Any signifcant SSD changes imminent? Good time to buy?

Hello, I'm not sure where this thread went, but I did end up buying an Intel X25-M. Everyone's feedback was useful, and I did end up reading a bit more about SSD's.

I fully agree that an SSD purchase does depend on situation. And the decision is not as clearcut from either the pro-SSD or pro-HDD perspective. SSD's are a relatively new consumer technology, many early adopters have been burnt (e.g. early jmicron related stuttering) and most pc users should stick to mature HDD technology.

Buying an SSD is still a risk - as a consumer technology, it is still not mature, and expensive. "But" I think the major limitations and risks are now understood and major companies are actively addressing them with firmware upgrades, O/S adjustments or new features, and price discounts. Some of the results are seen in the feedback on this board. No one is aware of any seriously exciting new breakthrough SSD products, so I see prices decreasing "gradually" rather than dramatically decreasing.

So the risks are a lot lower now, but still there - but it's my hard earned money to risk and my hobby. So thanks to the efforts of the "early adoptors" and feedback, I'm now far more confident I'm not buying either a significantly flawed technology, or a lemon of a product, when I go with the Intel X25-M today.

Good choice. When I set out to buy an SSD, I was thinking... Intel or OCZ, so one thing that I thought was how much I'd be paying. Buying 2 x 60 GB Vertexes was too expensive, and the 80 GB Intel would offer me just the amount of space I was looking for (for starters, anyway). I'm not disappointed with my purchase, and will be glad to add another one in later, when prices drop at least 50%.
 
Let us know how you've liked it after you've had the chance to play around with it!
 
Good time to buy?

That's the original question of this thread, after it got derailed by pc1x1's attempt at justifying his WD Velocity Raptor hard drive purchases instead of going for Solid State Disk instead.

(Off topic but I need to say it: pc1x1, you made a poor decision and you're wasting your time in the futility that is trying to justify your purchase to others. Folks that are long time forum users have seen the same exact behavior from many others before and all the justifying and reasoning doesn't lead anywhere until the person realizes or has the truth smacked into this face that his decision was an unwise one. Stop it, you're wasting everyone's time, get over the past and think how you can fix things. At this point if you have a chance return the hard drives and get a good SSD instead or eBay the drives and hope to recoup your costs that you can put away towards a good SSD now Intel X25-M or OCZ Vertex possibly.)

Back to the question. I've been progressively upgrading the storage in my old computer for a while now, starting with WD Caviar 250GB 7.2K, to 320GB, 500GB, and now WD Caviar Green 1TB 5.4K. I've been using the same OS install since I believe Windows XP SP1 came out and been doing progressive updates every since. I've been had automatic defragmentation on my drives for a long time and I have enough space to be constantly at 0% fragmentation, but all the updates have shifted all the files all over the platters that I'm guessing most of the files for each application are peppered all around the drive. I'm hoping that Windows XP's PreFetch is doing it's job but frankly without NCQ support on my old nForce 4 chipset I'm a bit behind the curve.

Just recently I helped my friend build an Ultimate HTPC and I recommended that since he's building a whole new system that he should jump on the SSD bandwagon. He took my advice and bought the Intel X25-M SSD 80GB drive back when it cost $363 USD ($4.5375 per GB). I updated it for him to the new 8820 firmware to resolve the issue with the old version of those drives that would slow down as they fill up in capacity. He's been using it every since.

Just recently I got a little sick of my system being unresponsive every time I wanted to open up a new application while using another one, or even starting up a new command prompt, and having to wait 5-10 seconds to open up a simple TextPad editor. I finally looked at the drive and it dropped to $314 USD ($3.925 per GB). I've been reading all the relevant AnandTech Storage reviews and updates so the OCZ Vertex drives came into my view also. Looking at the OCZ Vertex 120GB for $375 USD ($3.125 per GB) it is a very good drive for the price.

However comparing the Intel X25-M and the OCZ Vertex I feel that the price of the latter is a little high for the low preformance if offers for random read and write speeds as compared to the Intel. The usage pattern for the operating system drive is mostly composed of random reads and writes so this is the most important metric of all when it comes to system responsiveness. The emphasis that people put on the super fast sequential read and write numbers is completely wrong for the vast majority of usage patterns for any hard drive, unless you spend your entire time writing and reading entire hard drives full of data such as uncompressed video or audio files.

I do have multiple WD Green 1 TB 5.4K hard drives full of media files and frankly on my gimped PCI SATA I controller I hit 44 MB/s sustained speed when I do a while drive copy or large media files copy, but I do that only once every few months so I don't care if it taxes 5x as long than it would with an SSD or not since I just queue up the copies with Total Commander and let them rip while I do other work.

Now the small difference in price $69 USD for the drive ($0.80 per GB) or 20% difference in price looks like it would favor the OCZ Vertex drive, but when you consider the random read and write performance of this drive you see that it is almost 958% slower than the Intel.

18643.png


The random read results are very good from the Vertex drive, since it is only 30% slower than the Intel drive meaning that the Intel drive reads more data in that time.

18641.png


But when you look at the latency you see that it is ~40% slower there also. That means that it is nearly half as responsive as the Intel drive. Close to being twice a slow.

18642.png


Even with the newest firmware update for the OCZ Vertex drive it is still 350% slower in random writes than Intel.

18704.png


If you consider the fact that the SSD will be used mostly for random read and writes during the OS boot and usage cycles then you will realize the importance of this metric. Then if you look at the fact that for a mere 20% price difference per gigabyte that you pay for the Intel drive over the OCZ Vertex you get 350% more performance in the one metric that counts you suddenly realize how much superior the Intel drive is right now. Where else can you get that kind of a performance boost for a product with such a small difference in price.

I know that looking at aggregate performance scores from SYSMark 2007 the difference between the OCZ Vertex and Intel X25-M get a lot closer to each other, but you have to consider that some of the parts of this benchmark such as the Video Production and 3D part heavily favors sequential read and write access where the OCZ Vertex drives is able to catch up to the Intel and skew the result towards its favor. Be aware though that aggregate and average statistics have a bad way of fuzzing the real statistics for individual parts because the hide all the peaks and valleys, the important ones that expose weakness in certain areas, such as the random read and write performance.

If I consider my intended usage for an SSD it will be primarily to hold the OS, Program, and Game files and most of the access there will be random as I start new applications or activate new features or load resource files. Very little will be sequential since I won't be using my SSD to store any sizeable media files, which will remain on my HDD drives that give me great price/performance ratio for sequential data access and large media file storage.

So in conclusion, I bit the bullet and finally joined the SSD crowd by buying the Intel X25-M 80GB MLC SSD drive for $314 USD (@ $3.925 per GB) because I feel that compared to my WD Caviar Blue 320GB 7.2K drive it will make a big difference in the responsiveness of my operating system, program loading, and game loading.

I am also aware that Intel is going to be releasing new products in the Q3/4 of 2009 in a few months but I'm unwilling to wait that long since the price of these new releases will of course be at a premium since they will be new to the market. I know that in 1-year the price of my 80GB drive will drop 50-70% but in that time I will be happy with being able to use my SSD for that whole year and get my money's ($300 isn't exactly a fortune) worth for vastly improved system responsivness and lower stress levels from having to wait for everything for so long as I do now.


So now is a good time as any to buy an SSD!
 
Last edited:
Old benchmarks with old firmware there Jak, especially the first one. And why do you think random write speed is the most important metric? As long as random write speed isn't ridiculously low, like it was with the JMicron drives, then it doesn't make much of a difference. I would argue access time and read speed are much more important, unless of course the write speed is terrible (cough... JMicron).
 
Last edited:
Old benchmarks with old firmware there Jak, especially the first one. And why do you think random write speed is the most important metric? Actually, as long as random write speed isn't ridiculously slow, like it was with the JMicron drives, then it doesn't make much of a difference.

Last one isn't an old bench mark, new firmware.

JMicron controller drives, be serious? I'm glad that I held back from SSD's until the whole JMicron issue became exposed and I was able to limit my buying choices considerably thanks to the great reviews. Still going from HDD to even a bad JMicron controller SSD without partition alignment will still have a huge performance increase so folks will be happy with the results.

Also good thing that the Intel slowdown over time firmware issue was also exposed and resolved because of reviews.

The random read performance is probably more important for regular usage since the operating system spends more time reading data and loading files than than random writing files but both of these metrics combined is really what is important. I just want to emphasize that sequential metrics are bunk and shouldn't be used to make decisions.

I've waited, read, and then weight my choices and I came down on the Intel as the choice after a bit of a price drop.

Updated: One more thing the people waiting for larger capacity SSD drives for lower prices are waiting needlessly. I just recently had to increase my Windows XP partition size from 20 GB to 40 GB to install Visual Studio 2008. Including the OS 25GB and ~30GB of games that I have installed (mostly ones I don't play anymore), that 80 GB drive is roomy enough to suit my needs and growth for a while, until next year until SSD's double or triple in capacity for same price.
 
It's not the latest firmware though.

Yes, I know that there was a new release afterward but are you implying that the newest OCZ Vertex firmware somehow matched or surpassed Intel's random read/write performance?

Because it seems like you're implying that without being specific about what performance improvements it includes, except for fixing the corruption issue with the past firmware update.

Reference the numbers please.


JakFrost said:
I just want to emphasize that sequential metrics are bunk and shouldn't be used to make decisions.
I wouldn't go that far. While sequential reads/writes occur less often, they do occur.

Really, why not? Everyone seems obsessed about hundred megabyte a second sequential metrics when the fact is that both HDD and SDD do well in sequential access with the SSDs being a few times faster than HDDs. The real difference between these two technologies really shows with random access where SSDs are hundreds of times faster than HDDs. There is a huge difference in the orders of magnitude between these technologies when comparing their performance for sequential and random read/write performance or latency.

These different SSDs compete really in the random performance category since the vast majority of access destined for them will match that kind of an access profile, with a little bit of sequential thrown in. Since all of them do well in sequential anyway, then I say disguard that metric when making a choice and focus where the real differences are... random performance.
 
Yes, I know that there was a new release afterward but are you implying that the newest OCZ Vertex firmware somehow matched or surpassed Intel's random read/write performance?

Because it seems like you're implying that without being specific about what performance improvements it includes, except for fixing the corruption issue with the past firmware update.

Reference the numbers please.

No, that's not what I was implying. I was just pointing out that those numbers may not still be accurate.

After browsing the OCZ forums, I found this iometer benchmark showing 4K random write speed with firmware 1370, so it seems there is some improvement with the new firmware.

44860838.jpg


I also noticed that firmware 1571 will be released shortly, so we'll see what that brings.

Really, why not? Everyone seems obsessed about hundred megabyte a second sequential metrics when the fact is that both HDD and SDD do well in sequential access with the SSDs being a few times faster than HDDs. The real difference between these two technologies really shows with random access where SSDs are hundreds of times faster than HDDs. There is a huge difference in the orders of magnitude between these technologies when comparing their performance for sequential and random read/write performance or latency.

These different SSDs compete really in the random performance category since the vast majority of access destined for them will match that kind of an access profile, with a little bit of sequential thrown in. Since all of them do well in sequential anyway, then I say disguard that metric when making a choice and focus where the real differences are... random performance.

You can't base everything on 4 KB reads and writes, which are used to measure random performance. That's just the very end of the spectrum (the minimum block size), 2+ MB being the other end, which is used to measure sequential performance. What about 8K, 16K, 32K, etc? At some point, the Vertex passes the X25-M, and I believe that happens much closer to the 4 KB end than the 2 MB end.
 
In your original post that I responded to, you were trying to figure out why someone would buy a Vertex over an X25-M. I'll tell you why...

Vertex 60GB - regularly $180 after rebate
Intel X25-M 80GB - $315

60GB is enough for a lot of people to install their OS and apps on, and they don't feel that the extra 20GB and slightly better performance are worth $135.
 
In your original post that I responded to, you were trying to figure out why someone would buy a Vertex over an X25-M. I'll tell you why...

Vertex 60GB - regularly $180 after rebate
Intel X25-M 80GB - $315

60GB is enough for a lot of people to install their OS and apps on, and they don't feel that the extra 20GB and slightly better performance are worth $135.

Rebate Time!

That's true, I was not aware of the big rebate for the Vertex 60GB but the list price on Newegg.com for $229 -$30 = $199 at $3.316 per GB makes it seems expensive in cost per GB, being more expensive than the 120GB version. Even if you can get it for $180 at $3 per GB from another retailer then it is a pretty good price, and this is slightly less expensive than 120GB. (I'm reference Newegg.com since I usually shop there but lately I've noticed other vendors, such as ProVantage that offer lower prices on some things, like my HP LP2475w monitor for example.)

This OCZ Vertex 60GB drive with its price and performance does fit a certain budget for some people who don't want to spend $300+ for an SSD and who can do with 60GB with an attractive sub $200 price. I can definitely see folks who would go for 2x60GB in RAID0 for $360 to get better performance to boot than a single Intel.

OCZ Vertex Considered and Recommended

I myself did consider the OCZ Vertex 60GB as a preferred alternative to Intel, especially in that Ultimate HTPC that I built for my friend. The Intel being a second choice only because of the large price premium at the time for the Intel drive. In the end I showed my friend the two choices and he chose to pay more for the Intel drive a few months back when the price was $4.5 per GB for the Intel

I myself had to make the same decision this week and I briefly looked at the prices per GB of the Intel X25-M 80GB and OCZ Vertex 120GB and decided that the small additional premium of 20% was worth the larger 40/350% increase in performance for random read/write performance.

Oversight In A Hurry

I have to honestly admit that I did not really consider going with the 60GB or 2 x 60GB because I was so focused on the 80 vs 120 numbers. I could have gone with 2 x 60GB for RAID-0 but it just didn't occur to me at the time since I was so thrilled to be able to afford $300 for an SSD upgrade that I just jumped at the Intel first, salivating at the mouth.

Was it the best choice to make verus the OCZ Vertex 60GB? I honestly don't know but buying the Intel certainly does not feel like a wrong choice because I'm going with the top performance leader that is only now slightly overpriced by 20%, and not 40-50% as just a few months ago.

Would I have gotten more performance from 2 x 60 GB for $360 instead of 1 x 80 GB for $314? That is quite possible but a few benchmarks probably would still not catch up, such as random write performance since that's like 350% out. Even with the newest firmware of 10 MB/s versus Intels 23.1 MB/s the difference is still 230% percent, closer with 2x 60 GB but still not there.

RAID-0 Striping - With A Punch In The Face

However, even as folks exalt RAID-0 setups like the greatest thing in the world, I'm running a 3.5-year old nForce4 chipset on a flaky DFI motherboard and I have routinely seen my RAID-1 setups get split into two separate arrays for no reason at all. For months everything is fine, one day the systems boots with an error and the array is split and doubled but the hard drives are perfectly fine with no errors and identical data. If this happened with my RAID-0 stripe set, say goodbye to my OS and hello reinstall/reimage from backup. I've been able to fix my RAID-1 mirror set problems by having to recreate the array and completely resync the drive for ~5-hours. How much time would it take me to do a completely reinstall or reimage from backup and restore data files if my RAID-0 went belly up one random day?

For me at this time with a limited budget it seems more prudent and safer to go with 1 x 80 GB and hope that by the time I move to the new Intel Core i7 or i5 if it's out by the time I have extra money to upgrade my system and ditch this DFI Expert Motherboard with shitty Silicon Image controller and all the other baggage and problems of AMD Opteron system.

OCZ The Underdog

Just because I chose Intel for myself I don't consider myself a fanboy needing to push the Intel X25-M brand on to people or look down upon OCZ Vertex drives. I think that the many recent convictions against Intel show that they are a corporation with no scruples that abuses its monopoly powers to step on the little guys at every turn and do tricks to keep them from gaining a foothold in the market. They did that with AMD for processors and they will do that with OCZ for SSDs.

I am all for OCZ as the underdog rolling out attractive products to punch Intel in the face and put them in their place. I don't like the prices that Intel puts on their products because if it wasn't for the little competitors to this monstrous giant Intel would raise the prices for all peripherals to unheard of levels to suck everyone dry.

I just hope that OCZ with work more with Indillix to improve their SSD products and controllers to hopefully take the crown away from Intel forcing some real competition in the market. I as the consumer want to see them fight it out and offer us better products in exchange. OCZ is still sticking its fingers in the pudding trying to figure out all the little engineering problems seeing how they are releasing a new firmware every month or so. Frankly, right now I'm not interested in their recipe since I don't want to take a taste only to spit it out and reformat my drive with every new firmware. I'm patient enough to let them work on it until they get it right with new products.

As things stand now Intel has a champion product but OCZ is biting at its heels with the Vertex line. Vertex 2 is just around the corner and the real fight is about to begin. The SSDs that we have seen now are really only the beginning and now the market is starting to take shape with the players getting into their places.

However, I'm not interested in continuing to wait for the next round of fights because the fight will last a long time before things settle down to a controlled little brawl. I waited long enough since last year and I just bought my first SSD now, but later I'll need another one with more capacity so that's where I'll be looking for my next replacement.

Let them fight!!!
 
Last edited:
Supposedly since firmware 1275 or 1.1 they managed to nix that whole "flashing the drive erases all data" thing. While there's still a risk, apparently the Vertexes can now be flashed without the need for a jumper and without loosing your data.
 
My advice would be to at least opt for an SLC drive if you're going for an SSD. Flash media is degrading rapidly in quality, with data retention rates (was 10 year, now >1 year) and write cycles (was 100k, now >3k) for new Flash chips dropping all across the board. SLC chips are the only ones with <10k cycles and data retention of <1 year at this point.

True, you're unlikely to hit the write cycle limit with 'average' usage, but why risk it by going cheap on an SSD?
 
My advice would be to at least opt for an SLC drive if you're going for an SSD. Flash media is degrading rapidly in quality, with data retention rates (was 10 year, now >1 year) and write cycles (was 100k, now >3k) for new Flash chips dropping all across the board. SLC chips are the only ones with <10k cycles and data retention of <1 year at this point.

True, you're unlikely to hit the write cycle limit with 'average' usage, but why risk it by going cheap on an SSD?

This information is the opposite of what I read myself with engineers going crazy trying to kill cells to test longevity, only finding that it takes a huge amount of write cycles to make that happen.

There is of course a difference between bargain binned SSD's from some minor manufacturers like Transcend, Supertallent, RIDATA and the major SSDs from others like Intel, OCZ (Vertex, Summitt), Samsung.

Do you have any references to what you are saying here because I'm having trouble verifying your facts about longetivity?

I found this yesterday but I had a much better reference to what I was saying, but I can't find it right now.

Ask an Intel Solid State Drive Engineer

Q. How can I be sure my SSD will be reliable?

A. There are multiple issues here and I'll try to touch on each. First of all, flash memory is a well established technology, so long term studies have been done and the results are fairly well known. Flash cells do have a limited number of read/write cycles and this is one of the major reasons that SSDs haven't been popular until recently. Fortunately, the industry has progressed to a point where the limits are far away enough that engineers can make a useful product out of it. The Intel SSD datasheet claims 5 years minimum useful life and what they mean is regardless of your usage, your drive will function for the full duration. You might get far more than 5 years if you aren&#8217;t quite so rough with it.

The Intel SSDs implement the ATA SMART monitoring feature set, and it is probably your best source of information for signs of failure. I don&#8217;t work in failure analysis so I don&#8217;t have the details, but you can be sure that the protection of user data is top priority.
 
I'll take my information from an Intel SSD engineer over an uncited, unsourced article in some random japanese tech journal, thanks.
 
I'll take my information from an Intel SSD engineer over an uncited, unsourced article in some random japanese tech journal, thanks.

Tech-On is published by Nikkei Business Publications and has an online and magazine presence. NBP is one of the biggest publishers in Japan. They're hardly some 'random' Japanese organization. Please do some research before you make yourself look foolish.
 
Tech-On is published by Nikkei Business Publications and has an online and magazine presence. NBP is one of the biggest publishers in Japan. They're hardly some 'random' Japanese organization. Please do some research before you make yourself look foolish.

If they're anything like Maximum PC then I wouldn't trust them. Just sayin'. :p
 
http://www.hardforum.com/showthread.php?t=1425800

That's the thread I posted a few days ago, linking to a series of articles at Tech-On (Japanese tech site) on recent developments with SSDs and their future.

I re-quoted the relevant sections about reliability in the articles.

TechOn - SSDs Challenge HDDs, but Quality a Problem [Part 1]

Cost vs Reliability
SSD prices may continue to drop in the future, and thanks to high-density packaging and other technologies from the semiconductor industry, will continue to shrink in size, thickness, etc. The future indeed seems rosy for equipment manufacturers.

There is, however, a problem that still has to be overcome before SSDs can achieve that future growth. Evolution toward lower cost will be made possible by finer manufacturing technology, multi-level architecture and other advances, but these changes will also degrade NAND Flash memory quality.

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.

Different Characteristics

The assured number of rewrites, data retention time and other characteristics are the values of the NAND Flash memory (Fig 9). The life of the SSD is not entirely determined by these numbers, because of extensions made possible by the controller IC to keep the number of rewrites per memory cell equal (wear leveling), by error correction coding (ECC) and other measures.

These measures help, but they don't solve the problem. In addition to controller IC algorithms, tweaking the operating system (OS) and other items, etc, the life of the SSD will vary greatly with just how it is utilized by the equipment manufacturer, and the operating environment, including temperature, altitude and vibration frequency (Fig 10).

The biggest difference between SSDs and HDDs, as a source at a major manufacturer of SSD controller IC explains, is that "The number of rewrites, data retention time and other characteristics are limited. SSDs cannot be used in the same way as HDDs, which have always been considered as effectively unlimited." It would be best, it seems, not to think of them merely as HDDs without any moving parts.

Okay, this article talks about the the declining number of rewrites as manufacturers look towards smaller fab processes to make chips and also towards doing multi-bit cells, 2-bit, 3-bit, etc. It seems perfectly understandable that making the chips smaller and then trying to pack more bits per cell is going to lead towards finer signals and more problems in bit retention and higher difficulty in distinguishing bit values. I can see how they talk about 100,000 rewrite cycles for 90nm chips with 1-bit cells, and then go 3,000 for 30nm 3-bit cells and even less for smaller chips and higher capacities per cell.

I think that this article tries to spread a little F. U. D. to get a bit of a shock value with posting this information about declining reliability of SSDs while selectively ignoring the differences between the products that various SSD manufacturers make. They only make a little reference to data shuffling that the current SSDs do to prevent certain the cells from being over utilized and spreading the writes around evenly.

Not all SSDs are created equal and this is what is expected so it always goes back "caveat emptor" (buyer beware).

However, my original advice still stands and that is to buy your SSD from a reputable manufacturer. I feel confident that buying from Intel or Samsung are pretty safe bets. OCZ is becoming a leader in the field but I'm more concerned about their product lines since they did sell JMicron controller based SSD before they moved to the new Indillix. But would you really buy an SSD from an unknown company like SuperTalent, RiDATA, Transcend, or another generic rebrander offering massive capacities for dirt cheap prices and then put your operating system with all your documents on one of these drives? I sure wouldn't.

I trust my data to an SSD from a major manufacturer more since I know that they have their reputation on the line and at least the bigger manufacturers do their own testing to determine longetivity and reliability before they sell such products on the US market. Litigation is expensive, but more expensive is the loss of reputation if your new SSD suddenly dies and loses all the data and suddenly everyone is blaming your product for losing data and your brand name is tarnished forever.

I'm confident that the current generation of good quality drives is up to par in being able to handle server and desktop workload for years of service. I only need about 1-3 years of time for this drive to work since by that time the SSD market will be drastically changed and my little 80GB will be forgotten as the floppy disk is now. I'm sure that Intel over engineered the drive and speced to do a lot more than I will be using it for. I'll be running a SMART hard drive monitor utility to monitor the drive and inform me of any changes in the health status also just to make sure that things aren't degrading too quickly for me.

I bought an Intel drive because I'm not fully certain about OCZ Vertex line as the answer. OCZ is releasing a lot of firmware versions meaning that they are still stirring the pot and aren't finished cooking the product. Hopefully, they'll improve their product line in the near future and give Intel a run for the money to lower their damn high prices. I want to see competition.
 
So now is a good time as any to buy an SSD!


For use in a desktop, what do you think would be a better purchase:

(2) X25-M 80GB in RAID 0 on ICH10R

or

(1) X25-M 160GB

Or, what is a better X25-M purchase right now for any use: 2 X 80GB or 1 X 160GB?
 
Last edited:
For use in a desktop, what do you think would be a better purchase:

(2) X25-M 80GB in RAID 0 on ICH10R

or

(1) X25-M 160GB

Or, what is a better X25-M purchase right now for any use: 2 X 80GB or 1 X 160GB?

2x80gb as you'll have faster reads/writes. The ICH10R is pretty good up to 750~ read/writes.
 
Thanks for your opinion, that is what I am likely to do.

Would anyone take a 160GB X25-M over 2 X 80GB if it was the same exact price?
 
True enough, JakFrost :) Buying from a reputable vendor is never a bad idea.
 
Thanks for your opinion, that is what I am likely to do.

Would anyone take a 160GB X25-M over 2 X 80GB if it was the same exact price?

Only if it was going into a notebook or sff case with very limited room.
 
Thanks for your opinion, that is what I am likely to do.

Would anyone take a 160GB X25-M over 2 X 80GB if it was the same exact price?

Yeah, but only because my motherboard doesn't have a RAID controller. :D Then again... if I was rich enough to afford 2 of these, I guess I could afford a RAID controller as well.
 
I am assuming that for RAID use on an ICH10R, setting AHCI vs IDE would not be relevant?
 
You really would only go IDE if you planned on Windows XP-only, since it's not-AHCI compatible outside the box. In all other cases you should go AHCI-mode. Some of the SSD controllers are optimized to take advantage of NCQ (which the Intel controller is).
 
I thought just installing the RAID drivers would lock you into SATA-AHCI mode with no IDE avaliable.
 
Depends on the chipset but some motherboards have a feature in AHCI/RAID called IDE compatibility that's separate from running IDE mode. (My EVGA X58 has this).

Not sure if that would result in RAID without NCQ.
 
Thanks.
It is a P5Q Deluxe.
I will check, but sounds like I would not want RAID with IDE mode.
 
Please take your motherboard raid questions elsewhere since this is a topic regarding SSDs.

However, for your information most recently released motherboard enable AHCI and NCQ if you set the drive mode to RAID. Intel recommends to set board with their ICH chipsets to RAID from the start and load their drivers during OS install to give you AHCI and NCQ support from the start and if you later decide to do a RAID-1 array with your OS drive you can do so without installing any additional drivers.

AHCI vs IDE &#8211; Benchmark & Advantage

Many SATA controllers can enable AHCI either separately or in conjunction with RAID support. Intel recommends choosing RAID mode on their motherboards (which also enables AHCI) rather than the plain AHCI/SATA mode for maximum flexibility, due to the issues caused when the mode is switched once an operating system has already been installed.
 
Depends on the chipset but some motherboards have a feature in AHCI/RAID called IDE compatibility that's separate from running IDE mode.

I C, says the blind Old Hippie! :)
 
Does anyone have any concrete info on how RAID0'd SSDs affect random read/write performance and IOPS? I seem to recall seeing some benchmarks that suggest it negatively affects both, but I may just be imagining things.
 
Thanks for your opinion, that is what I am likely to do.

Would anyone take a 160GB X25-M over 2 X 80GB if it was the same exact price?

I just did. I'd rather not have to deal with RAID, as it's just another level of complication that I really don't want to deal with. RAID arrays are also tied to their controller, whereas with one drive I can always use another PC to pull data off the disk if I need to. Besides, I think that the drives are pretty freakin fast as is, and things like access time aren't helped by having more drives. One slot, one cable, no complications...on my primary PC, I prefer one drive.

That said, I do have an HTPC with 5 x 750 GB drives in RAID 5 for storing data. I believe in RAID, I just don't feel like dealing with it on my main PC when one SSD seems to be fast enough.
 
Back
Top