Subjective Review
Unfortunately this drive did not get the test of “real world performance” why? Because 128 GB is the bare minimum recommended here for a desktop OS drive. Currently the 24/7 desktop is using 130 GB of a 250 GB and as such could not be imaged over to this drive. 256 GB and large drives will be reviewed when time allots. On paper it looks to do well, not a screamer, but better than the alternative to not a screamer. Vertex 4 is somewhere between a SandForce and a Marvell drive.
No award or points will be affected by this.
The Conclusion
Should an Indilinx driven Vertex series drive perform like previous generations? If it bears the name Vertex, it should. In fact when the Vertex moved from 34nm to 25nm OCZ allowed people to return then rev 2 Vertex 3 because it did not have the exact same performance of the initial revision. So is the Vertex 4 worthy of The Vertex name? In many ways it is. Small file size write are off the charts much like the previous generations. Reads were really low in small sizes (also typical).
ATTO performance was all over the place, the move to each next cluster size would report as much as 20 MBps different performance in writes and over 100 MBps in reads! While we were capable of hitting 540 MBps read speed, it was not the norm, most size varied below that. Most SSDs will hit there speed and do that the whole time maybe get incrementally faster.
Another “issue” that is being incorrectly bandied about on the web is the linear write performance. Indilinx/OCZ has optimized the firmware to improve write speed over that of the typical Marvell controller. In a SUSTAINED write-up to 50% of empty NAND remaining on the drive the performance is amazing. After passing the 50% mark the firmware drops into storage mode and write speeds drop to under 100 MBps. Many people seem to think that if the drive is half full write performance drops.
That is not entirely accurate. Yes performance does drop when the drive is 50% full. However, this is true of most drives. The thing to note is it does not drop to below 100 MBps. Only if you are writing more than 60 GB of information in one shot to an empty drive, or sustained 30 GB file to a half full drive will you see the performance drop. Even then it is only when you pass those numbers that the performance drops. If imaging an SSD from another SSD, So when empty the first sustained 60 GB of a 90 GB image will be at 440 MBps, the last 30 GB will drop to 90 MBps or maybe a little higher.
A Windows 7 install that has been optimized for Windows 7 (256 MB static pagefile, disable hibernation 0% restore and recycle bin) takes up about 7 GB of space. So after application installs, 20 GB is easy to do. So copying a Steam folder of 30-40 GB over to your SSD, will be done at probably 60-140 MBs. Why? Not because the SSD is the limit, but your HDD that you backed up the information too is not fast enough to provide the data to the SSD. If coming from a 6-8 drive RAID array.. 440 MBps.
32QD Performance was really impressive, hitting nearly 75K IOPS in writes. What does this mean to the average person? We know that windows file copy is a single threaded application (unless you use RoboCopy and only then from a multi drive RAID array), what it does mean, is that you can perform copies while installing applications, surfing the internet and downloading files from the internet all without a slow down, that is what makes SSDs so amazing. Try doing all that with your HDD. A good comparison is the i7 920. The first CPU capable of transcoding a DVD into a digital file format for storage while playing a heavy physics game while you wait. You can do stuff like that now.
In some ways it is really incredible what OCZ managed to do with a “Marvell at heart” controller and IMFT NAND that has pretty much become the 25 nm standard. Amazing write performance at small file sizes is something that has been the SandForce realm. However the ATTO performance was inconsistent at larger file sizes. The truth is, with these synthetic benchmarks, no one will ever see this kind of performance anyway. Only die hard storage thrashers can expect to actually see the maximum capability of the drive. So OCZ optimized for the typical user. The gamer, the bencher, the seekers of snappy performance. While we didn’t hit .02 access times, we did hit .03 and that is an order of magnitude better than most drives. Also, unlike typical Marvell powered drives, the firmware improvements have done a lot for narrowing the gap between SandForce snappiness without sacrificing Marvell smoothness.
If you are looking for a drive that can bridge the gap between SandForce speed and Marvell’s smooth delivery. The Vertex 4 is the only drive that does that. Excellent small write speeds, good Garbage collection and an incredibly low access time. The Vertex 4 128 GB gets a thumbs up as one of the best small density drives reviewed here at TechREACTION.
The OCZ Vertex 4 receives the Techreaction.net Silver Silicon Award!
For more information on OCZ products, make sure to check out their website at OCZtechnology.com and for personal help from knowledgeable staff and members hit up their website at the OCZ forums.








1 Comment
Try using the latest AMD AHCI drivers with the Vertex 4. I have a Phenom II X4 955 with a Vertex 4 128GB SSD and the AMD drivers doubled the lower ATTO scores compared to the Microsoft AHCI drivers.