Time To Upgrade: 10 SSDs Between 240 And 256 GB, Rounded Up

[DemolitionDemon]

[citation][nom]mapesdhs[/nom]Some info for what it's worth, I've been using OCZ drives for a while, been very happy so far; Ihave V2E, V3, V4, various capacities. However, as part of building a 3930K setup for a friend foruse with After Effects, I recently bought a Samsung 830 256GB and I have to say I'm veryimpressed. After doing a full Win7 install, the 830 showed a totally smooth performance profile inHDTach, more than 400MB/sec right across the range of the test. None of the OCZ models I'vetested have ever maintained performance like this. As a result, I changed the 2nd drive (forthe AE cache) from an OCZ 120GB MAX IOPS to another Samsung 830 256GB (persuaded myfriend the extra cost was worth it).I do like OCZ drives, especially the V2E series for systems that do not support TRIM (ie. UNIXmachines), but the Samsung really is rather good (it was 130 UKP total from ebuyer).Ian.[/citation]


Not sure I follow.

I mean OCZ statistics show as being much higher in the 256GB section than Samsung does, could you elaborate further on what you mean by your results showed Samsung coming out on top, I am interested as the price difference is $30 so either one is worth a shot over my old 7200 1TB.

 

[mapesdhs]

What I meant was, after doing a full Win7 install and testing with HDTach, the performance profile across the medium showed
highly varied graphs for the OCZ models, with many peaks & troughs, whereas after doing an install onto the Samsung 830
the HDTach result was a completely smooth 400MB/sec right across the graph, and higher max rates overall. Somehow, the
Samsung is maintaining consistent performance much better. I've not seen this with any of the OCZs.

And btw, the results for OCZ models do not always show them as higher than the Samsung.

One caveat though: my OCZ models are all 128GB or less, so the 830 does have the advantage of more space to play
with re moving data around, but even so, the results were startling. Note that by full install I mean about 35GB used.

Just for reference, I own one OCZ Vertex2E 120GB 3.5" (my 1st SSD), ten OCZ Vertex2E 120GB 2.5", four Vertex2E 60GB,
two Vertex3 120GB, one Vertex3 60GB, nine Vertex3 MAX IOPS 120GB, one Vertex4 64GB, one Corsair F60 60GB (similar
to the 60GB V2E) and one Corsair Force Series3 60GB. The two Samsung 830s are in a 3930K system I'm building for a friend.

Ian.

 

[DemolitionDemon]

Thanks, I will have to look into this further, but I'm not completely convinced that your results are completely valid due to OCZ and Samsung have better specifications the more storage they have hence that I feel a vertex 4 256gb would be on par with a Samsung 830 256gb, I will have to ponder some more before making a choice, at the moment the Samsung is $30 AUD cheaper at the moment with the OCZ claiming better speeds, so I'm left on the balancing board of cheaper or slightly better.

 

[blazorthon]

Results like that are completely irrelevant if you're comparing different capacities. That's not fair at all as the difference between a 128GB and 256GB capacity can be a considerable difference in performance behavior regardless of brand.

 

[blazorthon]

[citation][nom]DemolitionDemon[/nom]Thanks, I will have to look into this further, but I'm not completely convinced that your results are completely valid due to OCZ and Samsung have better specifications the more storage they have hence that I feel a vertex 4 256gb would be on par with a Samsung 830 256gb, I will have to ponder some more before making a choice, at the moment the Samsung is $30 AUD cheaper at the moment with the OCZ claiming better speeds, so I'm left on the balancing board of cheaper or slightly better.[/citation]

I prefer Vertex 4 over Samsung 830 overall, but if the price difference is that huge, then go for the Samsung 830 over the Vertex 4 without a second thought about it. There's no incentive on the Vertex 4 to pay so much more at the same capacity.

 

[mapesdhs]

[citation][nom]blazorthon[/nom]I prefer Vertex 4 over Samsung 830 overall, but if the price difference is that huge, then go for the Samsung 830 over the Vertex 4 without a second thought about it. There's no incentive on the Vertex 4 to pay so much more at the same capacity.[/citation]

That's why I eventually recommended the Samsung to my friend, it was a lot cheaper than the V4, and I don't like the way the
V4 behaves when it fills up.

Btw, I don't think the capacity difference explains the results I've seen. When doing an install onto a 120GB V3, there's
_plenty_ of space available for the unit to maintain performance in the equivalent manner when only about 35GB is used
and there's some overprovisioning anyway, but it doesn't result result in the kind of performance profile shown by the 830,
not even after a TRIM and some idle time to let it funk away. Note it's not a mbd issue, I have numerous different boards
and systems.

DemolitionDemon, since when was 'feel' any more valid a basis for a buying decision? ;D

I did a lot of research on what model to recommend to my friend, and initially it was to be a Sandisk Extreme 240GB for
the system disk and a MAX IOPS 120GB for the AE cache, but then something odd happened with the supply of Sandisk
units (dried up here) and the prices shot up 10% in just one week. Plus, certain sellers were doing an offer on the Samsung
(ebuyer had hundreds in stock, selling them for 130 UKP each with free shipping, which even now is 37 less than the V4
and 29 less than the Sandisk here), so that was that. Then the good price on the MAX IOPS vanished (they had been 80
each on Amazon, but stopped - my order for one was cancelled, though I managed to get 4 for my own use earlier), so I
recommend just getting a 2nd 830 for the AE cache.

Ian.

 

[DemolitionDemon]

Haha I should have explained further however that reply was from my phone which took forever.

To me $30 isn't a huge difference, and I'm not sure we are trying to compare cross currencies but $30 is 2 hours worth of hard work with my current job if that helps compare.

What I was trying to get at is that you tested with a 128GB Vertex 4 which specs are 550MB/s read and 430MB/s write, where as the Vertex 4 256GB version is 560MB/s read and 510MB/s write, and so along with the increase in storage is an increase in performance, and I assume the increase is in multiple areas however have to research that further also.

This is why I am unsure with what you have concluded as I believe Samsung also follows in the same idea that the 256GB performs better than the 128GB, therefore comparing a 256GB to a 128GB would seem unfair.

So this is why I sit on the fence wondering if I should go a Samsung or OCZ, at the end of the day price isn't an issue, but if you do say the Samsung is better performing despite what the specs are and that includes both hard drives being of same storage, then Samsung may just be the way to go, either way it is an upgrade from my old mechanical drive.

 

[DemolitionDemon]

As I cannot edit a post that I am aware of.

"256GB to a 128GB would seem unfair."
wouldn't* was what I meant.

 

[blazorthon]

[citation][nom]mapesdhs[/nom]That's why I eventually recommended the Samsung to my friend, it was a lot cheaper than the V4, and I don't like the way theV4 behaves when it fills up.Btw, I don't think the capacoty difference explains the results I've seen. When doing an install onto a 120GB V3, there's_plenty_ of space available for the unit to maintain performance in the equivalent manner when only about 35GB is usedand there's some overprovisioning anyway, but it doesn't result result in the kind of performance profile shown by the 830,not even after a TRIM and some idle time to let it funk away. Note it's not a mbd issue, I have numerous different boardsand systems.DemolitionDemon, since when was 'feel' any more valid a basis for a buying decision? ;DI did a lot of research on what model to recommend to my friend, and initially it was to be a Sandisk Extreme 240GB forthe system disk and a MAX IOPS 120GB for the AE cache, but then something odd happened with the supply of Sandiskunits (dried up here) and the prices shot up 10% in just one week. Plus, certain sellers were doing an offer on the Samsung(ebuyer had hundreds in stock, selling them for 130 UKP each with free shipping, which even now is 37 less than the V4and 29 less than the Sandisk here), so that was that. Then the good price on the MAX IOPS vanished (they had been 80each on Amazon, but stopped - my order for one was cancelled, though I managed to get 4 for my own use earlier), so Irecommend just getting a 2nd 830 for the AE cache.Ian.[/citation]

Vertex 3 had over-provisioning. Vertex 4 does not. Regardless, it may have simply been a blunder in whatever firmware version you used. Which version did you have?

Also, I'd avoid pretty much any SandForce SSD, but especially those from OCZ. Other than those made by Intel, they all have reliability issues, but OCZ was often especially bad about this with their SandForce SSDs.

 

[blazorthon]

[citation][nom]DemolitionDemon[/nom]Haha I should have explained further however that reply was from my phone which took forever.To me $30 isn't a huge difference, and I'm not sure we are trying to compare cross currencies but $30 is 2 hours worth of hard work with my current job if that helps compare.What I was trying to get at is that you tested with a 128GB Vertex 4 which specs are 550MB/s read and 430MB/s write, where as the Vertex 4 256GB version is 560MB/s read and 510MB/s write, and so along with the increase in storage is an increase in performance, and I assume the increase is in multiple areas however have to research that further also.This is why I am unsure with what you have concluded as I believe Samsung also follows in the same idea that the 256GB performs better than the 128GB, therefore comparing a 256GB to a 128GB would seem unfair.So this is why I sit on the fence wondering if I should go a Samsung or OCZ, at the end of the day price isn't an issue, but if you do say the Samsung is better performing despite what the specs are and that includes both hard drives being of same storage, then Samsung may just be the way to go, either way it is an upgrade from my old mechanical drive.[/citation]

Overall, I find the Vertex 4 SSDs to be better performers than Samsung 830 at the same capacity, especially in writes, but in the real-world, most people simply don't use workloads where the difference is noticed much if at all. That's why the price difference breaks the deal for me. The Samsung 830 also uses less power, but they are both so low in power usage compared to some of the most power efficient HDDs that it's not like it's generally a big deal except in a laptop where the difference in battery life can be somewhat noticeable.

 

[DemolitionDemon]

Cheers guys, I may just have to bite the bullet and just buy either one, I'll make a deciding factor when I get paid haha.

 

[mapesdhs]

DemolitionDemon writes:
> ...What I was trying to get at is that you tested with a 128GB Vertex 4 ...

Actually no, I was using a 64GB V4. Feel free to scream even more, but you're missing my main point...


> which specs are 550MB/s read and 430MB/s write, where as the Vertex 4 256GB version is 560MB/s read and 510MB/s write,

... namely, it's not max speeds I was really getting at, it was the way the 830 recovered from what had been done
to give a completely smooth performance profile, even though I had given it no idle time (just the usual Empty
Recycle Bin to fire off a TRIM). This performance recovery doesn't happen with any of the OCZs I've tested.


> and so along with the increase in storage is an increase in performance, ...

Max performance numbers? Yes. Basic controller behaviour? No.


> ... Samsung also follows in the same idea that the 256GB performs better than the 128GB, ...

Of course, but it seems to _recover_ better from usage, that's what I was getting at.


> ... So this is why I sit on the fence wondering if I should go a Samsung or OCZ, ...

Despite my being a long time fan & user of OCZ models (see my earlier summary of what I own), I'd say
Samsung or Intel for the reliability factor, even though personally I've never had a problem with any of my
OCZ drives.

Btw, if I yak about UK pricing, just multiply by 1.6 to get an approx. $US equivalent (not strictly accurate,
but sufficient for discussions).


> ... either way it is an upgrade from my old mechanical drive.

As blazorthon said, in reality it's unlikely you'd ever notice the difference between a 256GB V4, 830 or Vector
in normal use.

One thing btw, I really do like OCZ's Toolbox update program. By contrast, I found the way Corsair does it to
be absolutely horrible. Samsung's app is ok though.


> As I cannot edit a post that I am aware of. ...

I thought this until recently, but actually you can. Look at the top of the list of comments, there's a link that
says, "Read the comments on the forums". Click on that link, find your post, look to the bottom-right of your
post and you'll see three icons. Click the middle icon for a 'quick edit' mode, or the left icon for a full editor.


> Cheers guys, I may just have to bite the bullet and just buy either one, I'll make a deciding factor when I get paid haha.

I'm sure you'd be happy with either model. 8)


blazorthon writes:
> ... Which version did you have? ...

Latest 1.5.


> ... Also, I'd avoid pretty much any SandForce SSD, but especially those from OCZ. Other than those made by Intel,
> they all have reliability issues, but OCZ was often especially bad about this with their SandForce SSDs.

OCZ certainly had problems at first, but so did _everybody_; Intel messed up too remember (that 8MB issue).
These old fw issues have long since been fixed (though it's true OCZ seemed more happy than most to release
units with such issues, though who knows whether that was deliberate), which is why Sandisk do so well with
their models that are based on the earlier controller.

And I guess it's all very much personal experience, some people have had just one OCZ and it's died, whereas I
have loads of them and not one has gone wrong so far. I've read equivalent comments across numerous sites
about pretty much every brand of SSD, except Intel and Samsung.

However, from the perspective of how companies seem to deal with the thoroughness of their testing prior to a
product release, it does seem as if Intel and Samsung are that the top of the pack. This is why, even though I'd
be perfectly happy with the *performance* of a V3 or V4, I'd still recommend an 830 to a friend if reliability is a
key factor.


> Overall, I find the Vertex 4 SSDs to be better performers than Samsung 830 at the same capacity, especially in writes, ...

I did think about the V4 256GB carefully (I had been considering one for my own 2700K and 990X systems),
but in the end for my friend's AE build I just wasn't keen on the issues related to what the V4 does when it
fills up rapidly. Naturally for normal desktop use this shouldn't be an issue with a whopping 256GB SSD, but
for use with AE as a cache? Doing a hefty HD job could indeed fill it up real quick - even the lowly 10-second
HD benchmark test I'm using creates 6GB of cache data very fast.

This harks back to your earlier comments, and those from Andrew Ku, namely that every person's usage & criteria
are different.


> ... That's why the price difference breaks the deal for me. ...

That was the final clincher, the V4 was a lot more here at the time. IIRC, the 830 was 130 UKP while the
V4 was about 165.


> The Samsung 830 also uses less power, but they are both so low in power usage compared to some of the most

True; in a system like the one I'm building though (3930K, Quadro 4000, 64GB RAM, SAS RAID, etc.), the SSD's
usage really isn't relevant in the grand scheme of things.

Ian.

 

[DemolitionDemon]

Cheers for clearing that up and I think I'll have myself a new Samsung, cheaper and if it is indeed more reliable then I don't have to play around, having this not go right on the car is enough I don't need to be messing with the computer haha

 

[blazorthon]

[citation][nom]mapesdhs[/nom] Really long post that I don't want to reply to and make my post much longer than it needs to be, so I'm cutting it down [/citation]

Sorry, but I fail to see how comparing a 64GB SSD to a 256GB SSD like that could in any way be fair. Behavior is not necessarily identical along the entire line and at least according to my own little tests, my Vertex 4 128GB simple doesn't have issues like you reported, although I admit to not running benchmarks much as I don't want to waste endurance cycles. I think that you are greatly underestimating the performance difference that a 64GB drive would exhibit versus a 256GB drive. Heck, the 64GB drive wouldn't even nearly fill all of the SSD's channels and would have very little room to move things around as you fill it up because of that and also it's low capacity.

Sure, OCZ wasn't the only one that had problems, but among the main players, they still usually had among the worst reliability of the SandForce users due to some issue(s) with OCZ's power delivery circuitry. Intel had much better reliability overall. Having issues and having anywhere near as many as OCZ are two very different things.

I've owned many SSDs from OCZ and a few other companies and can say that it's not just my failure rates that I'm looking at (I realize that as a single person, even if I had several dozen SSDs, is simply too small of a sample size for overall failure rates of everyone else to necessarily line up with mine), also those around the web. Vertex 4 seems to me to be the first recent heavy attempt from OCZ at delivering superb reliability and performance in a single drive with usually great pricing and they backed it up with superb customer service overall too. I've had a few Vertex/Agility 3 drives fail (only a few out of several, but still) and only one Vertex 4 issue (was a DOA on my first unit; I didn't even need to pay return shipping and they sent a replacement very quickly). Vertex 4, seemingly thanks to Marvell, is also very reliable like the Samsung SSDs and other modern Marvell-based models such as Crucial M4.

 

[mapesdhs]

[citation][nom]blazorthon[/nom]Sorry, but I fail to see how comparing a 64GB SSD to a 256GB SSD like that could in any way be fair. ...[/citation]

How a controller recovers from the device being written to *should* be reasonably comparable, irrespective of capacity. You're
being far too dismissive IMO, and I don't know why. It's almost like saying the 64GB V4 must be garbage while the 256GB is
brilliant; I don't buy that notion at all.

And again, I wasn't talking about max I/O rates which of course are going to vary greatly by capacity. Stop focusing on
'performance', that wasn't my main point. Ach, never mind...

Ian.

 

[blazorthon]

[citation][nom]mapesdhs[/nom]How a controller recovers from the device being written to *should* be reasonably comparable, irrespective of capacity. You'rebeing far too dismissive IMO, and I don't know why. It's almost like saying the 64GB V4 must be garbage while the 256GB isbrilliant; I don't buy that notion at all.And again, I wasn't talking about max I/O rates which of course are going to vary greatly by capacity. Stop focusing on'performance', that wasn't my main point. Ach, never mind...Ian.[/citation]

How the drive behaves most certainly can depend on the capacity. For example, even with earlier firmware versions, the 64GB and 128GB capacity models for the Vertex 4 would slow down to performance levels of previous versions of firmware at certain capacities being used, yet the 256GB model didn't exhibit such behavior much, if at all. The same is true for the 512GB model to an even greater extent, but it doesn't really improve on the 256GB model in performance.