We've just posted a news article on an upcoming solid state drive based on DDR memory.
Looks like we are going to see some movement in this space and of course we are interested in your opinion on those devices. A option to accelerate read/write process or still very much somthing you would not consider to use at this time?
I'm hoping this technology becomes mainstream... and fast. I could really use drives like this at work where I don't need more than a 4-5 GB harddrive for the computers we supply to our customers.
The main thing for me is the fact that these drives will use less power, create less heat, run faster (obviously) but most importantly... have a longer lifespan than conventional Hard Drives which when placed in a bad environment don't last very long.
Hello Wolfgang. I'd be more interested in something like the HyperDriveIII, only with more slots, battery backup good for four hours, and standby power directly from the power supply.
Yes, I understand the problem with making the controller chips fit with more DIMM's, I'd like someone to fine a workaround. 16 DIMMs could be fit to a 5.25" drive bay device if the company were to put the controller chip on an upright card extension. Timing issues with unequal length traces? Slow the modules down! Even at 100MHz the DIMMs could easily saturate an SATA 3gb connection.
Heck, someone could even divide the board and have two SATA 3gb connections.
What's to gain from using PCI-Express x1 anyway? It's not any faster than SATA 3gb, in fact it's slower "in theory" and I seriously doubt there's that much difference in practice.
I am very interested in this technology. Two of these configured in a RAID 0 for a total of 16GB would be pretty nice. Using two PCIe lanes should up the bandwidth too, although, it would still bottleneck. I also like the idea of more slots for more total memory (like the 5.25 drive bay idea mentioned by earlier). Only, the power adaptor with no batteries worries me. What if you need to move the computer? An external battery sounds like a bit of a hassle... Guess I'll just have to wait for carbon nanotubes, when volatility is no longer an issue. Hopefully it will help with the space issue too...
I am very interested in this technology. For expandable scientific computing on a budget, it can't be beat. It would certainly be cheaper to add this card and some RAM onto a workstation than it would to buy a whole new workstation with more ram slots or faster harddrives.
There are a couple of adaptations I'd like to see to make the product more useful. I'd like to see the card adapted to use an 8x slot on an SLI board or even a 16x slot for applications where graphics speed isn't important. It also might be useful to see this card adapted to expresscard slots. Then, because it would be an external device, you could make the enclosure whatever size you see fit, it could work for laptops or desktops (if there are adapters out there), and potentially fit much more memory inside (think external optical/hard drive enclosure).
I am very much waitting where this will go as i have selected this technology to be my next performance system drive.
My system drive will, at most, occupy no more than 10 GB (personal data reside on separate disks) , so this is a viable option for me.
However, I expect a few things to be resolved first:
1) A good, on-board, battery. I don't want to the hussle of trying to find another piece of equipment to support the first one I bought.
2) Maximum performance interface. Right now that seems to be SATA II
(300). Why neither iRAM nor HyperOS not DDRdrive support it, is beyond me.
3) Memory supply. Like with the battery issue, i don't want to be guessing what the right memory will be. The manufacturer must have available memory packages of 2GB RAM sticks and at cheap prices too (as the device does not even require PC2700 to run, even PC2100 is sufficient).
This would be amazing and expensive no doubt. The hardrive is one of the slowest component of the computer, with these would dramatically increase the pc's performance. Wow, it's just amazing to think having a hardrive that is as fast as the memory. Man, I could boot up my computer and start playing BF2 in let say 15 seconds?
I am also very intregued about solid state drive's. Even running sata in raid 0 doesn't seem to quench my thirst for responsivness.
For a heavy multi-tasker, large photo editor, gamer, or even small video editor, having your intire windows install(including the page file) on a solid state drive would definitly have it's benefits in removing the bottle neck that slow down performance.
Although expensive, i imagine running two ram disk cards of any kind in raid 0 with 4gb on each would cost roughly the same as running 2x 36gb WD raptor's in raid 0. Granted one has much more space, but it isn't solide state. And really, more and more people are running their operating system on a small partition such as 10-20gb to facilitate easy backup's and rebuild's/reformats. Then keeping the bulk of their data on a massive storage drive, or the other partition in the HD.
There are talk's about solid state growing in existing disc hard drive's. I would like to see the day when HD's consist of ~5gb solid state partitions for the OS, and programs to be installed in, with the remainder of the files being stored in the tradition disc section of the HD.
I only hope this technology matures quickly. As a power user/ over clocker any performance mod is of great interest to me. Why must this tech be available only as an add on card--I would think a remote device would work better than an SSD PCI-e card with a (lol) remote power source . Perhaps a specialized workstation mobo that incorporated SDD technology would make the physical limitations of only 8 GB a moot point--an extended ATX board could sacrifice some real estate for double stacked DIMMS of SSD--likewise power management features enabled in BIOS for such a board could provide juice from SSD complian PSUs (in the future). Like so many mobo features we now take for granted, what was once only available on a card (SCSI, RAID, FIREWIRE, USB, LAN, NIC etc) is now part of a chipset of other onboard controller.
Unfortunately, it seems like this futurity is a good ways off and likely to follow the same path as all those other emergent technologies.
I would hope mobo mfrs would have the vision to make a leap forward to support SSD with onboard controllers in highend workstations.
Of course what seems like a lot of storage now will seem miniscule by the time this tech matures (in much the same way 40 gig HDDs were "huge" 5 years ago...)
Thanks for spotlighting this inevitable progression to solid state.
I second the notion that a relatively small solid state drive to boot the OS and run most commonly used applications will be tremendously useful, and still allow conventional hard drives to be used for storage. Actually conventional drives can be speeded up by building a RAM table of storage location and record ID information, followed by searching in RAM for the location of a record, and finally using the location information to go directly to the record for retrieval. I wrote an app once in VB6 and used DAO (data access objects) to do this very thing. Speed was amazing, but not surprising given that searching on the hard drive (after the initial RAM table buildup) was eliminated. There are many, many apps in scientific areas that need tremendous speed (MRI image processing, protein folding, weather forecasting, to name just three) and I hope solid state drive technology will be pushed hard and as fast as possible. As to when I jump in with a purchase I am looking right now at systems with PCIe slots and the solid state drives to fill them. After I have done a reasonable review (which should be soon) I am likely to go ahead and take the plunge.