long.

Re:Finally, decent write speed from Intel ...

[Neil Boekend]

People keep reading my "something in between your HDD and your memory" as cache. It's not what I meant. I meant it is a separate device, not a cache for the disk. It should contain the OS and some software (as you do it). Caching is a less-than-perfect option, if you have the option to use the SSD for low-latency things directly.

Re:Finally, decent write speed from Intel ...

[haruchai]

Well, instead of saving up for a another big SSD, I'll go buy whatever it is you're smoking!! A petabyte SSD / HDD?? At equal cost? Just the thing I need to store my holodeck scenarios.

Re:Finally, decent write speed from Intel ...

[jbolden]

My first hard drive was 20m and ran about $2k. I'm looking at a personal 4 disk storage unit which is 16t (5x4 - parity) at about the same cost. That's a million fold increase in storage in under 30 years. Pushing that even further. The most common data interchange was a single 128k 5 1/4" inch floppy, that disk would hold most of a person's actives files the way a student might use a large USB key today.

  The only reason people don't have an iTunes like player for their television with thousands upon thousands of high def TV shoes and movies is they don't have the storage yet.

Re:Finally, decent write speed from Intel ...

[Nick Ives]

PC133 was more like 1GB a sec. 133Mhz * 64bit bus / 8 / 1024^3 ~ 1GB/s.

Re:Finally, decent write speed from Intel ...

[Nick Ives]

That's because "something in between your HDD and your memory" is a good definition of a certain type of cache.

What you should have said is that SSDs should be used to store the OS and common programmes.

Re:Finally, decent write speed from Intel ...

[Anonymous Coward]

I'm thinking of making a hybrid filesystem that stores the first X bytes (meg or 1/2 meg or so, configurable) of the file on the SSD, and the rest on a spinning disk. This also eliminates the discrepancy between large and small files. Need to get off of my ass...

Re:Finally, decent write speed from Intel ...

[Nick Ives]

The only reason for buying a SSD over a Winchester drive is random I/O. That's always been the case and will continue to be the case for the foreseeable future.

SSDs aren't going to compete with traditional rotating media on either cost/GB or raw throughput (compared to RAID) because you can do both those very cheaply with traditional media. The big win is in small random small page I/O, which is what makes up most of an average desktop load.

Now that I'm used to having my OS on an SSD, I find it agonising booting up systems that run on traditional media. A traditional 7200RPM drive can easily drop into the single digit MB/s during system boot up and initial log on. Modern "green" HDDs will drop below 1MB/s when dealing with sustained random 4k I/O! You can mitigate that by not having programmes start up automatically and disabling services that you don't really use, or you can install an SSD and just have that problem disappear.

Re:Finally, decent write speed from Intel ...

[haruchai]

I don't think that the spinning rust disks can maintain the impressive growth rate we've seen to date. Of course, I'm far from the 1st to say this but I do think we're close to a tipping point for SSDs (finally!!).

Re:Finally, decent write speed from Intel ...

[jbolden]

I don't know. Toshiba has already prototyped bit-patterned recording which is good for another factor of 10. So maybe the move from 5t to 50t is mainly driven off that technology. Seagate is playing around with Heat-assisted magnetic recording which gets a factor of 100. I still think the direction is up in terms of storage at least for another decade. But then again who saw the crash in CPU performance?

  This subthread started with the question "How much longer will these be 20x the per GB cost of a HDD?"

I responded

They spread will come down but there will be a spread until long after no one is using HDD. http://en.wikipedia.org/wiki/File:Disruptivetechnology.gif [wikipedia.org]

I'd agree with your 20x right now at say the 256 g - 500 g levels. But lets say when SSD is around 3t the spread might only be 10. At 20t the spread might only be 5. And maybe at 20t people start to switch in mass. This at first drives the spread back up close to 10 but within 2 years, the spread is like 2-3 with 35t SSD costing the same as 100t HDD. At that point only the most demanding users, those people who need huge amounts of cheap storage, exabytes worth still buy HDD. That's called "a retreat to quality". That small market kills their ability to get the innovations they need and so likely SSD becomes cheaper when it hits 200t or something.

Those are rough numbers but lets say around the time you are buying 1p SSD hard drive it will be cheaper than the HDD alternative.

So we are talking about a different issue, I was talking about the cross over point when SSD gets less expensive. As far as

Re:Finally, decent write speed from Intel ...

[Tibia1]

You are forgetting that when a big line of different capacities is released (all on the same speed from the same manufacturer..), then the only variable with any influence on price is capacity in GBs. The idea is to get them to buy the higher capacity 240GB SSD by offering a slightly better $/GB. Also, people have many programs to install these days and are looking especially at the capacity, when considering their additional xGBs of games they would like to put on their SSD. Besides, when a higher speed comes out (It will keep doubling at least), everyone knows about the speed and the entire market is reassessed.

Re:Finally, decent write speed from Intel ...

[ShakaUVM]

Right.

But I've done the "small/fast OS drive" + "large/slow data drive" (well, it's actually going to be pretty fast for streaming media) before, and it's a total pain in the ass when your OS drive isn't big enough for all the stuff that installs to it by default. Then you're spending half your time uninstalling apps, and reinstalling them on the data drive when you run out of room (if you can find the disks) and shuffling apps you want to run faster onto the SSD, and basically spending much more time fucking about with your system than you get benefit from in terms of increased boot speeds. I reboot, what, once every month or so? I can go get a soda. It's way better than having to micromanage.

Basically, my threshold is about 500GB for an OS drive - that's large enough to not have to worry about installing apps onto your secondary drive, and with space to spare. At 250GB, you won't even be able to have 20 apps (especially games) installed at once on your computer, which is just crippling to me. But 500GB SSDs are still outrageously expensive. So when I build my new machine, it'll probably be on top of another RAID0 of two high speed HDDs with automatic mirroring onto the third drive.

Re:Finally, decent write speed from Intel ...

[Nick Ives]

Most games are fine with living on a traditional HD. You only have to worry about the initial load times and then after that, everything will be cached in RAM anyway.

Just have all your OS and regular apps on SSD and install games on your RAID. The few games that have random small I/O patterns (for me, that's my Baldur's Gate 1+2 Tutu + mods install that has loads of crap in the override folder), put them on your SSD. Most games just have resource files containing relatively large assets so your regular disk will be reading a few MB at a time before seeking. Those resources files are often optimised to put common assets close together in the assumption they'll be on optical media on a console, which obviously has much worse random I/O performance.

"Thunderbolt SATA bus interfaces"?

[Chuck Chunder]

Somebody is confused. Thunderbolt is DisplayPort and PCIe bundled together.

The SATA 3 ports on Cougar Point platform have nothing to do with Thunderbolt.

Re:"Thunderbolt SATA bus interfaces"?

[Anonymous Coward]

Somebody is confused. Thunderbolt is DisplayPort and PCIe bundled together.

The SATA 3 ports on Cougar Point platform have nothing to do with Thunderbolt.

I'd vote this up if I had any karma. Completely correct AFAIK, and I'm in a position to know.

Re:"Thunderbolt SATA bus interfaces"?

[keith_nt4]

The summary is probably oddly worded. I think the summary just meant this new SSD will be able to keep up with the sata 6 speeds from a "thunderbolt" interface. I just saw a podcast about "thunderbolt" recently (tekzilla I think?). Looks like the interface will be that of DisplayPort but be able to do video/hdd/network/audio/other peripherals (and power also) whilst daisy chaining a number of devices. Not fiber optic like originally talked about (maybe later) but still fast.

Re:"Thunderbolt SATA bus interfaces"?

[2themax]

Yes, somebody is probably confused, but I think we will eventually see hard drive enclosures with a Thunderbolt port on the outside and a PCI-e to SATA 3.0 bridge chip inside.

Re:"Thunderbolt SATA bus interfaces"?

[Anonymous Coward]

Thunderbolt is the commercial name of Light Peak and the MacBook Pro implementation uses a connector identical to the Mini DisplayPort kind. Thunderbolt is protocol agnostic, so it is more than DisplayPort and PCI-E bundled together. Thunderbolt is also the name given by Intel, not Apple.

Re:"Thunderbolt SATA bus interfaces"?

[Redlazer]

I believe it's intended to be used for external hard drives as well. Lacie and WD have signed on to develop compatible hardware - I imagined at the time it would be in that vein.

Is there any reason you can't transfer hard drive data over PCIe?

Re:"Thunderbolt SATA bus interfaces"?

[Chuck Chunder]

Thunderbolt is protocol agnostic, so it is more than DisplayPort and PCI-E bundled together.

Thunderbolt only supports two protocols, DisplayPort and PCI-E. Other controllers can hang off the end of the PCI-E channel and drive other protocols from there but Thunderbolt itself is certainly only DisplayPort and PCI-E.

Re:"Thunderbolt SATA bus interfaces"?

[benow]

Optical is still in the spec, but probably only to be used for long spans.

Re:"Thunderbolt SATA bus interfaces"?

[syousef]

Somebody is confused. .

With Intel's naming (lack of) conventions I'm surprised anyone ISN'T confused.

Re:"Thunderbolt SATA bus interfaces"?

[Anonymous Coward]

I suppose that you could design the controller on the back of a HDD or SSD to "directly" support the PCI-E part of Thunderbolt. On the other hand, given that everyone has standardized on SATA for "low-end" drives, Thunderbolt-to-eSATA adapters (or cases that take Thunderbolt and have SATA adapters inside) seem like they'd be the path of least resistance.

Re:"Thunderbolt SATA bus interfaces"?

[Neil Boekend]

No there is no reason, as can be seen by the first card to appear in Google The PCI-e card simply needs a sata controller.

Re:"Thunderbolt SATA bus interfaces"?

[mabhatter654]

that's the only way to feed the darn Thunderbolt bandwidth! Even to max Thunderbolt you'd need at least 3 of these in RAID of some fashion, meaning something to control them, like Drobo.

Re:"Thunderbolt SATA bus interfaces"?

[mabhatter654]

that makes it simpler to implement as a device. Basically the consumer equivalent to 10Gb Ethernet/8GB Fiber Channel on servers that can speak Ethernet or Fiber Channel drive... (and virtualize anything else)

Consumers care about Video and Audio devices mostly, maybe a smattering of other things. This brings back the PCI enclosure again and should open up Macs to things like Robotics, test equipment, etc that requires dedicated hardware cards most Macs can't have.

Re:"Thunderbolt SATA bus interfaces"?

[Anonymous Coward]

this pcie card *is* a sata controller.

Re:"Thunderbolt SATA bus interfaces"?

[Yvan256]

Could this also possibly mean external GPUs? I wouldn't mind something half the size of a Mac mini stacked under if it meant a dedicated GPU with its own RAM.

The perfect compliment to Sandybridge

[Anonymous Coward]

Oh wait....

Re:The perfect compliment to Sandybridge

[Ster]

The perfect compliment to Sandybridge

Actually, yeah. The problem with Cougar Point (the Platform Controller Hub that goes along with Sandy Bridge, not the Sandy Bridge chip itself, BTW) is with the SATA 3.0Gbps ports. So to get the maximum performance of these new SSDs, you wouldn't use those ports anyway.

Wear usage?

[wisebabo]

I know this problem has (probably) been satisfactorily addressed but if one were to use such a super fast drive for an application that had extremely heavy usage (swap space for the OS or an program like Photoshop) wouldn't it cause those sectors to be read/written to many many times very quickly? Doesn't each "cell" have a limited number off times it can be accessed before it fails? (on the order of 100,000 i think). And wouldn't that case the drive to fail (sooner rather than later because it is so fast)?

Again, I'm sure the SSD drive manufacturers have looked at this problem very closely, I'm just concerned that's all. After all, even if your computer made only one error every billion instructions that would mean it would break down in less than a second!

Otherwise, I'm getting one of the 512GB drives for a smokin' fast laptop drive so I boot up lightening fast! (Time can never be recovered especially for an old fart like me).

Re:Wear usage?

[pz]

Again, I'm sure the SSD drive manufacturers have looked at this problem very closely, I'm just concerned that's all.

So, look up the specs, then. Current write cycles are over 1,000,000 per cell. Modern wear-leveling algorithms combined with extra blocks and ECC mean that it's more likely that some other component will fail before your SSD will.

Besides, if you were really concerned, and not just trolling, wouldn't you have the same issues with your hard drive, too? Doubly so in a laptop?

Re:Wear usage?

[Microlith]

Current write cycles are over 1,000,000 per cell.

Far from it. Depending on the litho, at 25nm for instance you're down to 3,000 program/erase cycles. And even at 34nm, you're still little better than 10k cycles. The overprovisioning and ECC required at these scales is massive.

But yes, studies have been done and it takes an industrial strength workload to kill an SSD. If one of these is in your home machine, you likely won't kill it. If you think you might, then you should already have practices in place to deal with disk failure.

Re:Wear usage?

[Anonymous Coward]

Besides, if you were really concerned, and not just trolling, wouldn't you have the same issues with your hard drive, too? Doubly so in a laptop?

What are you talking about?

Re:Wear usage?

[TrancePhreak]

This is why I believe in hybrid setups. The other replies to this are talking about mere millions of writes. Do you know how many writes the swap file gets every hour you use your computer? I too would be worried about a pure SSD setup. If you check out the how-to guides on installing SSD's in your machine, they almost all mention to set the swap on a non-SSD and to move your home directory off it too. (Home directory contains such things as temp web files!)

Re:Wear usage?

[Anonymous Coward]

There are a lot of interesting papers on the reliability decline of flash as the technology shrinks and becomes more affordable. Every year at ISSCC, companies like Samsung and Hynix will come and present a tiny portion of how their flash controller works, and as part of the presentation they talk about raw error rates of MLC NAND flash. They are easily 1e-6 per cell for 45nm, but at 32nm and 27nm it's a different story (~600x worse BER, with ~100-1000 write cycles per cell). Consumers don't usually see this side of flash because manufacturers will build in enough ECC and wear leveling to make the device useable, but this comes at the with a price in terms of power and capacity (that write cycle count is AFTER the wear-leveling).

Sadly most of the interesting papers require an IEEE subscription, but there's one from 2009 that gives a nice overview of the ECC required to get the overall error rate to something reasonable:

http://cyclicdesign.com/whitepapers/Cyclic_Design_NAND_ECC.pdf

Re:Wear usage?

[Kjella]

So, look up the specs, then. Current write cycles are over 1,000,000 per cell. Modern wear-leveling algorithms combined with extra blocks and ECC mean that it's more likely that some other component will fail before your SSD will.

Are you looking at MTBF numbers or something? Expensive 34nm flash has 10,000/cell, cheap 34nm 5000/cell and 25nm is down to 3000/cell.

And yes, you can kill an SSD that way, already done it. Granted, I tortured it in pretty much every way possible by running torrents and Freenet 24/7 and keeping it 90%+ full all the time. It died after about 1.5 years with 7000 writes/cell average, 15000 highest.

Fortunately for me I still managed to sneak it in as a warranty repair, even though they aren't supposed to do that on drives that are worn out. Now I've got a new SSD and I'm doing all my heavy lifting to rotating media keeping it as a fast OS/application disk. I'm guessing it'll last 10+ years that way. But if you want to kill it quickly, you can.

This is very much unlike rotating media. Google released a huge bunch of data on HDD reliability, what they found was that actual use didn't really matter. Hard disks lasted more or less X years spinning away at 5400-7200 rpm whether you read/wrote to them or not, as long as they weren't completely idle and could sleep. SSDs are exactly the opposite, each read/write takes a little bit of life out of it.

Re:Wear usage?

[aztracker1]

I'm paranoid... generally redundant backups over raid redundancy as Ive had raid1 with both drives dead relatively close together, one time before I had a replacement, my nas box has a spare next to it. Though my intel ssd died att 11 months in my desktop, I still won't go back for my boot+os ...

Re:Wear usage?

[moogied]

Hi! Its called ram. Its where you store data that you plan to access and manipulate a lot before you save it.

Re:Wear usage?

[m.dillon]

Insofar as actual swap (as in paging physical ram in and out) goes, it depends heavily on the amount of ram the machine has and the type of work load. Mostly-read workloads with data sets in memory that are too large can be cached in swap without screwing up a SSD. A large dirty data set, however, will cause continuous heavy writing to swap space and that can shorten the life of SSD-based swap very easily.

A system which needs to page something out will stall badly if the pageout activity is rate-limited, thus you just don't want to use an SSD in any situation where huge, continuous pageout rates are expected. Huge pagein/re-pagein rates due to largely static in-memory data sets are fine. Huge pageout rates are not.

It is a different story for filesystem meta-data or data caching. You should be able to control worst-case write rates in those sorts of setups (at least you can with DragonFly's swapcache), making the SSD a huge win under virtually any circumstances.

-Matt

Re:Wear usage?

[TheEyes]

But yes, studies have been done and it takes an industrial strength workload to kill an SSD. If one of these is in your home machine, you likely won't kill it. If you think you might, then you should already have practices in place to deal with disk failure.

Just as important to note is the failure mode for flash memory is for it to become read-only; in other words, it simply becomes impossible to delete what is written on your drive, which is a perfect reminder to get a new one. Given that this sad event will be nearly ten years from now, it should be dirt cheap to buy a replacement drive.

When you do, though, don't forget to remove the metal shell on the old drive and cook it in the microwave for a minute or two to destroy your old data. It's not like you're going to be able to sell the drive used anyway.

Re:Wear usage?

[Anonymous Coward]

I thought I'd find you here. Quit procrastinating, you've got a release to work on!

-Not Matt :-P

Re:Wear usage?

[Anonymous Coward]

If your swap file is constantly being written to, you should spend your money on memory, not SSDs :) It really doesn't get that many writes.

"Mere millions of writes" -- I don't think you know what you're talking about. A normal user probably wouldn't get million writes per cell during _his_ lifetime... I realize that's not even relevant to the question here because these new denser SSDs don't have that kind of write-counts but you brought that number up.

Re:Wear usage?

[Tapewolf]

Just as important to note is the failure mode for flash memory is for it to become read-only;

Are you sure? I was under the impression that it worked like EPROM, where the bits were set high by the erase cycle and data was written by grounding the bits which needed to be zero.
That being the case, it's more likely that the data would be corrupted (since it would fail to set bits anymore) which is actually the sort of thing one of my old USB keys started to do.

'Course, there might well be logic in the controller to detect this and put the drive into read-only mode when it runs out of non-defective blocks, but I don't think that would be an inherent property of flash drives and certainly not one you'd want to rely on.

Re:Wear usage?

[Anonymous Coward]

I've never run into any source that says anything different from this. The big problems that arise are, first this will show up during an erase-write cycle, so you only discover this when things explode. Second and very dangerous is the claims of 100,000 write cycles are including wear-leveling, hence after 100,000 writes or the first failure shows up, the entire drive is 100% toast (unlike disks where you can typically still read and write 90% of the data).

Re:Wear usage?

[Rockoon]

This is why I believe in hybrid setups. The other replies to this are talking about mere millions of writes.

There are no replies talking about mere millions of writes. The one reply that did mention 1 million was talking about per cell and is also very wrong. Furthermore, we do not speak of write limits with SSD's. We speak of erase limits. We should learn the difference before being critical. SSD's write in 4K Pages but erase in Blocks. OCZ is currently using 128 Pages per Block, so Blocks are 512K on their devices.

Do you know how many writes the swap file gets every hour you use your computer?

You are attempting to invoke a nebulous unknown to support your argument for hybrids. When you know how much you are writing per hour on average, and how that compares to the limits of these drives, then you might have support for your belief in hybrids. Right now all you have given is voodoo hand-waving. Swapping has been considered bad for a lot longer than SSD's have been around, because disk reads and writes were slow, so you are essentially only hitch-hiking on that traditional negativity here. On modern SSD's, reads and writes are much faster, so that traditional reason for hating swap doesnt just naturally cross over.

My guess is that you are severely over-estimating how much writing you are doing per day while also severely under-estimating how much writing can be done on these drives. The erase-limits of these drives will last years even if you routinely erase and refill them entirely each day.

Re:Wear usage?

[Anonymous Coward]

you'd set the bits high for a secure erase, but normal deletions would just remove the entry from the filesystems index (as is done on rotary drives). The data itself is still there until overwritten.

Re:Wear usage?

[Soft Cosmic Rusk]

Uh... Sorry, but the data is in the microchips, not in the metal shell. Cooking it is more likely to destroy the microwave than the data!

Re:Wear usage?

[Anonymous Coward]

unless you have lots and lots of porn

Re:Wear usage?

[TheRaven64]

Do you know how many writes the swap file gets every hour you use your computer

Not specifically, but my OS - like pretty much any modern OS - gives me a disk I/O total since the last reboot. I last rebooted 11 days ago. Since then, I've written just under 50GB, and read 27.6GB. I don't think reads wear out flash, so we'll ignore the second number. That works out at 1660GB written per year, if the last 11 days have been representative. Assuming perfect weak levelling, that's 6.4 rewrite cycles per year for a 256GB SSD. A drive that can 'only' handle 3,000 rewrites will therefore wear out after about 450 years. If it does a tenth as well as that, then it will last for almost as long as hard drives have existed.

The advantage of hybrids is the same as the advantage of CPU cache - you get almost the same performance as a very large SSD for a much lower price. Given infinite funds, you'd build a computer that had gigabytes of SRAM, but it's much more expensive than DRAM, so you only have a few megabytes of SRAM and a lot of DRAM, which gives you almost as good performance but costs vastly less (the relative complexity means that the lower bound for SRAM is about six times the price per bit of DRAM - in practice it's higher). If someone else is paying, I'd take the pure SSD solution. If I had to balance price against performance, then ZFS's L2ARC is currently likely to be the best solution (unless it's a laptop, where space is an issue).

It's worth noting that Oracle is currently looking at using rotating disks as L3ARC, with tapes as the persistent storage. Writes remain fast, because ZFS is copy-on-write, so every write to the tape is just an append operation, but L3 cache misses are very expensive (you need to seek the tape, which can take several seconds, rather than several milliseconds) - for write-heavy workloads, it can give very good performance per dollar.

Re:Wear usage?

[Anonymous Coward]

I've had my first USB keys (bought in 2003) fail me last year. They basically became unreadable all of a sudden. There were warning signs a few years before (some bad sectors).

I don't use SSD as permanent storage, and certainly not for backup purposes...

Re:Wear usage?

[klui]

No, it doesn't work that way. Each cell contains a charge that is the pattern for a single or multiple bit. If wear leveling doesn't move the cell to another while the drive is turned on, the charges will leak and your pattern will either be all ones, all zeroes, or somewhere in between. The directory structure will also be corrupted. Yeah, you won't be able to write new data, but you probably won't reliably read what's already written.

Re:Wear usage?

[blind biker]

So, look up the specs, then. Current write cycles are over 1,000,000 per cell.

That's only almost true (it's about 100.000 reqrite cycles per cell) for SLC cells. SSDs unfortunately use MLC cells with a rewrite cycle number of 5000.

But hey, never miss an opportunity to call someone a troll, right?

Re:Wear usage?

[EmagGeek]

Current write cycles are nowhere near 1E6. 34nm flash is good for 3000-5000 PE cycles, and 22nm flash is good for between 1000-2000.

Re:Wear usage?

[EmagGeek]

>> When you do, though, don't forget to remove the metal shell on the old drive and cook it in the microwave for a minute or two to destroy your old data.

That's a great way to contaminate your microwave with all kinds of toxic chemicals that will later get into your food.

Re:Wear usage?

[DarkXale]

Sectors die in mechanical hard drives as well - and judging by my drives - at a far higher rate than they do in SSDs. Plus with mechanical drives, you also have to cope with mechanical failures bringing the entire drive down. A badly timed bump in a laptop for example can cause severe drive damage either immediately, or make the drive fail significantly sooner. Many people say hard drives rarely die; but thats hardly true in reality. Its true that they're not that likely to die for your average person - but hard drive deaths occur plenty when used in more active machines. If you see few or no mechanicals die with your use, you'll never see a SSD die with your use. I've owned my SSD for about a year and a half - but it would only take 6 hours to do the same amount of data writes I've done during that entire time, which is about 2.1TB of written data. (SMART info) On average - I replace around 3 mechanical hard drives a year within my entire family due to a drive crash, 6 in 2010 though. In testing, and in enterprise - those drives go through days and weeks (months!) of continuous writes at max speed. For your average person which uses their machine far less than me - theres not a chance their SSD will encounter significant wear issues.

Re:Wear usage?

[Anonymous Coward]

+100

I was scared too after buying my first Intel X25-M G2 160 GB some months before. I even moved my browser cache to a ramdisk. That's just bullshit. According to the Intel blah Drive Information tool, my SSD ate 400 GB writes since about 5 months. That's peanuts. All your other devices will either fail or be replaced before your SSD craps.

Re:Wear usage?

[jbolden]

Its a large number of writes not reads. Something like swap space gets allocated and even after it no longer reflects memory it stays empty for "a while". As long as "a while" is an hour or more on average you are fine.

Re:Wear usage?

[Anonymous Coward]

When you do, though, don't forget to remove the metal shell on the old drive and cook it in the microwave for a minute or two to destroy your old data.

If you're looking to burn down your house, you can just as easily start the fire the old fashioned way, with matches. So long as your drive's somewhere near the flames, your data will still be securely rendered inaccessible.

Re:Wear usage?

[drinkypoo]

Swapping to flash is dumb. But more importantly, with modern memory prices, swapping is dumb. You're not going to end up in swap unless an application misbehaves, and then the swapping is going to bring your system to a standstill. Better to run out of memory and let the OOM system kill processes. You can get 4GB SODIMMs now. Swap? That's over.

Re:Wear usage?

[T-Bone-T]

Too bad Newegg lists 17 SLC SSDs. They are also insanely expensive at $10/GB. What was that about trolling?

Re:Wear usage?

[Rockoon]

Assuming perfect weak levelling, that's 6.4 rewrite cycles per year for a 256GB SSD. A drive that can 'only' handle 3,000 rewrites will therefore wear out after about 450 years. If it does a tenth as well as that, then it will last for almost as long as hard drives have existed.

There is another factor that you arent considering, and that is write amplification. In an ideal world, the write amplification factor will be 1x (no amplification), but could be (for OCZ's current lineup) as bad as 128x (each 4 KB write causes a 512 KB erase.) I havent seen an actual write amplification study for various workloads, but the anecdotal reports are less than 2x.

I think what most people dont understand is that once we are talking about 40+GB drives, the erase limits really do become trivialized by the sheer enormity of the amount of erases (measured in KB) that these devices can sustain. Even 100000 erase cycles isnt very impressive on a 1 MB drive (100 GB of erases), but on a 40 GB drive even 1000 erase cycles is extremely impressive (40 TB of erases.) This, combined with the fact that most people way over-estimate how much data they are writing...

Re:Wear usage?

[Rockoon]

According to the Intel blah Drive Information tool, my SSD ate 400 GB writes since about 5 months.

~1 TB per year, and that thing (if its the 40 GB model) should withstand ~200 TB of erases.

Laptop Backup Times...

[BoRegardless]

ought to be severely reduced if you can pass the info off at those data rates to a similarly fast external drive.

Then if you want to archive to a "slow" spinning hard drive, the external SSD could supply the data at the slower rate of the HD

Re:Laptop Backup Times...

[Neil Boekend]

It depends: You could use a Raid 5/6 NAS and have high disk speeds again. Something like this. The max speed of a thunderbolt connection is 800MB/s so, assuming they are WD black 2TB 3GB/s drives, you can get an average of 109*5 (raid 5 with 6 disks) = 545 MB/s (host to disk speeds. Host to buffer doesn't matter).
That's, effectively, 5 disks to go slightly over their read speed. Then again, you'd get 10 TB (=9.09 TiB) on the external enclosure, while the biggest of these will be 250 GB.

Why not battery backed RAM straight on the bus?

[joesteeve]

With some lazy writing to solid-state-chips.. :D

Yeah, I am dreaming.. Sigh!! :-/

Re:Why not battery backed RAM straight on the bus?

[Anonymous Coward]

With some lazy writing to solid-state-chips.. :D

Yeah, I am dreaming.. Sigh!! :-/

While you're dreaming, dream up an OS written from the start for such modern advances.

Ok, sure, any such idea will be obsolete before you even finish coding the basics, let along get it to a usable state, but somebody has to break the chains that bind us!

Re:Why not battery backed RAM straight on the bus?

[edmudama]

The DDRDrive X1 almost fits your design. It's not on the memory bus, but on the PCI-e bus as a storage device. Bit pricey though per gigabyte.

Re:Why not battery backed RAM straight on the bus?

[joesteeve]

Checked it out. pretty cool :)

But still, 'persistent' memory gives too many opportunities.. I've been dreaming since morning about it. :D

Re:Why not battery backed RAM straight on the bus?

[joesteeve]

While you're dreaming, dream up an OS written from the start for such modern advances.

I am actually.. :D But well, can only dream.. sigh!

Just imagine:
- A power-cycle would leave the machine exactly where it was. Well, that would be a problem with buggy software. But, for such cases we could have a hard-reset of some sort..
- Absolutely no waiting for shutdown/power-up/etc.
- Stuff like 'object serialization', 'state persistence', 'configuration persistence' etc.. would just be there. We wont have to deal with the whole lot of parsing, etc. nonsense.
- Application startup would be blazingly fast.
etc. etc.

Yeah, those chains need to be broken. The 21st century feels like stone age :-/

Re:Why not battery backed RAM straight on the bus?

[joesteeve]

We actually did consider a similar option for an embedded system solution. A few MBs of "battery backed" RAM apart from the normal volatile memory. But well, the client ran away.. so .. :-S

Re:Why not battery backed RAM straight on the bus?

[mabhatter654]

That's how expensive SCSI controllers work. They have big read/write cashes of RAM to speed transactions and the RAM has it's own battery should power blink. That allows the captured transactions to be written to disk or recovered from journal on the next IPL. In today's world, that's a lot of hardware to throw at a simple problem though, definitely not for the "consumer" or even "prosumer".

Re:Why not battery backed RAM straight on the bus?

[symbolset]

We are onto flash-backed write cache now. No battery, just a supercapacitor. On power fail it writes to the flash and shuts down, so no battery to run out.

Re:Why not battery backed RAM straight on the bus?

[m50d]

- A power-cycle would leave the machine exactly where it was. Well, that would be a problem with buggy software. But, for such cases we could have a hard-reset of some sort..

- Absolutely no waiting for shutdown/power-up/etc.

You can get that nowadays with ACPI S3 suspend if you leave your machine plugged in. It's awesome, and kinda freaky.

- Stuff like 'object serialization', 'state persistence', 'configuration persistence' etc.. would just be there. We wont have to deal with the whole lot of parsing, etc. nonsense.

Please, no. Haven't we learned enough from .doc about the stupidity of trying to store data by dumping memory out and reading it back in?

- Application startup would be blazingly fast.

What would battery-backed ram do to help that?

Side note: most modern systems can fit a complete OS into RAM. You can tweak a live-cd linux or similar to do it. It does run very fast, but you find yourself missing all that "bloat", or I do - most of the modern resource-consuming features are actually quite useful.

Re:Why not battery backed RAM straight on the bus?

[Anonymous Coward]

Check out the old gigabyte iRam stuff. Half the performance, but 1/20 the price if you have some old ddr1 ram unused. You can raid them too via sata :)

Re:Why not battery backed RAM straight on the bus?

[joesteeve]

- Application startup would be blazingly fast.

What would battery-backed ram do to help that?

We can completely skip loading the application and its data parts from a slow disk, since all of them are already addressable directly in memory. We only have to deal with the 'initialization time' taken by the applications.

3rd generation X-35M?

[scheme]

What happened to the 3rd generation SSDs that Intel was supposed to release this month? They were supposed to be using 25nm flash and offer roughly twice the space for the same price as the G2 drives. Using a new controller and upgrading things a bit seems a poor substitute for that.

Re:3rd generation X-35M?

[Anonymous Coward]

If Intel was the only 25nm producer then they wouldn't simply double the space. They'd increase the space or lower prices just enough to beat the competition, and then keep the remaining, fatter profits for themselves. I mean, that's how I'd do it. But maybe things work differently where you live.

Re:3rd generation X-35M?

[Anonymous Coward]

Educated guess: Rumored issues with Intel 25nm Flash and/or new gen Intel controller. One rumor indicates that Intel decided very late in the game that they could not release another SATA2-based SSD drive with competition already going with SATA3 (6Gb/s) so the "postville refresh" G3s got deep-sixed and this Marvell-based product is a temporary solution until Intel can get their own "G3.5" stuff ready - and since it is a major redesign, that is 6-9 months away at best.

So no low price (per GB) SSDs anytime soon.

for McAfee

[Anonymous Coward]

Obviously the extra speed was simply to ensure decent performance for McAfee.

$584 for 256GB

[Joe The Dragon]

$584 for 256GB

Re:$584 for 256GB

[Sabriel]

$584 for 256GB

That's $584 for 250GB, in lots of 1000.

Re:$584 for 256GB

[TheLink]

AFAIK this new Intel SSD is using the same Marvell controller as Crucial's RealSSD C300, which goes for about USD490 for the 256GB drives in lots of 1 unit. Just look at Amazon, or even Crucial's own site.

So I'm not sure what makes the Intel drive better than the Crucial one which has been around for many months (and gone through some pains and fixes... ).

A supposedly nerd site like Slashdot linking to low-info press release or marketing article on Computerworld is stupid.

Re:$584 for 256GB

[mabhatter654]

that's not awful.

The point of these is to feed devices like Drobo anyway. So three would get you 460GB with one failure. That would also nicely fill up a Thunderbolt connection to one or two computers with speeds and latency way faster than the built in drive.

Re:$584 for 256GB

[Lennie]

Software/tuned hardware is what makes the difference.

Intel M25, etc. was better than it's competitors at the time because of the software, because it's performance did not degrade over time.

Re:$584 for 256GB

[TheRaven64]

There are very few workloads where that would be sensible. In most cases, you'd get almost the same performance, much more capacity, and a significantly lower cost, by using one 256GB SSD and three 1.5TB hard disks in a RAID-Z + L2ARC configuration. Your data is checksummed on disk and is distributed across two drives, with the SSD being used for level two cache (with RAM being used for level one cache).

Re:$584 for 256GB

[rwa2]

I would still stick with the recommendations from: http://www.hardware-revolution.com/best-hard-drive-best-ssd-december-2010/

For the same price or less, you can build a bigger / faster / cheaper RAID of OCZ Vertex2 SSDs. And you wouldn't even have to upgrade your motherboard to 6Gbps SATA 3.0

Re:$584 for 256GB

[toddestan]

I have a first generation X25-M and performance definitely degrades over time. Maybe the G2 drives with Trim fair better, but on my drive the stuttering eventually got so bad I just went back to using a hard disk.

Re:$584 for 256GB

[Lennie]

You do know you can 'wipe' them ? don't you ? You just have to do it every few months, maybe even weeks.

Re:$584 for 256GB

[toddestan]

Yes, I know that. But having to image the drive to something else so I can wipe it, then restore the image gets bothersome after a while. Naturally, the SSD is the boot drive, and Intel's erase utility is of course for Windows, so there's the extra fun step of having to remove the drive from the PC and using another PC (that runs Windows and has a spare SATA port) to actually do the wipe. I found it easier to just restore the image to a standard HDD then move on.

Finally

[Anonymous Coward]

A drive that can run McAfee properly.

Wonder why Marvell chip

[RightSaidFred99]

I wonder why they outsourced their controller chip, previous incarnations used an Intel controller.

Re:Wonder why Marvell chip

[pipedwho]

I wonder why they outsourced their controller chip, previous incarnations used an Intel controller.

Marvell acquired some of Intel's sub-divisions (including their wireless and embedded groups) - this happened a couple of years ago when Intel decided to focus heavily on their main processor and support chipset lineup. Given that, there is very likely a tight integration between the two companies, so no real reason for Intel to duplicate work and design their own controller chip.

Re:Wonder why Marvell chip

[RightSaidFred99]

They already had a controller chip, and the Marvell split off was well before their SSDs. Could be some Marvell ties there somewhere maybe, my guess is SSD controllers just became commodity (those things are coming out like every month now) and performed well enough (they didn't a few years ago) to use now.

Re:Wonder why Marvell chip

[Ster]

Where are you seeing that they're using a Marvell controller, I didn't see that in any of TFAs?

Re:Wonder why Marvell chip

[Anonymous Coward]

Where are you seeing that they're using a Marvell controller, I didn't see that in any of TFAs?

Look again. The Intel spokesperson says they used Marvell because it was available. Might have been on the second page...

IOPS?

[Weaselmancer]

IOPS?

It's important to know.

Only makes sense

[m.dillon]

Now that 6GBit/s SATA ports are becoming commonly available on motherboards it's only natural that SSDs follow. Normal HDs can't really take advantage of a 6Gb/s port but SSDs can. These high speed ports will make port multiplier enclosures more useful as well.

There's certainly a lot of use for this sort of thing. SSDs can already replace far more expensive DRAM (and the far more expensive motherboards needed to support lots of DRAM) for numerous problem sets, including mostly-read database accesses. This will significantly reduce the cost of server architectures for large swaths of problem areas.

What we are seeing is a natural progression.

-Matt

Re:Only makes sense

[jrobot]

A modern DDR3 part at 800 MBit/s x 16 bits/part = 12.8GBit/s
Granted, you add a lot more complexity to your board, but DRAM is not going away

Re:Only makes sense

[afidel]

Try 32GBytes/s for a dual Nehalem system with DDR3 1333 pieces in banks 1 and 2 on each CPU.

Re:Only makes sense

[Neil Boekend]

With the limited write cycles for SSD DRAM will not go away. Could you imagine your RAM to crash permanently after a long weekend of gaming?

Re:Only makes sense

[Arrepiadd]

He's just saying "future servers can have huge SSDs and small amounts of RAM instead of 128 GB of ram to store that database you need fast access to", not that SSDs will replace RAM entirely.

Re:Only makes sense

[m.dillon]

All our current servers use a small 40G Intel SSD to cache copies of filesystem meta-data, plus data for often-accessed files and databases. Since the data sets being served are far, far larger than the ram in the machine this leads to a massive improvement in efficiency and a massive reduction in the load on the hard drives. Actual write activity is minimal, the SSDs will easily last 20 years at their current wear rate.

It all comes down to the workload. A very large swath of workload types can use SSDs RIGHT NOW without shredding them to bits, so it's pretty stupid not to use them when a measily little $110 Intel SSD saves you $2500 worth of dram, server motherboards, not to mention a massive savings in power costs by not having stuff 32G of ram into a server.

Some people just don't get. It's ok, I don't take it personally.

-Matt

6Gbps throughput is impossible

[Anonymous Coward]

With 8b/10b encoding, throughput absolutely cannot be more than 4.8Gbps (600MB/s).

Re:6Gbps throughput is impossible

[omfgnosis]

Oh. Er. Why?

Re:6Gbps throughput is impossible

[Anonymous Coward]

As said, 8b/10b encoding You send more than just the data itself across the link - and that uses some of the available (1.5 / 3.0 / 6.0 gbps) bandwidth.

Comment removed

[account_deleted]

Comment removed based on user account deletion

is it 6gb/s or 500mb/s ?

[Latinhypercube]

is it 6gb/s or 500mb/s ? Make your fucking mind up.

Re:is it 6gb/s or 500mb/s ?

[Neil Boekend]

The SATA speed is 6Gb/s. The disk speed is 500MB/s. You can write to the disk itself at "only" 500MB/s. Since SATA doesn't have that exact speed they chose the next best thing: the 6Gb/s (=768 MB/s) SATA connection.

Re:is it 6gb/s or 500mb/s ?

[CastrTroy]

Seems unimpressinve when compared to the PCI-E cards that are out there. For instance check out this PCI-E SSD. Sure they're expensive, but they're much faster then you will ever get over SATA. To me, it makes almost no sense to be continuing down the SATA route and touting faster speeds when we already have a much faster way of transferring data.

Re:is it 6gb/s or 500mb/s ?

[drinkypoo]

It does seem more than a little daft to not simply come up with a PCI-E-for-disks standard, and give us a nice little flexible PCI-E cable. However, since SATA disks are supposed to work on SAS I guess there's still some reason for SATA to exist...

Re:is it 6gb/s or 500mb/s ?

[blackraven14250]

Did...did you even look at the price before you started typing...????

Re:is it 6gb/s or 500mb/s ?

[CastrTroy]

There is cheaper stuff out there, with less performance, although it still blows SATA out of the water.

Another breakthrough in marketing the next tier..

[Anonymous Coward]

Same old superior specifications at unsaturated retail market prices...its always more worth buying, yet again, and again. Incrementally strung out technology, as usual. Social engineering, not technology, turns Prometheus into a bic lighter.

ugly opportunity for malware

[SethJohnson]

If one of these is in your home machine, you likely won't kill it. If you think you might, then you should already have practices in place to deal with disk failure.

How about a vicious piece of malware? Could a piece of code be written to circumvent the wear-leveling algorithm and carpet-bomb your SSD with repetitive writes so that it's worthless overnight? Could be a real PITA in cases like the Macbook AIR where the SSD drive is built into the mobo. It's not a case of just paying for a new SSD to replace. On the plus side, this type of hardware failure just limits additional writes to the device. You should still be able to retrieve your data.

Seth

Re:ugly opportunity for malware

[Neil Boekend]

The same piece of malware would destroy the contents of an HDD, since it overwrites all sectors. How often do you see that?

Re:ugly opportunity for malware

[dargaud]

Yeah, I wrote some bootloader code recently to write a kernel to a flash without wear leveling (very low level embedded stuff) and I was really nervous the first time I ran it: if I'd goofed the code and written to the same cell at each iteration, it would have taken about one second to fry the flash !

Re:ugly opportunity for malware

[T-Bone-T]

From what I've read, it would take constantly overwriting the entire drive continuously for years to cause it to fail.

Re:ugly opportunity for malware

[Yvan256]

I don't know about the first model(s), but the latest MacBook Air uses SSD modules, it's not built into the motherboard.

Re:ugly opportunity for malware

[Anonymous Coward]

He's not talking about the contents - he's talking about the actual drive itself. Overwrite the SSD so much that the drive itself can literally not be written to anymore. (Basically, "bricking" the SSD.)

No new SLC?

[afidel]

It's been almost 30 months since the X-25e was launched and now the max 64GB capacity looks pathetic next to the competition so when will Intel launch their next generation SLC based SSD?

no prob under ubuntu

[Anonymous Coward]

on my uncles computer (i set it up with ubuntu as he doesnt play computer games and mainly just torrents movies) he's filled another TB so another was installed and i benchmarked it. A recient seagate sata TB and it was reading at over 90MB/s on his P4 with an ancient sis mobo from about 2002, i think it was about one of the first sata mobos on the market. so no problem with ubuntu, and no problem playing h264 on this ancient system with an agp ati 9550

i think the parent is full of it. im pretty happy using ubuntu on several computers including this laptop which is more recient and while xp cant even be made to boot with the esata ports enabled ive never seen any problems with functionality or speed with ubuntu, hotswapping esata drives, and dumping TB's from through one esata port and out another, try that with usb docks and you'll be waiting for a week. unfortunately xp blue screens with the esata enabled in the bios. so if i want to use windows eg for to run the tax pack program to lodge once a year, i have to enter the bios. and i returned it for a hw warranty issue and the tech agreed that xp just couldnt be made to work with esata on this laptop, a Dell Latitude E6400.

Re:no prob under ubuntu

[ls671]

I have never heard of "sata mobos". I heard about motherboards that had a SATA controller hardwired into them. Using the right driver for that specific controller might help in gaining speed.

http://slashdot.org/comments.pl?sid=2016546&cid=35345062

It uses a Marvell controller, not Intel controller

[MojoStan]

Having been a somewhat early adopter of SSDs, I got bitten a couple of time by the JMicron bug

Everything I've read so far suggests that if you are buying SSDs you want to go with Intel.

Note that this new Intel SSD is the first Intel-branded SSD that uses a non-Intel controller. It uses the same Marvell controller used in the well-regarded Crucial RealSSD C300.

I've also read about Intel's great combo of performance and robustness, but that reputation is mostly a result of Intel's controllers. JMicron, a manufacturer of SSD controllers, got its buggy reputation from early JMicron-based SSDs. Marvell's controller performance has been proven in many reviews, but Intel's "endorsement" gives me more confidence in Marvell's reliability and robustness.

Don't forget about Sandforce/OCZ

[0111 1110]

Sandforce has already announced its new sata3 controller. On paper it looks like it will have much faster sequential writes than Intel, but it sounds like it will also have a shorter lifetime and shorter data retention times due to the use of 25nm NAND. Intel is wisely sticking with 34nm. It may be more expensive to manufacture, but is superior tech. I can only hope that OCZ changes their mind and decides to at least offer a more expensive 34nm version. OCZ won't be shipping their Vertex 3 drives until Q2 so Intel will have a big head start in the market.

The NAND industry seems to be doing its best to encourage ignorance on the disadvantages of smaller process sizes from the consumer POV and the ignorance seems to be widespread. Getting the facts on this issue can be a bit difficult. Here is a good thread on the topic.
http://forums.anandtech.com/showthread.php?t=2142742

The following post sums it up better than I could. Note his point about data retention times as well. That is a point that is often ignored when the focus is solely on write cycles.

As flash cells are shrunk, they become less good. This is a fundamental feature of the technology. The overall volume of the cell becomes smaller, so less electrons can be stored in the cell (so the signal picked up by the electronics is weaker and less clear, so you get a higher error rate) and the insulating barriers around the cell must be made thinner, in order to save space - allowing the electrons to leak out of the cell more easily (reducing power off data retention time). The thinner insulation also wears out more quickly (reducing life cycles)

It's difficult to define a 'fundamantal' limit for flash, because it may be possible to work around poor performance, and as yet unknown new manufacturing techniques and semiconductor materials may be developed. However, it has been suggested in the scientific literature that 18-22 nm, is the realistic limit. Beyond that, the performance/reliability/lifespan of the flash would be too poor, no matter how much wear levelling, and how sophisticated the ECC codes were.

Enterprise grade SSD flash, will need higher specifications than flash for toy cameras. Enterprise applications are unlikely to tolerate 18 nm flash with 100 write cycles and one lost sector per 100 GB of data stored. However, this probably would be acceptable for toys or throwaway devices.

Some more coverage of the topic:
http://techon.nikkeibp.co.jp/article/HONSHI/20090528/170920/

NAND Flash memory quality is also beginning to drop. Chips manufactured using 90nm-generation technology in 2004-05, for example, were assured for about 100,000 rewrites and data retention of about a decade. As multi-level architecture and smaller geometry are introduced, quality is showing a sharp decline. The 30nm 2-bit/cell chips expected to enter volume production in 2009-10 may well end up with a rewrite assurance of no more than 3,000 cycles, and a data retention time of about a year. The first 3-bit/cell chips are hitting the market now, with only a few hundred rewrites.

http://hardforum.com/showthread.php?t=1502663

Flash memory works by trapping electrons. Over time these electrons leak away, until the charge is too small for the data to be read any more. With smaller feature sizes (34 nm instead of 45 or 65 nm) this leakage is more significant and fewer electrons can be stored per bit, thus the time during which the stored value can be maintained is decreased.

http://www.corsair.com/blog/force25nm/

Re:Don't forget about Sandforce/OCZ

[loosescrews]

I can only hope that OCZ changes their mind and decides to at least offer a more expensive 34nm version.

Is a 32nm version good enough for you? They are going to offer one of those called the Vertex 3 Pro. Source: http://www.anandtech.com/show/4186/ocz-vertex-3-preview-the-first-client-focused-sf2200/2

Re:Don't forget about Sandforce/OCZ

[0111 1110]

Is a 32nm version good enough for you?

Yes. Thanks for the link. A lot could change between now and Q2, but that is nice to see. I suspect they would reserve the right to change to different memory at any time. I will keep my fingers crossed that I will be able to purchase a 32nm SF-2000 series drive that maxes out the sata3 interface in both sequential reads and writes. Still, unless consumers become more aware that smaller process size is a bad thing for NAND SSDs I think all of the manufacturers will eventually be peddling NAND at 25nm and beyond. Their financial motivation is just too strong.

Re:Don't forget about Sandforce/OCZ

[TheDarkMaster]

Thanks for the info :)

Re:Don't forget about Sandforce/OCZ

[Anonymous Coward]

Anyone who thinks that Flash cells leak electrons through insulating barriers around the cell, and thus loses information, is not qualified to comment on flash technology. Electron leakage at that scale is a property of DRAM, which therefore needs a refresh timer and permanent power. But Flash stores bits for years, not milliseconds.

Yes, on a glacial scale, electrons will leak out the floating gate of a flash cell. But there's a reason we talk about the write cycles of Flash: they're far more damaging to the floating gate than the theoretical tunneling of electrons. Writing happens at 10-13 Volt, which provides the energy to break structures.

Also, the notion that CPU's don't come down in price after a die shrink is missing the point. They come down all the time, which is made possible because die shrinks definitely push down the costs. But the AMD/Intel competition introduces an extra effect on CPU prices and similar effects occur in RAM and Flash. That means that prices and cost do not go down at exactly the same time. Companies may take a higher profit after a die shrink, if the market allows it, only to decrease prices later when the competition catches up.

Re:Don't forget about Sandforce/OCZ

[Anonymous Coward]

OCZ gets horrible ratings on every site I have been to. Good performance but very poor reliability. They seem to break quite often.

I wouldn't consider them until they get their act together. SSD's in general are very unreliable so I would rather stick to a company that seems like they are trying to mitigate that (like Intel I guess).

Re:Don't forget about Sandforce/OCZ

[0111 1110]

The exact mechanism through which unpowered 22-25nm MLC NAND storage loses track of its data is interesting, but somewhat beside the point. The point is that there is an inherent limit to the technology in terms of data retention and we have now basically arrived there at 25nm.

You seem to be arguing that the loss of the drive itself through the destructive write mechanism is more significant. In many cases this is likely to be true, but many end users may not realize that leaving their drive unpowered for 8 months will result in data loss/corruption. These process shrinks are starting to blur the line between volatile and non-volatile storage.

As for the cost issue I think that cost reductions are a good thing in theory. When there is competition available in the market to drive down prices. Unfortunately, we haven't had such competition in the CPU market for a while. Which is why I used it as an example of how cost reductions don't necessarily result in lower prices at the consumer level. Of course there is competition in the SSD market. So the comparison is a bit unfair. Yet, I remain skeptical. Just because a company could lower their price to increase sales (and start a price war) doesn't mean they always will. As long as none of the competition attempts to get too ambitious with market share prices can stagnate for quite a while. Back when AMD was still competitive (in terms of performance) with Intel in the processor market they attempted to beat Intel on price for a while. Then they just stopped. They decided that the loss of profit margin wasn't worth the additional market share. Both companies happily maintained similar price/performance ratios for quite a while that way.

Hardware TRIM support?

[Golbez81]

What would really be nice is to get TRIM support on hardware controllers. I've been waiting on TRIM drivers now since 2008...

Sod SATA

[Eunuchswear]

Give us fucking SAS already.

Re:Sod SATA

[Anonymous Coward]

You could use an interposer: LSISS9252. Then these should not be far away: Hitachi Ultrastar SSD400S which contain Intel flash.

Re:Sod SATA

[Eunuchswear]

You could use an interposer: LSISS9252.

That's what I do.

Doesn't work reliably with multiple initiators. Two seems ok. more gives trouble. (Disk connected to two machines, and using multipath).

Then these should not be far away: Hitachi Ultrastar SSD400S

Been not "far away" for a couple of years now.

Re:Sod SATA

[picoboy]

Give us fucking SAS already.

Sadly, SAS SSDs will probably never be mainstream (i.e. cheap and ubiquitous). I'm assuming that's what you mean by "Give us fucking SAS already."

On one hand, Romley with its integrated SAS ports may help seed the market for SSDs using SAS protocol.

On the other hand, NVM Express (http://www.intel.com/standards/nvmhci/) is pushing SSD designs toward direct attach on PCIe. The reality is that Intel doesn't care about all the beautiful things that SAS brings, such as long cables, hot plugging and multiple initiators. They're happier just getting the intermediate protocol controller out of the way and pushing the protocol stack into software running on an x86, thereby reducing cost and keeping more margins for themselves.

Don't hold your breath for SAS, unless you want to pay enterprise storage prices.

Re:It uses a Marvell controller, not Intel control

[Lennie]

"mostly a result of Intel's controllers"

Actually, the controller is just the hardware. It is the software running on it which makes it smart. Intel owns the software I would think.

Firmware updates

[torako]

I have an Intel X25-M SSD in my Mac Pro and use an OCZ Vertex 2 in my Macbook. Both work very well, although the OCZ feels faster. One thing to watch out for though, is the way firmware updates are handled: Intel offers a bootable ISO image that will update the firmware very easily and without any hassles whatsoever.

OCZ on the other hand only offers an .EXE tool (32bit only!) that needs an Internet connection and only works if your SSD has an MBR partition style and at least one NTFS formatted partition on it. This makes updating the firmware on my Macbook (which has no Windows partition and uses a GUID partitioning scheme) a nightmare that requires a total wipe + installing a Windows system with Internet access.

Given that firmware updates for SSDs where at some point necessary to ensure data integrity (The Vertex had a bad firmware that produced lockups at some point, e.g.), this whole process just sucks.

Re:Firmware updates

[EmagGeek]

You need to go to the OCZ Technology Forums and get the MAC and Linux tools for updating OCZ Firmware.

Re:It uses a Marvell controller, not Intel control

[jbolden]

Thanks for that tip. I've been buying computer and printer memory from crucial but it was great to red the review on A&D and see how well their drives scored.

disappointing

[buddyglass]

The Intel X25 series was so much better than the alternatives when it came out, I was expecting more from its successor. The 510 doesn't even use an Intel-developed processor. It uses one from Marvel, which will eventually be used by other SSD manufacturers as well. And the 510 is already slower than drives based on Sandforce's newest chips, such as OCZ's Vertex 3. Lame, Intel.

Random reads in movies

[tepples]

And more importantly, your MP3's and movies do not require the random reads and writes which is an SSD's greatest strengths.

Unless, of course, you're producing music and movies. Nonlinear video editing, for example, requires a random read for each transition. And with some of the effects common in the Dadaist video remixes commonly called "YouTube Poop", there can be often several transitions in a second of video.

In addition, your games usually need random reads.

Wow cool...

[Anonymous Coward]

But, why wouldn't they start promoting Thunderbolt by creating drives that can achieve 10+ gb/s.

Wow, they achieved a bandwidth for a standard that came out 4 years ago from a solid state drive.

6Gbit/sec? No

[JimboFBX]

Oh sure this gets 6 Gbit/ sec just like the Crucial C300 got 6 Gbit/sec.

Oh wait, neither got/get that. The drive is spec'd at 450 MB/sec reads (still nothing to sneeze at). 6 Gbit/sec is 750 MB/sec. The 6Gbit/sec is just the maximum speed of the interface.

I would expect a news for nerd site to not make this kind of mistake tbh. It only took a simple Newegg search to find the spec since the second page of the article indicated that this drive is indeed already for sale. Some might say this posting kind of seems like a slashvertisement in disguise.

Anyways, the C300 128 gig is 50 bucks cheaper than the 510 series 120 gig but the intel drive definitely wins on speed if it truely hits those benchmarks in typical use scenarios. That price might drop even further when the C400 (or whatever the next crucial drive is) comes out.