SSD Thread (Benchmarks, Brands, News, and Advice)

What advantages does a V3 have over a 510 in a typical, single user scenario?

When the V3's supposed advantages are taken advantage of (in a server setting) it will shut itself down (DuraWrite/LTWT) and spend the rest of its life navel gazing (until the next SE brings it temporarily to life).

There is nothing but 'hype' in the new SF and V3 offerings (and I can see they belong together from a mile away).

The industry (manu's and reviewers) are trying to kickstart the SSD fever they've been promising for so long - but they're going about it the wrong way: a slightly cheaper price is not what will attract and keep serious buyers; a strong (lie-free) statement of what the drives can actually do for us will.


Although this is old and maybe biased to OWC products, this is what people should be considering before buying an SSD:

See:
MPG - SPECIAL REPORT: Real World SSD Performance - Real-World SSD Performance Considerations (Introduction)

 

SF 2xxx series may be different from 1xxx series. We don't know yet about the DuraWrite effect in it. You comment is valid for the 1xxx, not sure about the new one.

 

I have nothing to add to all the information that can be found online. It's all pretty well described there.

PS. the article you're linking to is about OS X, which doesn't have TRIM. (at least at the time of writing of the article)

 

Phil, about the TRIM, I know.

But neither do the SF drives in Win7 either.

 

I've done several heavy duty tests on Sandforce drives and never noticed a performance drop. Seems to me like TRIM works fine.

Do you have a source that says otherwise?

 

Source?

I've read of a few others right here on this forum that have voiced the same concern/conclusion.

 

No links, no source... too bad.

In my experience Sandforce drives hold their performance well, except for the famous throttle bug, that only affects sequential write speeds for incompressible data.

Edit: This article confirms that TRIM does work on Sandforce, as long as you're using Intel RST drivers:
http://www.behardware.com/news/10962/sandforce-trim-listen-up.html

 

Toms Hardware just released an article with the Intel 320 and Crucial C400/M4,

Crucial m4 And Intel SSD 320: The Other SSD Competitors

 

Anandtech has their review up too:
The Intel SSD 320 Review: 25nm G3 is Finally Here - AnandTech :: Your Source for Hardware Analysis and News

And Techreport: http://techreport.com/articles.x/20653

 

Seems all are coming out with the reviews, also add

Storagereview Intel SSD 320 Review (300GB)

Legit Reviews Intel 320 Series 300GB SSD Review w/ 25nm Flash!

PCPerspective Intel 320 Series 300GB SSD Performance Review

TheSSDReview Intel 320 Series 300GB SATA II SSD Review

 

From the horse mouth, i.e. OCZ. They have publicly said that TRIM in SF drive does not have effect in removing the DuraWrite brake

 

That's another issue. The DuraWrite brake doesn't mean that TRIM doesn't work.

 

Do we really need to spend that much time or these 'terminology' thing ?

For most users, they expect TRIM to be able to restore performance(almost immediately). When people said 'TRIM' doesn't work(I assume tiller included), it means they don't see the restoration in speed. No one care how it is done internally.

 

What you call DuraWrite ONLY affects the speed of sequential writes and ONLY for incompressible data.

That's only a small part of overall performance. TRIM affects the overall performance of the drive.

So it would be fair to say TRIM doesn't work fully.

Does that mean that TRIM doesn't work on Sandforce drives? of course not. TRIM works fine on Sandforce drives except for sequential writes of incompressible data. Those are limited to about 80 MB/sec.

 

mhh, luckily they used the cheaper 25 nm..

i'm quite disappointed.... in a relative way, intel has improved greatly in writes from G2, but in the end they got only now the same performances of sandforce old generation.

I can only imagine how much work intel has done to achieve theese results, but other sellers haven't rounded their thumbs.

I presume that it will be a fight between the brands, one claiming to have the best endurance, the other will show great benchmarks.

Some people like phil who have been in good feeling with sandforce, will tend to trust again it, as many other people will do.

Others like me who tasted the quality of intel products, will keep buying them.

i still don't understand why they keep prices sooo high; maybe they don't have understand that keeping the market closed to few enthusiasts, SSD won't take off.
Sorry if i sound like captain Obvious but i talk as an average customers would do; we have already outperformed hard disk from some years, research must go on, it's right, but please calm down prices

 

I could totally buy a 600GB G3 instead of my X220...

 

As a company, why would Intel just give better prices? They wouldn't and they don't. Well understood from a business perspective.

They didn't waste any time on this 'refresh' - with the controller developed over two years ago, they simply enabled some features it already had to remain/seem still competitive today.

HDD's have hardly been surpassed yet: less than .1% of the storage market is SSD based. Still a long, long road to go.

 

To me its the demand, while i was yesterday waiting for the 320s, i was spoting the 120gb G2, and they sold like 8 in less than 6 hours that i was checking, the drives sell very fast, specially new anticipated drives, probably they dont have the capacity to do so many drives, and as long as people clear the shells with their current price scheme they will mantain them high. Wait a week or two, you will see the 320s go up in price, my guess is $600 for the 300gb.

 

aristocracy is always surrounded by the masses
sorry for my bad english, i wanted to say that actual SSD are already great improvements.It's time to increase this 0,1 %...

 

splinterpc, lol...

the saying that comes to my mind is: the bigger the lie, the more believe.

 

Corsair Force Series F60
Firmware: 2.1b

 

Excellent Laptopmagreview found by Cloudfire.
Intel SSD 320 (300GB) Review - A Review of the Intel SSD 320 (300GB)

Many real world tests, done on a notebook. Intel 320 vs. Vertex 2, Samsung 470 and C300.

Intel 320 doesn't look that impressive.

 

That review shows how fast SSDs actually are in programs we all use compared to 7200 RPM HDD

 

where did you get firmware 2.1b?

 

Did anyone noticed that Intel 320 SSD have same controller as G2?

Does it means that G2 is able to write much faster than ~100MB/s?

 

Theoretically, yes, but Intel will not be providing firmware that uncaps the G2.

 

According to Tom's HW, the G3 has more cache and a faster clock which could be the main contribution factor for the speed increase. Sure they may have tweaked the algo as well as other wise they can't get the 30% improvement in endurance on 25nm flash but that I believe only affects RANDOM write, mostly.

That could also be the reason why newer gen controllers(SF 2xxx, Marvell) suck more power(they are faster CPU in a sense).

 

As I know Intel SSDs doesn't write user data to cache, so it shouldn't matter at all, but new algorithm could be an answer.

 

You're right, no user data in cache - but the bigger/faster cache is used for live housekeeping. So it definitely helps in getting better write performance (along with the updated algorithms).

 

I read that the spare area is about 13 %, still low compared to sandforce;
With 25 nm shrink, i would know if it's adeguate

 

Maybe the 'old' SandForce (~28%), with the current SF stuff its around 7% so Intel is still 'better'.

 

Where would they get the extra spare in 320 ? a 80G would still be using 10x8192MB NAND, 160 double that. So the spare would still be 7% or so.

the only thing I see that is different is the 300G and the 600G.

 

"The 300 GB SSD 320 sample in our Bakersfield, CA lab has 20 NAND packages, and each package adds 16 GB to the drive's capacity. In total, our 300 GB has 320 GB of raw NAND flash, and like other consumer drives, the SSD 320 series employs ~7% overprovisioning. Hence, this drive is listed as the 300 GB model (279 GiB in Windows). Data parity is already accounted for in the ~7% of over-provisioning."

 

yes, that is what I meant. The extra seems to be only applies for the 300/600G but not 160 <.

 

Do you mean the 7% overprovisioning (21GB) or the extra 20GB on top of that?
160GB do also have 7% overprovisioning hence the 149GB available flash.

 

no. I don't think we are talking about the same thing.

Each NAND in the 160GB case is 8192x1024x1024, and there are a total of 20 such chips inside. But the drive capacity is 160x1000x1000 exposed. The 149GB inside Windows use the (160x1000x1000) / (1024x1024) => 152GB or so.

This was always the case. So I that is still about 7% provision.

For the 300GB, it would be 16384 x 1024 x 1024 x 20. That however only expose 300x1000x1000 rather than the 320x1000x1000. That extra are for the additional provision.

 

Looks like the 120GB Vertex 3 just showed up at NewEgg. Out of stock immediately, of course.

 

I already had it in my cart and it went out of stock while I was checking out. Today's Newegg coupon for $25 off would have been too nice.
The big question now...wait, or use the $25 off against a vertex 2???

 

At this point, get a vertex 3, vertex2 use 25 nm so...get a controller optimized for them

 

Just noticed that Amazon changed the status of my order: Delivery estimate for the 240GB Vertex 3 is April 5. I placed my order on March 10, for reference for anyone else who ordered through Amazon.

 

^^ mine is April 6th
hope it's true

 

I just put my new Intel 320 series SSD im my system.

Here's a first run of CrystalDiskMark, I haven't done any system tweaks yet, it's a fresh install of Windows 7 Professional, 64bit.

 

Guess you could pump your 4K up by applying either those or those tweaks...

 

Grats, I'm jelous... Still thinking on cancelling the intel 320 in favor of the vertex3, hope you post some benches to see if the real production is as good as the pre production reviews we seen.

 

Another run, this time with the stamatisx tweak applied and Throttlestop. (Intel 320 series 300GB drive)



Nice improvement imo.

 

I agree, nice improvement - but, do you notice this while using the system normally (not just benching...)?

The 'numbers' are not what have me interested (they're more than good enough) - what has me interested in the 320 series is the capacity that is available (300GB+ is minimum for 2011).

Thanks for any feedback on how the drive 'feels'.

Cheers!

 

Yep, I definitely will.

 

Is it true that benching your SSD a few times can kill it?

 

Is it true that driving your car sucks the gaz out of your tank ?



An SSD is a myriad of flash cells (like the ones on an SD - MMC - and so - media cards) and those cells get used as you write to them. Basically, the more you write to a cell, the sooner it will get used, worn, and at the end, won't be able to retain any more information you would write to it.

Now, thing is that those cells have been made in such a way that they can endure a lot of writes before they actually die. Still, every single write is a step toward it's end of life...


Therefore, manufacturers of SSDs introduced wear leveling, which is an algorithm that guides the data to be written to the disk toward the cells that have had the less writes at that time. This is to prevent some cells to die WAY before others. However, if you start benching your SSD like crazy, if you run 50 iterations per day, each comprised of 1000MB, surely, you're helping your puppy to reach it's end of life sooner than it should...


So, NO, benching your SSD a few times won't kill it; however, it's obvious that the more you write to it, the sooner it will die, but hey, under normal conditions, your SSD should last 5 ~ 10 years, so you should not worry about that.

If you would like to get more infos on this, may I kindly suggest you take a look here, and there ?

Cheers !

eYe

 

Very Nice Sir !