Corsair Unveils Fourth Generation SSDs with LAMD Controller

Status
Not open for further replies.

dalethepcman

Distinguished
Jul 1, 2010
1,636
0
19,860
27
With all the issues the last Sandforce based drives had, I can understand Corsair wanting to switch controllers. Hopefully they can still offer a > $1/GB ratio on the smaller drives.
 

zulutech

Distinguished
Mar 30, 2011
139
0
18,680
0
[citation][nom]dalethepcman[/nom]With all the issues the last Sandforce based drives had, I can understand Corsair wanting to switch controllers. Hopefully they can still offer a > $1/GB ratio on the smaller drives.[/citation]
Don't you mean < $1/GB?
 
[citation][nom]carvedinside[/nom]And later on we will find out that LAMD controller is in fact a rebranded Marvell 88SS9174[/citation]

The standard Neutron series does seem similar to Vertex 4 in performance specifications. If this LAMD controller is similar, then that would explain that. If the standard Neutrons are similar to Vertex 4, then Neutron GTX drives should beat Vertex 4. Perhaps a Vertex 4 MAXIOPS will be made that will be similar to the Neutron GTX series? Granted, it's all very loosely based speculation, but the numbers do look curiously familiar.

[citation][nom]ubercake[/nom]Does the GTX model only work with Nvidia cards?Who came up with that marketing scheme?[/citation]

IDK why you were voted down so badly, I thought it was something worth mentioning. Are all of you other people just going to ignore the similarity in the name? I think that they should have used a different name. Maybe something with a similar context. What about going with another particle, perhaps the Neutrino? Other decent options might include Neutron Star or something like that. Heh, Pulsar and/or Magnetar.
 
[citation][nom]TheBigTroll[/nom]the capacities look a lot like how sandforce drives are arranged....[/citation]

The capacities are multiples of 60 or 30 instead of 64 or 32 because that means that some of the capacity is reserved for the controller for TRIM and that kind of stuff. It should help with longevity, reliability, and probably also keeping performance more consistent over time.
 

DRosencraft

Distinguished
Aug 26, 2011
743
0
19,010
6
The specs seem similar to the Mushkin Chronos series. I don't know about the reliability on either side though. It'll be interesting to see how it works out.
 

s3anister

Distinguished
May 18, 2006
679
2
19,060
37
[citation][nom]ubercake[/nom]Does the GTX model only work with Nvidia cards?Who came up with that marketing scheme?[/citation]

The "GTX" moniker first appeared on Muscle cars and has since been used for the branding of a lot of products similar in the way that "GT" and "Super" have.
 

alidan

Splendid
Aug 5, 2009
5,303
0
25,780
0
[citation][nom]blazorthon[/nom]I'd say sometime between the years 2014 and 2018 seems reasonable for that much of a price drop.[/citation]

all i want is the chip size on the waffer, not the black shell, and i could figure out how much we are screwed in pricing...

if the cost per square mm is close to break even price, we wont see 35 cent per gb drives till the 10-14nm mark, if there is a jarge proffit margin, we may never see cheap ssds, as hdds will always out store ssds (till we get 3d chips at least) and thats all most people care about. with an ssd being a boot for most people and just a boot, that means they can stick around the 1$ per gb range and have no need to go lower.
 

drlawyer

Honorable
Mar 13, 2012
2
0
10,510
0
After learning (the hard way) that most consumer motherboards - even so called "high end" or "gaming" boards - utilize the Marvel "it's SATA III except that it's really kinda not" controller, I find myself looking for a (cost-effective) alternative. Am I reading this correctly, that the Neutron will feature an on-drive controller? Or is it simply compatible with the LAMD controller if you happen to have a system with one installed?
 
[citation][nom]drlawyer[/nom]After learning (the hard way) that most consumer motherboards - even so called "high end" or "gaming" boards - utilize the Marvel "it's SATA III except that it's really kinda not" controller, I find myself looking for a (cost-effective) alternative. Am I reading this correctly, that the Neutron will feature an on-drive controller? Or is it simply compatible with the LAMD controller if you happen to have a system with one installed?[/citation]

The controller for an SSD controls the NAND chips and lets the SATA controller on the other end of the SATA cable use the SSD. These drive's can't have the SATA controller on-board because the chipset would need to be modified (and perhaps the very SATA interface) to let the drive have the controller. This would also increase latency dramatically and there would be other drawbacks. A comparable situation would be moving the memory controller of a CPU back onto the motherboard instead of on the CPU. Latency would increase greatly.

There would also still be the problem that the SATA ports on the motherboard (or expansion card connected to the motherboard, or any other way of using a SATA bus on a computer) would still need to be interfaced all the way to the CPU without a bad link being in that chain. Theoretically, it would mean that the quality of the SATA controller on the drive would be able to better dictate the throughput of the drive, but it would have larger latencies. If you use the Intel or AMD SATA3 controllers instead of any poor Marvell controllers, throughput is not a problem like it is on some of those Marvell controllers.

Also, the Marvell controllers are truly SATA3, they simply don't have throughput that can match the Intel or even the AMD SATA3 controllers. SATA3 is a specification and the Marvel controllers must abide by it to work with SATA3 devices. Saying that they aren't quite SATA3 is like saying the Celeron G440 isn't quite a CPU just because it has a mere ~1/3 of the dual threaded performance thatthe next Celeron up, the Celeron G530, has.

Point is, the controllers on SSDs are not SATA controllers, they are NAND flash memory controllers. They are more comparable to the memory controller on the LGA 775 boards and similar arrangements. Sure, they're a memory controller, but since the chip with the controller is separate from the CPU, there needs to be yet another connection from the CPU to that north bridge that has the memory controller. In this case, that would be the CPU's side of the FSB and the SATA controller is akin to it in this comparison. Sure, it's doable, but it would require a reworking of SATA and although it could mean that each drive has more consistent throughput a little more regardless of what SATA port that you put it in so long as the port is SATA3 compatible, but at the cost of latency, increased complexity of implementation, and all the work that would need to go into implementing it.
 
Status
Not open for further replies.

ASK THE COMMUNITY

TRENDING THREADS