Personal computing discussed
Moderators: renee, morphine, Steel
DPete27 wrote:I agree. With the limitations of SATA III, I'm not sure the new controllers will be anything groundbreaking. Heck, even PCIe SSDs don't offer that much additional performance that will give the average consumer a night-and-day change like going from a HDD to SSD.
The biggest (and most important IMO) improvement to look forward to is for SSD prices to drop again. It's been...what?...over a year since SSD prices have changed significantly? Sounds to me like TLC is one of the only options available today to drive that price down.
kamikaziechameleon wrote:I think we will see $200 Terabyte SSD's before we see our next massive performance jump.
Milo Burke wrote:Then again, Intel created a fantastic product with S3700. ... the S3700 was the first of a new generation focused on consistently high IOPS. But I don't believe anyone has followed the S3700 along that path, or that many have realized how special the S3700 is. Maybe if more websites publish TechReport-style thorough reviews over SSD speed consistency, the enthusiast world would wake up to the inconsistency of current-gen SSDs, and more manufacturers will focus on making consistent products.
bandannaman wrote:Interesting. What kinds of workloads do you think would benefit from this consistency, and which ones suffer from inconsistency? Many enthusiasts focus on gaming performance, which doesn't seem oriented at consistency as a high-value thing. For myself, I'm way more interested in virtualization, e.g. running development SQL databases and full-blown IDEs inside VMs, which may well benefit from consistency and other things not routinely measured.
I view the evolution of "affordable" SSDs as falling across three distinct eras. In the first era we saw most companies focusing on sequential IO performance. These drives gave us better-than-HDD read/write speeds but were often plagued by insane costs or horrible pausing/stuttering due to a lack of focus on random IO. In the second era, most controller vendors woke up to the fact that random IO mattered and built drives to deliver the highest possible IOPS. I believe Intel's SSD DC S3700 marks the beginning of the third era in SSD evolution, with a focus on consistent, predictable performance.
Chrispy_ wrote:I don't think SATA3 is holding back consumer SSD's that much.
jihadjoe wrote:I get mixed feelings regarding process shrinks for SSDs. On the one hand a smaller process means more GBs per $, but OTOH it also means less reliability.
Milo Burke wrote:I feel like we've been pushing the limits for SATA III for a while now. I don't know if we need SATA IV or the commoditization of PCI SSDs or some other format, but I feel like we need an interface upgrade before we can have much appreciable improvements realized.
Milo Burke wrote:bandannaman wrote:I believe Intel's SSD DC S3700 marks the beginning of the third era in SSD evolution, with a focus on consistent, predictable performance.
jihadjoe wrote:I get mixed feelings regarding process shrinks for SSDs. On the one hand a smaller process means more GBs per $, but OTOH it also means less reliability.
The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable. If someone could make a 4TB SSD for an affordable price simply by going to 3.5" or 5.25, they would have SURELY done it by now.
geekl33tgamer wrote:The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable. If someone could make a 4TB SSD for an affordable price simply by going to 3.5" or 5.25, they would have SURELY done it by now.
He has a point - I thought about this some time back and got into a mini-debate over it with another guy at the office. Why isn't there larger capacity SSD's in 3.5" cases, as we never reached any firm conclusions...
The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable. If someone could make a 4TB SSD for an affordable price simply by going to 3.5" or 5.25, they would have SURELY done it by now. Doing so would allow them to completely dominate a huge emerging market. I suppose it's possible that they have all been in collusion thus far (like we saw with DRAM pricing some years back), but I think in this case the payout would be far too lucrative for everyone to play nice with each other.
kamikaziechameleon wrote:The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable. If someone could make a 4TB SSD for an affordable price simply by going to 3.5" or 5.25, they would have SURELY done it by now. Doing so would allow them to completely dominate a huge emerging market. I suppose it's possible that they have all been in collusion thus far (like we saw with DRAM pricing some years back), but I think in this case the payout would be far too lucrative for everyone to play nice with each other.
Affordable, well I mean more affordable to do that I think that to say put years of R&D into shrinking 4 plus terabytes into a 2.5 inch drive. I mean that just my thought. Given the insane margins they sell SSD's for you know that fab price is fixed, cost of research is generally astronomical compared to producing proven tech. I think that they could have given us consumer 4Tb drives if they cared to.
The Egg wrote:kamikaziechameleon wrote:The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable. If someone could make a 4TB SSD for an affordable price simply by going to 3.5" or 5.25, they would have SURELY done it by now. Doing so would allow them to completely dominate a huge emerging market. I suppose it's possible that they have all been in collusion thus far (like we saw with DRAM pricing some years back), but I think in this case the payout would be far too lucrative for everyone to play nice with each other.
Affordable, well I mean more affordable to do that I think that to say put years of R&D into shrinking 4 plus terabytes into a 2.5 inch drive. I mean that just my thought. Given the insane margins they sell SSD's for you know that fab price is fixed, cost of research is generally astronomical compared to producing proven tech. I think that they could have given us consumer 4Tb drives if they cared to.
My point is, I don't think form factor has anything to do with anything. We've had PCIe card SSDs all along with plenty of physical PCB space, and you still don't see anything over 1TB. Manufacturers aren't stopping at 1TB and throwing up their hands because it won't fit in a 2.5" drive.
Airmantharp wrote:It's all in the controllers, period.
If we had controllers that supported more packages and more individual die (as lanes or channels, just like DRAM), we would have higher capacity SSDs at only marginally greater than incremental costs. Just note that the returns for this technology is going to be relatively small, because as mentioned above, the gains for most users are effectively negligible.
the wrote:There is means to scale this: put the flash on DIMM modules. Just like DRAM, users would be able to both increase performance by enabling more channels on the controller and continue to increase storage capacity when number of channels is maxed out on the controller. This could also improve reliability if a flash stick goes bad, it could be replaced without affecting the others.
The Egg wrote:Your theory of the 2.5" form factor being the limitation for SSD sizes is laughable.
The Egg wrote:My point is, I don't think form factor has anything to do with anything. We've had PCIe card SSDs all along with plenty of physical PCB space, and you still don't see anything over 1TB.