That comparison was astounding. To think of all the seconds I'm wasting....
Here's hoping the prices aren't as astronomical as they probably are.
Yeah, we need this tech. The mechanical drive is severely bottlenecking our system performance. This video drives the point home more than any bar graph ever could.
I wonder how these would perform against the current range of SSD's?
I always thought that the performance gained by an SSD was minimal and only for those bleeding edge people. After watching that video...makes me want one. As Snark mentioned...prices are probably super high.
I always thought that the performance gained by an SSD was minimal and only for those bleeding edge people. After watching that video...makes me want one. As Snark mentioned...prices are probably super high.
This SSD will push down the prices on tons of disks which are 70-85% as fast, meaning a second or two slower. Today's performance SSDs will still deliver the results demonstrated in the video.
I wonder how the reliability will hold up to an abusive writing use with a photoshop or premier paging file?
The current standard is to use a RAID or second HD for the paging file. These newer generation of SSDs seem to have as high or higher (MTBF) than traditional drives. Some of the best enterprise SAS drives cap out at 1.5 million hours MTBF vs SSD with 1.5-2 million hours MTBF depending on the drive.
The MTBF is an even more worthless figure for SSDs than it is for mechanical disks. What we need to know from manufacturers is their estimation of how many write cycles the SSD is capable.
An even more useful figure for customers would be gigabytes per day (a function of write cycles), or years of service assuming one complete rewrite of the NAND per day.
For example:
A 60GB SSD might have NAND with a write endurance of 35,000 cycles. This means that 60GB can be erased and rewritten 35,000 times, for a total of 2,100,000 gigabytes. 2.1 million gigabytes divided by 365 days is 5753 days, or 15.7 years.
In other words, you could write 60GB a day for 15.7 years without killing the SSD. Higher capacity SSDs would last even longer.
These figures may even be a little conservative given the current state of MLC NAND endurance and write combining at the controller level.
These figures also assume that the user is a good user and has disabled or moved all unnecessary writes to the SSD: Windows Indexing Service, System Restore, Windows Defrag, NFTS Memory Usage, Date Stamping, Boot Tracing, Page File (moved to RAMdisk), Prefetching, 8.3 Names, and Superfetch.
Working on it, guys, but I hope to have something much larger in the works.
Take your time, I'm sure it will be worth it. My "by Wednesday" comment was only because my pair of SSDs arrive on Wednesday; I might pick your brain a bit in IRC if the Google fails me.
No regular consumer on earth is going to disable any of those services, much less can even tell you what they are.
Indeed. It would be lovely if OEMs would tweak Windows services on systems equipped with SSDs or if SSD manufacturers would release tools with simple GUIs to get the most out of their products without having to scour the interwebs and spend time in the command shell.
Thrax
🐌Austin, TX Icrontian
Comments
Here's hoping the prices aren't as astronomical as they probably are.
Yeah, we need this tech. The mechanical drive is severely bottlenecking our system performance. This video drives the point home more than any bar graph ever could.
I wonder how these would perform against the current range of SSD's?
This SSD will push down the prices on tons of disks which are 70-85% as fast, meaning a second or two slower. Today's performance SSDs will still deliver the results demonstrated in the video.
The current standard is to use a RAID or second HD for the paging file. These newer generation of SSDs seem to have as high or higher (MTBF) than traditional drives. Some of the best enterprise SAS drives cap out at 1.5 million hours MTBF vs SSD with 1.5-2 million hours MTBF depending on the drive.
the heavy write cycles with paging files or database driven applications warrant needing the write cycle figures.
Some estimates I've seen for MTBF only use a 33% write cycle
For example:
A 60GB SSD might have NAND with a write endurance of 35,000 cycles. This means that 60GB can be erased and rewritten 35,000 times, for a total of 2,100,000 gigabytes. 2.1 million gigabytes divided by 365 days is 5753 days, or 15.7 years.
In other words, you could write 60GB a day for 15.7 years without killing the SSD. Higher capacity SSDs would last even longer.
These figures may even be a little conservative given the current state of MLC NAND endurance and write combining at the controller level.
These figures also assume that the user is a good user and has disabled or moved all unnecessary writes to the SSD: Windows Indexing Service, System Restore, Windows Defrag, NFTS Memory Usage, Date Stamping, Boot Tracing, Page File (moved to RAMdisk), Prefetching, 8.3 Names, and Superfetch.
The <.0001% of us who understand what that even means don't figure into these equations.
Any useful SSD reliability metric will have to figure for the normal computer user.
Indeed. It would be lovely if OEMs would tweak Windows services on systems equipped with SSDs or if SSD manufacturers would release tools with simple GUIs to get the most out of their products without having to scour the interwebs and spend time in the command shell.
I also wrote a similar script using AutoIt which disables all the above features in one swoop.