Beta

×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Intel SSD 510 Series 6Gbps SATA Drives Tested

Unknown Lamer posted more than 3 years ago | from the sign-over-your-first-born-child dept.

Data Storage 79

MojoKid writes "Intel recently announced its 510 series Solid State Drive products. The new 510 series SSDs build upon Intel's successful X-25M series of drives by offering native support for SATA 6Gbs interface speeds, with maximum reads in the 500MB/s range and write speeds of approximately 315MB/s — huge improvements over the previous generation. The numbers are in and the new Marvell-infused Intel SSD offers impressive performance rivaling other 6Gbps SATA SSDs on the market but not as fast as the recently announced SandForce 2500-based SSDs like the OCZ Vertex 3."

cancel ×

79 comments

Sorry! There are no comments related to the filter you selected.

Fast enough to get first post? (-1, Offtopic)

Shikaku (1129753) | more than 3 years ago | (#35383262)

Yep.

Burn karma burn!

Degraded Performance (2)

smelch (1988698) | more than 3 years ago | (#35383278)

If it starts at 500MB/s reads, how long does it stay that way before performance may begin to degrade (possibly, in some cases... maybe)?

Re:Degraded Performance (0)

DigiShaman (671371) | more than 3 years ago | (#35383330)

From what I understand, it's not the reads that suffer but the writes over time. Fortunately, the TRIM command (with supporting OS) will do a background scrub to zero out the cells in preparation for the next set of user writes to occur.

Re:Degraded Performance (2)

arth1 (260657) | more than 3 years ago | (#35383508)

TRIM doesn't work with RAID, so unless you are fine with having no redundancy and downtime with drive failure, it won't help.

Re:Degraded Performance (1)

the_one(2) (1117139) | more than 3 years ago | (#35383566)

It works with raid if the card supports it or if the software raid supports it.

Re:Degraded Performance (1)

LordKronos (470910) | more than 3 years ago | (#35383686)

As far as I know, there is no raid setup (hardware or software) which supports trim. Please correct me if I'm wrong (with proof please). I know that about a year ago there was a bunch of fuss when people thought a new intel driver was going to support TRIM on SSD raid setups, but it was an error in the documentation later clarified. Other than that, I've heard nothing more on the matter.

Re:Degraded Performance (1)

Anonymous Coward | more than 3 years ago | (#35383922)

I found this in about two seconds with Google from last year. I'm guessing it exists.

http://www.hardwarecanucks.com/news/cpu/intel-chipset-driver-brings-trim-support-raid-setups/ [hardwarecanucks.com]

Re:Degraded Performance (1)

TimSSG (1068536) | more than 3 years ago | (#35385268)

From your link; http://www.hardwarecanucks.com/news/cpu/intel-chipset-driver-brings-trim-support-raid-setups/ [hardwarecanucks.com] [hardwarecanucks.com] I am not sure which person the quote supports. Tim S.

*** Correction: It should be noted that Intel’s announcement about TRIM support only applies to single SSD’s that are present in a system with a RAID array, and not to a RAID array of SSD’s. ***

Re:Degraded Performance (1)

varmittang (849469) | more than 3 years ago | (#35385882)

I believe what it says is that a drive that is not in the RAID can use TRIM, even if its on the same card or in a system that does have a RAID array. But any drives in a RAID setup can not use the TRIM.

Re:Degraded Performance (2)

LordKronos (470910) | more than 3 years ago | (#35385914)

Good job. Apparently, you bothered to read neither my entire post, nor the contents of your own link. I said there was a bunch of fuss about it a year ago that turned out to be incorrect. And when is your link from? 1 year ago. And in your own link it even says in the VERY FIRST PARAGRAPH that it doesn't apply to RAID.

So bravo, but maybe you should have spent more than "about 2 seconds" researching the matter.

Re:Degraded Performance (1)

Phreakiture (547094) | more than 3 years ago | (#35383606)

TRIM doesn't work with RAID

Looks like it might, actually. [bit-tech.net]

Re:Degraded Performance (2)

LordKronos (470910) | more than 3 years ago | (#35383734)

Nope, that was an error in the documentation:
http://www.intel.com/support/chipsets/imsm/sb/CS-031491.htm [intel.com]

Intel® Rapid Storage Technology 9.6 supports TRIM in AHCI mode and in RAID mode for drives that are not part of a RAID volume.

A defect was filed to correct the information in the Help file that states that TRIM is supported on RAID volumes.

Notice the date on that page.... "Last Modified: 26-Mar-2010", which is a few days after the flood of articles a few days prior (including the one you linked) claiming the driver would support TRIM on SSD arrays

Re:Degraded Performance (1)

slinches (1540051) | more than 3 years ago | (#35384668)

Who said anything about RAID? If you're concerned about TRIM support, don't use SSDs in an array. Individual drives are fast enough for most purposes and if you want redundancy, then just drop periodic images on a spare.

Re:Degraded Performance (0)

Anonymous Coward | more than 3 years ago | (#35386762)

Mod parent up. RAID redundancy is mainly introducing more complexity to your setup, and more likelihood of problems. RAID only protects against data loss from a failed hard disk. This is not a very common occurrence compared to data loss from user error, data loss from memory errors, data loss from controller bugs (which are more common in RAID mode), data loss from OS or driver bugs, etc.

In other words, RAID 1 does not exempt you from making backups and IMO is worthwhile only in very specific circumstances.

Re:Degraded Performance (0)

Anonymous Coward | more than 3 years ago | (#35386802)

RAID only protects against data loss from a failed hard disk. This is not a very common occurrence compared to data loss from user error, data loss from memory errors...

I've had many, many more hard drives fail in my experience than memory failures.

Re:Degraded Performance (0)

Anonymous Coward | more than 3 years ago | (#35384740)

Newer SSDs include garbage collecting controllers which can perform "trim" without even being connected to a controller.

Re:Degraded Performance (1)

toddestan (632714) | more than 3 years ago | (#35393854)

This only works if the drive can read and interpret the file system to know what's stored on it. This maybe will work with RAID 1, but certainly not with any of the other RAID levels.

Re:Degraded Performance (2)

arth1 (260657) | more than 3 years ago | (#35383488)

Read again. It's 500 MB/s maximum read speeds.

I am much more interested worst case speeds, cause that's when you feel a problem.

I'd also like to know the random access speeds, and in particular small random write speed; the typical bottleneck of SSDs in general, and the new 22 nm process in particular.

Re:Degraded Performance (1)

v1 (525388) | more than 3 years ago | (#35383628)

Read again. It's 500 MB/s maximum read speeds.

I am much more interested worst case speeds, cause that's when you feel a problem.

Since it's solid state, there are no seek penalties. Speeds decrease as the reads become more random over the surface of the drive, and in particular when you only want a small amount of data (a few k) where the drive has to read in an entire flash block which could be say, 64k. So the drive is reading a lot more data than its passing to the computer.

I'd also like to know the random access speeds, and in particular small random write speed; the typical bottleneck of SSDs in general, and the new 22 nm process in particular.

Again sequential is going to work better than random since it will be making 100% use of the read/write operation. Writing to a part of a cell for example, requires reading the entire cell, integrating in the changed data, and reflashing it, which is significantly slower than just reflashing the entire cell in one operation.

Cell size isn't going to be too big, and most operations are on a much larger scale. Launching a 2mb binary is going to be almost 100% sequential reading. Writing is often the same way, occurring in large enough blocks to not have any cell overhead. Then you're just limited by the speed of flashing.

Re:Degraded Performance (0)

Anonymous Coward | more than 3 years ago | (#35387684)

> Speeds decrease as the reads become more random over the surface of the drive

If your system is truly performing "random" reads then something has gone terribly wrong...

"Random" has a specific meaning, it doesn't just mean "out of sequence" or "non-linear".

ITYM "arbitrary reads".

Waiting for SSD price drop (1)

DigiShaman (671371) | more than 3 years ago | (#35383288)

Now that that's out, I'm waiting for a price drop on the Intel X25-M 120GB version. It's currently listed as $229.99 on NewEgg.com. I'm going to guess that the real sale will happen once the new Mac OS gets released. At least, that's what I plan on doing.

Re:Waiting for SSD price drop (2)

Shuh (13578) | more than 3 years ago | (#35385508)

I just bought one of those for $229.99! I guess I should have known no sooner than I had installed the drive, Intel would come out with a brand new SSD and everything would be cheaper. Also OS X 10.7 with TRIM support is just around the corner. Oh well. This wasn't my first run-in with premature acquisition, and it probably won't be my last.

Early adopter, huh? (1)

cefek (148764) | more than 3 years ago | (#35388382)

That's the cost of being an early adopter. You should be proud of that, man! How many of *them* never heard of Apple Newton - even though it had just a little too slow processor to handle it's software power? Raise your hand, if you know what it feels to have Gravis Ultrasound just to have it killed with Windows 95, or texting and talking with Matrix Neo's Nokia (old 7110) phone - just to realise that even tho the call answering effect is awesome, the phone wears and starts answering - and rejecting - your phones while in your pocket.

Man, I for one love to be ahead of the curve. And even if it means you got to pay the price for some time... still, it's worth it.

From the article... (4, Funny)

Cinder6 (894572) | more than 3 years ago | (#35383334)

The OCZ Vertex 3 drive you see pictured here is a 240GB model (224MB formatted) outfitted with 16 pieces of Micron NAND flash, totally 256GB—the additional capacity is over provisioned for wear leveling, data protection, and other functions.

224MB formatted? I knew SSDs used some of the space for redundancy, but that's just ridiculous.

Re:From the article... (4, Funny)

natehoy (1608657) | more than 3 years ago | (#35383380)

What, 0.93% of the drive capacity isn't enough for you? I bet you want a whole 1% of drive capacity all to yourself.

Greedy bastard...

Re:From the article... (1)

alienzed (732782) | more than 3 years ago | (#35383722)

Unknown Lamer must have had his head in the Cloud.

Re:From the article... (1)

blind biker (1066130) | more than 3 years ago | (#35383926)

So according to you, 224MB is 0.93% of 240GB?

It's funny because you fail as badly as the TFA itself.

Re:From the article... (1)

vux984 (928602) | more than 3 years ago | (#35384190)

And here I thought it was on purpose. Sarcastic satire...

I could easily imagine a future time when 99.07% of the advertised space on a storage medium was actualy reserved for 'file system / partition overhead, redundancy, wear leveling, error correction, government tax, your operating system recovery partition, secret NSA back door partition, and unlockable future upgrades via micropayments...) leaving 0.93% of the advertised space available for the buyer to actually use.

I admit I already feel a bit gipped buying a new PC with a 64GB SSD, and finding less than 20GB of available space on it, while I watch Windows update download a 1GB service pack that consumes another 5% before I've even finished registering the thing... :p

Re:From the article... (1)

XLazarusX (534555) | more than 3 years ago | (#35384416)

I think you mean gypped, and it's just as bigoted as saying you jewed someone down on a price.

Re:From the article... (0)

Anonymous Coward | more than 3 years ago | (#35385206)

I think you mean gypped, and it's just as bigoted as saying you jewed someone down on a price.

Doubly so, because "gypsy" itself is a misapplication of "Egyptian" to the Romani, who are originally from the Indian subcontinent and now live primarily in Europe.

Re:From the article... (0)

Anonymous Coward | more than 3 years ago | (#35384224)

In the interests of national security, your computer will be devoted primarily to scanning for viruses and detecting network intrusion. If there's any storage or CPU left over, you may attempt to run authorized applications on it.

Re:From the article... (0)

Anonymous Coward | more than 3 years ago | (#35384744)

You mean .093%? One could only hope for .93%

Re:From the article... (1)

natehoy (1608657) | more than 3 years ago | (#35385872)

D'OH! And you'd think I'd be better about that from my slide rule days.

Re:From the article... (1)

JesseMcDonald (536341) | more than 3 years ago | (#35383458)

Putting aside the MB/GB mixup, the larger 240GB figure does not include redundancy internal to the SSD. The drive reports 240GB of usable space; 224GB is what you have left over after subtracting partitioning and filesystem overhead (presumably based on FAT or NTFS). Redundancy, ECC, wear-leveling, etc. come out of the remaining 16GB (256GB - 240GB) not reported to the OS.

Re:From the article... (1)

tweak13 (1171627) | more than 3 years ago | (#35383558)

The drive reports 240GB of usable space; 224GB is what you have left over after subtracting partitioning and filesystem overhead

Actually, it's what you have left over after the power of 10 to power of 2 conversion. Yes, the manufacturers are lying about this with flash drives as well.

Re:From the article... (0)

Anonymous Coward | more than 3 years ago | (#35383952)

There is no lying, just a bunch of software made by people who are bad at math and a bunch of angry but not very bright people defending this shitty software.

Re:From the article... (0)

Anonymous Coward | more than 3 years ago | (#35386928)

There is no lying, just a bunch of software made by people who are bad at math

They're not bad at math. They're bad at the metric (SI) system.

Re:From the article... (1)

toddestan (632714) | more than 3 years ago | (#35393884)

Well, the manufacturers can't even be consistent. Flash memory cards and most flash thumb drives use the powers of 2 definition, The flash chips, if you look up the spec sheets, use power of 2. Yet when it comes to SSDs suddenly they use the power of 10 definition. Their marketing people should be dragged out and shot.

Re:From the article... (1)

JesseMcDonald (536341) | more than 3 years ago | (#35384312)

I admit that the correlation is suspicious (224 * 1024**3 ~= 240 * 1000**3). The difference between 240GB and 224GB could be a result of abusing SI's decimal prefix definitions with non-SI base units like "byte", if one assumes the article was in error regarding 224GB being available after formatting. The actual NAND flash appears to be measured in standard binary units (256GB total, or 16GB per device), which would be consistent with my own experience with raw NAND. Assuming two of the 16 NAND devices are set aside, there would be 224GB or about 240e+9 bytes left for the operating system to use, which the unscrupulous could attempt to pass off as 240GB of "formatted" space.

One the other hand, the difference could actually be due to partition and filesystem overhead. Seven percent of the reported drive size would not be unreasonable for many filesystems.

Either way, the 240GB and 224GB figures do not include the 16GB or 32GB of reserved space, respectively, which is internal to the SSD.

Re:From the article... (1)

GIL_Dude (850471) | more than 3 years ago | (#35383906)

Actually, Intel recommends overprovisioning the drive even further. For example, I have a 160 GB SSD in the notebook I am typing this on. It is provisioned to 128 GB (and formatted to 125 or so). http://cache-www.intel.com/cd/00/00/45/95/459555_459555.pdf [intel.com] . You do indeed see a speed increase by leaving more of the drive "unavailable" to BIOS. We've been using Ubuntu Live CD's and hdparm to set this.

Beautiful Devices (1)

Ancantus (1926920) | more than 3 years ago | (#35383412)

I am quite happy that these SSD are becoming so fast. Hopefully soon personal desktops will come out with one of these as well as a disk HD. These SSDs serve a really neat niche of providing mid-range speed (compared to RAM) while still providing a large enough memory to store meaningful stuff. ZFS is really taking the right idea with smart usage of the speed and storage capacity of SSDs.

ZFS and SSDs [sun.com]

Re:Beautiful Devices (1)

screwzloos (1942336) | more than 3 years ago | (#35383572)

Having a SSD for the OS and high demand applications, with a separate spinning hard disk for low demand generic storage is definitely the way to do it. I've been running a X25-M for my system drive at home for a couple years, and it is far and away the better way to do things. Coming to work and dealing with a one-drive system all day is a noticeable let down at this point, even if the rest of the machine is brand new.

It's too bad a setup like that is still considered rather exotic for typical users.

Re:Beautiful Devices (0)

Anonymous Coward | more than 3 years ago | (#35383612)

Fuck two drives. I hope hybrid eventually takes over.

Re:Beautiful Devices (1)

Low Ranked Craig (1327799) | more than 3 years ago | (#35383846)

The newest iMacs have this as a configuration option - a 256 GB SSD as the primary drive with a 1 or 2 TB secondary.

Re:Beautiful Devices (0)

Anonymous Coward | more than 3 years ago | (#35384542)

These SSDs serve a really neat niche of providing mid-range speed

I can tell you that installing a pair of the X25Ms has done amazing things for my digital audio workstation. I understand that's not a mainstream application, but man has it made a difference.

Typo (2)

Trieuvan (789695) | more than 3 years ago | (#35383428)

Vertex 3 is based on SandForce SF-2200

Suggestion for benchmarks (1)

timeOday (582209) | more than 3 years ago | (#35383432)

They need to start including a conventional 2.5" hard drive in these comparisons. You can say they're in a different class, but these stories are always accompanied by people making direct cost/GB comparisons to hard drives, so clearly they are still competitors.

Re:Suggestion for benchmarks (1)

h4rr4r (612664) | more than 3 years ago | (#35383666)

Why?
Unless you plan to use these in a laptop it does not matter. Heck, many SSDs do not even fit in laptops as they are too tall.

Re:Suggestion for benchmarks (0)

Anonymous Coward | more than 3 years ago | (#35385520)

Say what? 90% of SSDs out there are the same form factor as 2.5" laptop drives.

Re:Suggestion for benchmarks (1)

h4rr4r (612664) | more than 3 years ago | (#35385568)

Maybe, but that still means 1 in 10 at least will not fit into a laptop. Laptop sata drives are not as thick/tall as some/many SSDs.

Re:Suggestion for benchmarks (1)

.tekrox (858002) | more than 3 years ago | (#35388116)

Depends on the laptop too,

For example:
a Macbook/Pro 13" will only fit 9.5mm drives
The 15/17" Macbook Pro fits full 12.7mm drives.

Re:Suggestion for benchmarks (0)

Anonymous Coward | more than 3 years ago | (#35383822)

They stopped doing this awhile back because from the very beginning, it was clear that comparing the two are like comparing ride-on lawnmowers with Ferrari's. Check out tom's hardware: http://www.tomshardware.com/charts/hard-drives,3.html [tomshardware.com] You can't put SSD's along side regular HD's, but casual glances at the charts show that SSD's beat platter's by a large margin in every category, with the notable exception being some of the first generation SSD's that suffered from major slowdowns once they filled up- this has been resolved in every new SSD that I am aware of, and if you are an informed buyer it should not be a problem at all.

I don't know if you have been out of the loop, but the latest generation of drives is almost double the performance of the previous generation's drives. Ironically, this has made me hesitant to buy an SSD even though the small capacity drives are now under $300, and within my "splurge" budget. They still talk about cost per GB because the chasm is still very wide, ridiculously so, and you can't buy SSDs with the same capacity as platter drives.

It's nice to get excited about hardware again though, that hasn't happened since around 2002!

Re:Suggestion for benchmarks (1)

0123456 (636235) | more than 3 years ago | (#35384232)

They stopped doing this awhile back because from the very beginning, it was clear that comparing the two are like comparing ride-on lawnmowers with Ferrari's.

My netbook boots twice as fast after installing an SSD as it did with the old hard drive. I don't think there are many lawnmowers that go half as fast as a Ferrari.

I installed it in there because I wanted to improve boot times, so that's fine, but the only place where the average user sees a benefit from SSDs are bootup and starting applications; if they don't do that much then the benchmark numbers are irrelevant to them.

The whole thing reminds me of when I used to develop video drivers and we were judged on whether we could render an Excel spreadsheet three thousand times a second or only two thousand times a second. Very few people care because most would never do that.

Re:Suggestion for benchmarks (2)

Zan Lynx (87672) | more than 3 years ago | (#35385046)

Bootup, starting applications, using Firefox, actually getting work done like compiling software or scrolling long documents, running backups, creating or unpacking file archives. Application updates run very quickly, as do OS updates like the recent Windows 7 SP1.

There are thousands of human visible IO waits during every work day. A SSD quickly becomes an invisible part of your work flow. And then until it is taken away or you use another machine you won't realize how much you miss it.

Re:Suggestion for benchmarks (0)

Anonymous Coward | more than 3 years ago | (#35386206)

Average users also see gains from swap, desktop search, and battery life.

I think the swap benefits are huge, unless you also have ungodly amounts of RAM. Lots of time and frustration ultimately comes from disk thrashing.

Re:Suggestion for benchmarks (0)

Anonymous Coward | more than 3 years ago | (#35384376)

There's no real way to to it in a single graph unless you use a logarithmic scale instead of a linear one -- the performance difference, especially in numbers like seek time, is that big. No 2.5" platter-based drive can come even close, and even 3.5" Velociraptors are left in the dust.

At this point, I don't care about the cost/GB of SSD; I refuse to own a PC or laptop that doesn't have one. It's just too painful, and you can still buy cheap platter drives to hold the things that aren't so important.

Still waiting on lower prices.... (2)

schnikies79 (788746) | more than 3 years ago | (#35383598)

I haven't spent over $100 for a hard drive in years and I'm not going to start because it's an SSD.

Re:Still waiting on lower prices.... (2, Insightful)

Anonymous Coward | more than 3 years ago | (#35383854)


I haven't spent over $100 for a hard drive in years and I'm not going to start because it's an SSD.

I haven't spent over $5,000 for a horse in years and I'm not going to start because it's a car.

OCZ, meh (1)

Anonymous Coward | more than 3 years ago | (#35383738)

OCZ may be the speed king but they also seem like the failure king. Go read the reviews of OCZ SSD's on various sites, the one common thing seems to be that they all have a short lifespan.

Re:OCZ, meh (3, Interesting)

slaker (53818) | more than 3 years ago | (#35384030)

In my experience, the average RMA turnaround time for the four drives I've so far sent back to OCZ is not quite six weeks. It's not just that they die, it's that they die and the after-sale customer service is atrocious.

My single experience with getting an Intel SSD replaced was a three day turnaround from the day the defective unit shipped out.

Re:OCZ, meh (0)

Anonymous Coward | more than 3 years ago | (#35387976)

Agree. Mine vertex2 died after 20 days of use in a notebook. Ocz forum is FULL of dying disks problems. I was scared when i started reading their forum.

Will never buy ocz ssd disks until they start to actually care.

The other issue with ocz ssd is a firmware. Read their cigno to see long list of known issues. That list is becoming longer and longer.

Important question... (0)

Anonymous Coward | more than 3 years ago | (#35383868)

When you delete stuff off of it, does it really delete in a manner that the cops can't read it with forensics?

Re:Important question... (2)

hoggoth (414195) | more than 3 years ago | (#35384104)

Yes and no.
Deletions are more non-deterministic than with traditional hard drives.

When something is deleted, or even overwritten and scrubbed the old information is still there because the SSD's wear leveling firmware moves the write to a new physical block. This means you can't reliably erase or scrub anything on an SSD. Even if you fill the entire hard drive (DBAN) the SSD may be using the over-provisioning and not erase the old data.

However, after a block is deleted and a TRIM has been issued, the SSD firmware will take it on itself to quietly write all zeros to the block in the background when it has the opportunity (probably when the drive is idle for a fraction of a second). This is because to write on an SSD the device must erase then write the block. Doing this on demand would slow down writes, so the SSD erases TRIMed blocks when it can so on-demand writes can go faster.

So the answer to your question is both (1) You cannot reliably erase data on an SSD and (2) You cannot reliably recover deleted information on an SSD.

Re:Important question... (3, Informative)

Ungrounded Lightning (62228) | more than 3 years ago | (#35384914)

When something is deleted, or even overwritten and scrubbed the old information is still there because the SSD's wear leveling firmware moves the write to a new physical block. This means you can't reliably erase or scrub anything on an SSD. Even if you fill the entire hard drive (DBAN) the SSD may be using the over-provisioning and not erase the old data.

Except that the ATA "sanitize" commands are supposed to do whatever is necessary to render the data permanently inaccessible (even by disassembling the device) before returning.

And just 18 days ago we had an article on a paper from UCSD [slashdot.org] where they tested nine ATA SSDs and of them:
  - 4 worked correctly,
  - 1 was encrypted so they couldn't check it with their methodology,
  - 2 didn't work correctly, leaving some data accessible,
  - 1 LIED, saying it succeeded but doing nothing to erase the data, and
  - 1 didn't implement the command.

I suspect the grandparent posting was asking which category this drive will occupy.

Re:Important question... (1)

compro01 (777531) | more than 3 years ago | (#35384952)

You can blank an Intel drive quite easily. They support the ATA secure erase instruction and their SSD toolbox provides a simple way to issue it.

Only problem is it only works on a whole-drive level, not individual files.

Looks good for another upgrade in 5 years. (1)

cstanley8899 (1998614) | more than 3 years ago | (#35384304)

All this new hi-speed data transfer technology should be just ripe for my next upgrade. 1 year ago I built a Quad Core system with all the standard SATA II stuff. I can see in about 5 years that all these super fast drives and Thunderbolt may be standard. But as of now I am satisfied with what I have.

Re:Looks good for another upgrade in 5 years. (0)

Anonymous Coward | more than 3 years ago | (#35384550)

That's great, we really care. What is your brand of TP, please tell us while we wait here breathlessly anticipating your response!

Re:Looks good for another upgrade in 5 years. (0)

Anonymous Coward | more than 3 years ago | (#35384818)

That's great as long as what you do is CPU bound. Otherwise you blew a lot of money on worthless unused processing power when you could have made a similar investment in something that would speed up everything you do. All the CPU power in the world doesn't mean squat if it spends 99% of its time waiting on data to process.

what's the point? (1)

markhahn (122033) | more than 3 years ago | (#35384698)

I would never interfere with someone else's fetish, but what is the point of getting excited about SSD speeds? how many people who buy them have any sort of thought-out reason to get them, rather than alternatives? what's your workload that works if you can write at 315 MB/s, but fails if you're limited to a measly 250?

in general, the SSD market seems driven by fetish, and that's just fine. the whole auto market is fetish-driven. and apple, too ;)

what I wonder, though, is if there's someone out there designing, say, a complex website with load balancing, HA-failover, frontend/backend/storage specialization, etc who just needs a 315 MB/s storage device to win. it seems like current standard practice (inmemory nosql, fast private networks, etc) make pretty much any performance concern just a matter of turning the knob - usually just adding some nodes. but if that's the case, then there's little importance to 315 vs 250 MB/s.

Re:what's the point? (1)

PTBarnum (233319) | more than 3 years ago | (#35384750)

The apple market is fetish driven? That makes me view the popularity of Granny Smith in a whole new light.

Re:what's the point? (1)

slinches (1540051) | more than 3 years ago | (#35385100)

what's your workload that works if you can write at 315 MB/s, but fails if you're limited to a measly 250?

While it may not fail at the ~150MB/s that the striped HDD array I'm using gets, but doubling the read/write speed will cut the solution time nearly in half for much of what I do. I run FEA that is limited by continuous writes and reads in the 50-60 GB range. One of these SSDs would allow me to get almost twice as much done. Even better would be a PCI-e SSD that has write speeds approaching 1GB/s, but those would be tough to justify since they still cost more than the rest of the Workstation.

In a desktop, the read/write speed is less important than the random access times. I have a SSD as a boot drive on my home desktop and it makes freshly opened programs feel like they were already loaded in RAM.

Re:what's the point? (0)

Anonymous Coward | more than 3 years ago | (#35385114)

There are plenty of harddrive applications besides servers. For instance video recording, editing, and music production. RAIDS are great, but with these drives you can have a RAID 0 speed along with redundancy, using less drives, for one. Is there a huge difference between 315 and 250? Yes when recording uncompressed 1080p video from a camera, or playing it back, and you want to be sure you have some headroom. I haven't myself messed with them yet, but when the price is right to use them for data storage and not just a small boot drive, it will change a lot of things in terms of what can be done on a budget and without an external array.

Re:what's the point? (1)

Sabriel (134364) | more than 3 years ago | (#35385478)

It's not the increased bandwidth so much - though that helps too - as it is the decreased latency. An SSD doesn't have to wait for a platter to spin up or around at subsonic speeds before it can read the next chunk of data, and your IOPS don't nosedive from multiple tasks competing for priority over where the drive head has to move next.

Re:what's the point? (1)

CastrTroy (595695) | more than 3 years ago | (#35386392)

I'm much more excited on speeds from products such as PCI-E Based SSDs [newegg.com] . That one has 740 MB/s, with sustained write speeds of around 600 MB/s. I don't own one, and i'm not in the market for an SSD, but if I got one, I would definitely be going this route.

Speed is great, data integrity is more important (0)

Anonymous Coward | more than 3 years ago | (#35384954)

The great thing about Intel is that while their SSDs haven't been top of the line in terms of speed for a while, they have always been top of the line when it comes to data integrity. All the speed in the world doesn't mean squat if your drive sufferes from frequent data corruption (I'm looking at you SandForce and Indilix)

OCZ has tried three times to get me a drive (indilix based) that doesn't suffer from massive data corruption (in any of 2 different machines I've tried it in). My new Intel drive works perfectly in both machines.

Other coverage (4, Informative)

Nemyst (1383049) | more than 3 years ago | (#35385178)

As always, I find AnandTech's coverage [anandtech.com] to have a few nuggets of information that most other publications don't. It's well worth a read, particularly for those curious about TRIM performance and degradation over time. There's also a nice page on average reliability around different SSD manufacturers.

Anand concludes by saying that the 510 is one of the fastest drives around today, but only worthwhile on a 6Gbps interface. He points out that they've swapped excellent random performance in the older X-25 for excellent sequential performance in the 510. The Vertex 3 still comes out on top, but the 510 should be more reliable. If OCZ can make their new drives more reliable, Intel will have an uphill battle to fight.

Then there's also the other SSDs, since we've only heard from OCZ and Intel thus far.

Re:Other coverage (1)

AmiMoJo (196126) | more than 3 years ago | (#35388174)

What seems to let other manufacturers down is a lack of proper testing. So many products have been found to have major performance issues long term or in particular configurations that it makes it hard to consider a drive from OCZ et. al. over one from Intel. Intel may not have the best performance benchmark wise but their SSDs have proven to be stable and reliable.

Intel don't do asshat things like changing the type of flash memory without changing the model number either. It is a shame they have gone for the headline sequential performance though, since for OS and application use random is far more important.

Check for New Comments
Slashdot Login

Need an Account?

Forgot your password?
or Connect with...

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>