In our recent look at the Vertex 4 SSD equipped with firmware 1.4, we found that the 128 GB drive's write performance depended heavily on available capacity. OCZ disagreed, but then quickly rolled out firmware 1.5. So, we're back to test the new build.
A couple of weeks back, we followed up our OCZ Vertex 4 Review: A Flagship SSD Powered By...Indilinx? story with hands-on time in the lab using a newer firmware version. Normally, it'd be difficult for us to retest a piece of hardware we already reviewed just because the vendor introduced fresh software for it. However, OCZ made big claims about the gains expected from its 1.4 release. We wanted to know more.
In OCZ Vertex 4 128 GB: Testing Write Performance With Firmware 1.4, we confirmed notable improvements in sequential write performance using our 128 GB sample. However, we also discovered that, after writing to more than 50% of the drive's capacity, throughput fell off significantly.

OCZ provided us with an explanation for this behavior, simultaneously suggesting how it procured its extra performance and why speed tapered off after the 128 GB model, specifically, filled beyond 50%. To paraphrase, the company noticed that many folks were leaving lots of space unused on their SSDs. So, it could afford to change the behavior of its garbage collection algorithm depending on available capacity. Below 50%, it'd operate in a "performance" mode. Otherwise, it'd switch back to a more conservative "storage" mode.
One of our goals here at Tom's Hardware is to test the latest hardware thoroughly, so that you can be sure you have a complete look at performance before making a buying decision. OCZ's claim that the Vertex 4 should recover compelled us to collect even more data. We started with a new procedure suggested by the staff at OCZ. But then the company released a new Vertex 4 firmware: a beta version of version 1.5. A week later, it uploaded a final build of 1.5. It's currently available as a stable release on OCZ's support site. Now, we have a couple of new variables to put through our gauntlet.
Vertex 4: Firmware 1.5
OCZ lists the following enhancements to firmware 1.5:
- Improved sequential file transfer performance for 128, 256, and 512 GB models
- Optimized idle garbage collection algorithms to extend the benefits of performance mode by enabling the feature across a greater percentage of the drive
- Improved HBA/RAID card compatibility
- Further improved compatibility with desktop and mobile ATA security features.
And the fixes:
- Corrected a corner case issue where the "Remaining Life" SMART attribute could be reported incorrectly.
In essence, it appears that OCZ took feedback on its performance and storage modes to extend the benefits of the former. Naturally, then, we're interested in just how the drive's performance attributes have changed. And of course, the compatibility improvements are certainly good news to anyone who might have been having trouble with a Vertex 4 previously.
- OCZ Vertex 4 Write Performance, Revisited
- An Updated Test Procedure
- Firmware 1.4.1.3: hIOmon Disk I/O Ranger
- Firmware 1.4.1.3: AS SSD Compression Benchmark
- Firmware 1.5: HD Tune And hIOmon Disk I/O Ranger
- Firmware 1.5: AS SSD Compression Benchmark
- Firmware 1.5: Iometer Test File Monitoring
- Our Theory About How v.1.4 Changed Write Performance
- Firmware 1.5: Something Special For OCZ's Vertex 4
the rest of the pages show "Richard Hart"
how much idle time is needed for the data rearrangement to take place?
And what is the authors recommendation on a Vertex4? should a user buy Vertex 4 over a Samsung/Sandforce?
Does the performance mode mean that the wear leveling is constrained to the first bit of every cell and therefore the drive wears out quicker compared to normal mode?
It is actually occuring with update 1.4. Hmm, damn I need to do another clone disk before doing the update.
Do you dislike manufacturers improving their products without demanding that you pay more money for the improvements?
How many additional firmware updates are needed before OCZ gets it right?
such a stupid comment. don't download any more updates from nvidia or amd then!
i agreed and i gave you a thumbs up to bring you back to level insead of -1, some people are just ignorant.
The average user and even somewhat above average users probably won't be able to effectively use very high end SSDs any better than much lower end SSDs. Many people can use the performance advantage, but most people probably won't unless they really try to in an unrealistic situation, such as a storage benchmark. End users who do more than minor internet browsing and such can find the performance advantage of high end SSDs almost vital to their usage.
Are you also confused about 4 bedroom houses, cars, and gourmet cooking? Because there are some people who live alone, prefer to walk, and eat salads.
Does the performance mode mean that the wear leveling is constrained to the first bit of every cell and therefore the drive wears out quicker compared to normal mode?
I think you are mis-interpreting that a bit. I would look at it as an either or scenario -- either you keep the disk less than half full, or you suffer from the drive wearing out faster. Only 1 of the 2 will occur.
Having said that, it does raise some concerns for me when the drive is more than 1/2 full. Realistically speaking, I think the market would be better served if manufacturers simply go with 2 tiers of SSD drives -- one using MLC for slower but high capacity drives, and the other using SLC for faster but lower capacity drives. Using MLC chips and making them behave like SLC at the cost of significantly reduced life seems like a poor idea to me.
The lifetime is unlikely to change because of this. The NAND flash cells are almost always the longest lasting part of a drive by far. If anything, the drive would probably fail due to the power circuitry, firmware, controller, etc. etc. before flash cells start failing.
first let me call you ASSHOLE.
second, let me answer: SSD's are complicated and still evolving storage instruments, you can't find 2 web sites testing them the same as you do with video cards, results change with each firmware update and from brand to brand even when the same controller and nand is used. If the shit isn't confusing to you then pint a rose on your nose but I wasn't an early adopter and haven't followed the ins and outs over the last 4 years. Now that i am considering an SSD I would like to find out from people who have paid attention what the deal is.
No i don't get confused by 4 bedroom houses, cars, and gourmet cooking? that i've visited before, driven before or ate before.
Smart ass.
another patch or 10
this is why i stick to Intel SSD's - they WORK