SandForce Using Intel 25nm MLC Flash On SSDs

Quote:

---

While walking around the IDF 2010 technology showcase we stopped by the SandForce booth to see what new SSD (Solid State Drive) technology that they were showing off. SandForce didn't disapoint us as we quickly saw several very cool SSD products sitting around the booth. The first thing that caught our attention was the SSD pictured above. This is the first public demo of an SSD that we have seen in person that uses the latest Intel 25nm MLC TSOP NAND Flash!

---

Source: Legit Reviews

Just needs a SATA-3 connector to be perfect... :cool:

very interested

Quote:

---

Originally Posted by L1m1t.NL

Just needs a SATA-3 connector to be perfect... :cool:

---

Hopefully it will have by the time it ships to masses - would be a waste if not.

Am I the only one that would love a smaller driver (around 64GB) with a fantastic controller, TRIM + RAID support and quality SLC NAND?? I can't be...

Quote:

---

Originally Posted by ElSel10

Am I the only one that would love a smaller driver (around 64GB) with a fantastic controller, TRIM + RAID support and quality SLC NAND?? I can't be...

---

Dont get your hopes up, even Intels new extreme drives will use MLC this time around.

Btw do anyone know if the rumours of close to half the GB price is true with 25nm? Seems stupid that the GB prices for quality drives is almost the same as when I bought my Vertex 1

Quote:

---

Originally Posted by Mech0z

Btw do anyone know if the rumours of close to half the GB price is true with 25nm? Seems stupid that the GB prices for quality drives is almost the same as when I bought my Vertex 1

---

smaller chips are naturally cheaper, but it also means that supply is increasing (more chips per month created). however we are still way to early in the life of SSDs to know how much demand will go up as prices come down a hair, which can make the pricing drop, MUCH slower.

even if this is able to make chips like 50% cheaper, and 50% more available, it may only drop prices by 10% right away, and probably 20% more before the next node is used where we "should" get another 50% drop + 50% availability. (this is my assumption)

Quote:

---

Originally Posted by Mech0z

Dont get your hopes up, even Intels new extreme drives will use MLC this time around.

---

That is so incredibly lame.

Not really. 34Nm MLC nand has matured and controllers have improved to reduce wear. Adding more over provisioned nand capacity also significantly increases the life of nand cells and that is much easier to do with lower cost and higher capacity nand.

Here Micron talk about process maturity of 34nm to 30K write cycles
http://www.youtube.com/watch?v=NiV8ATBoqXI

Here Micron talk about MLC with 100K write cycles, which is the same as current SLC.
http://www.youtube.com/watch?v=lQAxNV3b6Ec

Most Sandforce based drives on the market use Intel 32nm nand so no surprise that they will use 25nm nand.

Quote:

---

Originally Posted by Ao1

Not really. 34Nm MLC nand has matured and controllers have improved to reduce wear. Adding more over provisioned nand capacity also significantly increases the life of nand cells and that is much easier to do with lower cost and higher capacity nand.

Here Micron talk about process maturity of 34nm to 30K write cycles
http://www.youtube.com/watch?v=NiV8ATBoqXI

Here Micron talk about MLC with 100K write cycles, which is the same as current SLC.
http://www.youtube.com/watch?v=lQAxNV3b6Ec

Most Sandforce based drives on the market use Intel 32nm nand so no surprise that they will use 25nm nand.

---

Well those videos sounds good but they are pretty old and I am still not really the wiser about what for example a Intel 32nm is good for? Or what about the upcoming intel are they 10k or 30k :confused:

That's kind of interesting. Even though these drives are the strongest competitors to Intel, they also benefit Intel by using their memory chips.

Does it even matter whose nand they use? It's mostly about the controller.

the difference between 25nm and 34nm is bigger than it looks, it is something more like 4x smaller

(25*25)/(34*34) = .54

so its 54% the size, close to half the size

that is true!

Quote:

---

Originally Posted by Intel

The new 25nm 8GB device reduces chip count by 50 percent compared to previous process generations

---

The whole SLC/MLC issue is the enthusiast OCD kicking in.

A 10,000 write cell allows you to write the entire storage of the drive 1 a day for 3-4 years. I mean who does that and even if you do, you are not going to have it that long.

Add that with the advanced wear leveling and add cache to that and who knows how long a drive that isnt punished will last.

I don't like the sandforce compression crap. It's smart, but it kills performance if you need to read/write incompressible data (IE compressed RARs). All current benches on them are misleading because 99% of them are done with highly compressible data (IE OS files or synthetic bench data).

Anandtech has an article up a while ago where they used an incompressible data file for IOmeter and the results were UGLY. The lack of a buffer really shows.

Quote:

---

Originally Posted by HKPolice

I don't like the sandforce compression crap. It's smart, but it kills performance if you need to read/write incompressible data (IE compressed RARs). All current benches on them are misleading because 99% of them are done with highly compressible data (IE OS files or synthetic bench data).

Anandtech has an article up a while ago where they used an incompressible data file for IOmeter and the results were UGLY. The lack of a buffer really shows.

---

Why I went with my C300 :)

Quote:

---

Originally Posted by turtletrax

Why I went with my C300 :)

---

Not worth the price premium. I see you have tons of spare cash so you don't care, but even if the C300 could max out SATA3 @ 500MB/s, diminishing returns have already kicked in for SSDs so OS/Game load times will not be halved vs 250MB/s.

Intel's G2 drives are still the best overall because of their reliability + random read/write speeds. G3 will improve on that, but I doubt it'll even be SATA3 :(

Quote:

---

Originally Posted by HKPolice

Not worth the price premium. I see you have tons of spare cash so you don't care, but even if the C300 could max out SATA3 @ 500MB/s, diminishing returns have already kicked in for SSDs so OS/Game load times will not be halved vs 250MB/s.

Intel's G2 drives are still the best overall because of their reliability + random read/write speeds. G3 will improve on that, but I doubt it'll even be SATA3 :(

---

I actually got my C300 for less per GB than a sandforce based drive...

I got the C300 for the craploads of transcoding I have to do and will be grabbing another for source/target transcoding. Has been working very well and have been impressed with it. I almost did go dual 160 G2s but wanted larger drives and wanted to stay away from RAID at all costs. Actually glad I went the way I did.

Quote:

---

Originally Posted by ElSel10

Am I the only one that would love a smaller driver (around 64GB) with a fantastic controller, TRIM + RAID support and quality SLC NAND?? I can't be...

---

I can have my choice of MLC or SLC NAND, I still choose MLC high capacity in RAID.. after a certain RAIDed speed there's not much improvement to client workloads :P

Quote:

---

Originally Posted by Mech0z

Btw do anyone know if the rumours of close to half the GB price is true with 25nm? Seems stupid that the GB prices for quality drives is almost the same as when I bought my Vertex 1

---

Well, that's because the NAND prices dropped like a brick in the past two years, after the manufacturers all over expanded like mad, creating a massive oversupply. Everyone says they've learned their lesson and prices are recovering as well.

Quote:

---

Originally Posted by iddqd

Does it even matter whose nand they use? It's mostly about the controller.

---

It does matter actually.. the timing parameters on NAND can vary greatly. Controllers can mask this to some extent, but it will have an effect on performance.

Quote:

---

Originally Posted by turtletrax

The whole SLC/MLC issue is the enthusiast OCD kicking in.
A 10,000 write cell allows you to write the entire storage of the drive 1 a day for 3-4 years. I mean who does that and even if you do, you are not going to have it that long.
Add that with the advanced wear leveling and add cache to that and who knows how long a drive that isnt punished will last.

---

10k write cycle flash once a day = 30 years, not 3

Quote:

---

Originally Posted by HKPolice

I don't like the sandforce compression crap. It's smart, but it kills performance if you need to read/write incompressible data (IE compressed RARs). All current benches on them are misleading because 99% of them are done with highly compressible data (IE OS files or synthetic bench data)
Anandtech has an article up a while ago where they used an incompressible data file for IOmeter and the results were UGLY. The lack of a buffer really shows.

---

It actually does have an internal buffer, just not external DRAM like Intel/Indilinx etc.
Compressible vs Incompressible depends on the application and workload..
There's ways to make any drive stumble or look bad if you give it its worst case scenario.
I like Sandforce for its balanced set of features- the compression and IOPS as well as the quality design circuitry they put into their designs, and their rigorous firmware testing programme.

Quote:

---

Originally Posted by iddqd

Does it even matter whose nand they use? It's mostly about the controller.

---

not all nand is created equal by a long shot.

Bad news, guys... :(

Quote:

---

Intel has started mass production of its new NAND flash circuits with 25nm technology, which will lead to substantially higher storage capacities. We now have reliable information that points to that 25nm products will not appear before February 2011, but in large quantities.

Source: Nordic Hardware

---

Oh well... guess that will coincide nicely with my new laptop based on Huron River. It better have SATA-3 by then!

Quote:

---

Originally Posted by HKPolice

I don't like the sandforce compression crap. It's smart, but it kills performance if you need to read/write incompressible data (IE compressed RARs). All current benches on them are misleading because 99% of them are done with highly compressible data (IE OS files or synthetic bench data).

Anandtech has an article up a while ago where they used an incompressible data file for IOmeter and the results were UGLY. The lack of a buffer really shows.

---

lol gotta love the hating going on here. the 1% of the time your computer uses majorly compressed data is the reason you call the BEST SSD's on the market crap?

50$ says even with the data compressed your CPU cant uncompress it faster than the SSD can read it!

Which in effect makes your point useless and the Sandforce weakness a bit of a red herring.

Quote:

---

Originally Posted by HKPolice

Not worth the price premium. I see you have tons of spare cash so you don't care, but even if the C300 could max out SATA3 @ 500MB/s, diminishing returns have already kicked in for SSDs so OS/Game load times will not be halved vs 250MB/s.

Intel's G2 drives are still the best overall because of their reliability + random read/write speeds. G3 will improve on that, but I doubt it'll even be SATA3 :(

---

Intel G2? lol no thanks

Intel vs Sandforce Raid 0

You're clearly biased towards sandforce with your Raid0 setup. If this compression method of accessing NAND is so great, why didn't Intel or Marvell or Samsung implement it? Were they all too stupid to consider it even though they literally have over a billion dollars of R&D resources?

Obviously you're not an SSD engineer so there are other cons about using onboard compression that you don't know about or haven't been made public. The only reason sandforce went this route is because it lowers production costs by cutting out the external cache chip. This is the only way sandforce chips could be price competitive since they're such a small startup company with no fab of their own.

If Intel decides to add compression to their G3 (or G4) SSD what will you think then?

Innovation from smaller companies drives the larger companies back to competition. Intel got comfortable with the performance of the X25-M first gen, and look what happened to the G2 - they had a firmware update increase the artificially capped write speeds.

Quote:

---

Originally Posted by HKPolice

Snip...

---

Don't be so brand sensitive, I'm a happy G2 owner, but it's obvious sandforce 'currently' has a far superior controller than Intel.

I hope G3 get's a serious boost because upgrading from a G1 to G2 was a waste of time, I wil certainly think twice before listening to Anands Intel hype again.

^ not sure what you expected, latency was the same, read / write speeds were nearly the same. If something isn't at least 20% faster in the things you are most affected by (IMO) its not worth upgrading to.

Even if a drive like the C300 is faster at reading and writing large sequential files, where it matters is at the 4-512kb file size and latency/IOPs unless you regularly work with large files in which case you'd likely be better served by a conventional disk :p

The C300 does happen to be significantly faster at the 4k file sizes than the intel drives though and I would like to swap out my 2x vertex 120's for a pair and retire these to another role

http://images.anandtech.com/graphs/c...5742/23532.png
http://images.anandtech.com/graphs/c...5742/23533.png

http://www.anandtech.com/show/2899/13

The Sandforce (Vertex 2) results don't look "UGLY" to me.

Quote:

---

Originally Posted by HKPolice

I don't like the sandforce compression crap. It's smart, but it kills performance if you need to read/write incompressible data (IE compressed RARs). All current benches on them are misleading because 99% of them are done with highly compressible data (IE OS files or synthetic bench data).

Anandtech has an article up a while ago where they used an incompressible data file for IOmeter and the results were UGLY. The lack of a buffer really shows.

---

so since we're talking about sandforce compression.

I read that the sandforce controllers are able to achieve a write amplification of as high as 0.5.

Does this have to do with the compression?

The SF results are artificial, since IOMeter data is compressible and thus IOMeter gives higher results.
That does not mean it's bad with non-compressible data, just not impressive like C300.

Quote:

---

Originally Posted by zoob

http://www.anandtech.com/show/2899/13

The Sandforce (Vertex 2) results don't look "UGLY" to me.
http://images.anandtech.com/graphs/c...5742/23532.png
http://images.anandtech.com/graphs/c...5742/23533.png

---

Quote:

---

Originally Posted by alfaunits

The SF results are artificial, since IOMeter data is compressible and thus IOMeter gives higher results.
That does not mean it's bad with non-compressible data, just not impressive like C300.

---

http://www.anandtech.com/show/3681/o...200-reviewed/4

Something I don’t get with the Sandforce drives. They compress data so that less data is written to the drive. Although less data is written the OS records the uncompressed data size. So if a drive is showing up as full on the OS in reality it isn’t. So why could this unseen capacity not be used for over provisioning to cut down on the amount of nand being used? :shrug:

Perhaps that's their way of cutting down on wear.

Quote:

---

Originally Posted by danielkza

http://www.anandtech.com/show/3681/o...200-reviewed/4

---

So?

Quote:

---

Originally Posted by alfaunits

So?

---

Did you read the article? :rolleyes:

Quote:

---

Originally Posted by Anand

In our Agility 2 review I did what you all asked: used a newer build of Iometer to not only write data in a random pattern, but write data comprised of truly random bits in an effort to defeat SandForce’s data deduplication/compression algorithms.

---

Quote:

---

Originally Posted by Ao1

Something I don’t get with the Sandforce drives. They compress data so that less data is written to the drive. Although less data is written the OS records the uncompressed data size. So if a drive is showing up as full on the OS in reality it isn’t. So why could this unseen capacity not be used for over provisioning to cut down on the amount of nand being used? :shrug:

---

Okay, what happens if you write 100% incompressible data to the drive?
SSDs require overprovisioning to function properly and efficiently...
Since Sandforce often has "extra overprovisioning" due to the data compression,
this simply allows it to function that much more efficiently.

Quote:

---

Originally Posted by Fuji

so since we're talking about sandforce compression.
I read that the sandforce controllers are able to achieve a write amplification of as high as 0.5.
Does this have to do with the compression?

---

Well, the write amplification on all other drives is >1.0,
the Sandforce gets as LOW as 0.5..
that means some data written of 1GB uncompressed size is being compressed to (on average) 0.5GB.
But Sandforce says they are seeing an average WA of 0.7 in most user cases.

Quote:

---

Originally Posted by zads

Okay, what happens if you write 100% incompressible data to the drive? SSDs require overprovisioning to function properly and efficiently...
Since Sandforce often has "extra overprovisioning" due to the data compression, this simply allows it to function that much more efficiently.

---

But the chance of that happening would be remote and if the drive was full you would only be reading from it anyway. Let’s say your drive is 80% full. How much of that 80% is actually written data? 40%? 60%?

Quote:

---

Originally Posted by zoob

Did you read the article? :rolleyes:

---

You forgot to include the sentence that followed that quote.
“What we saw was a dramatic reduction in performance”.

“………..The takeaway? While SandForce’s controllers aren’t immune to performance degradation over time, we’re still talking about speeds over 100MB/s even in the worst case scenario and with TRIM the drive bounces back immediately.”

Still good performance even in a worst case scenario, I guess the issue is that mileage can vary.

Quote:

---

Originally Posted by Ao1

You forgot to include the sentence that followed that quote.
“What we saw was a dramatic reduction in performance”.

“………..The takeaway? While SandForce’s controllers aren’t immune to performance degradation over time, we’re still talking about speeds over 100MB/s even in the worst case scenario and with TRIM the drive bounces back immediately.”

Still good performance even in a worst case scenario, I guess the issue is that mileage can vary.

---

The bolded text is exactly what I'm trying to get at - Even in the worst case scenario it can still outperform the Intel SSD in 4K Random writes. Even with an Agility 2 on random data on 4K random, it can outperform an Intel X25-M G2.

No doubt; compressed writes, both sequential and random are faster. Compressed reads are slower going by AS SSD Benchmarks; ~ 210 vs 250. In some instances the Sandforce is faster and in others it’s slower. Either way no difference to the majority of users. Both drives are fast.

Regardless, they'll definitely serve their purpose as an OS drive. Longevity of these SandForce drives is what I'd worry about.

So I said... it's not bad, just not as impressive as C300.
It would be a hazard if it could not outperform an intentionally gimped >1yr old X25-M ;)

Quote:

---

Originally Posted by zoob

The bolded text is exactly what I'm trying to get at - Even in the worst case scenario it can still outperform the Intel SSD in 4K Random writes. Even with an Agility 2 on random data on 4K random, it can outperform an Intel X25-M G2.

---