- Packaging and Specification
- Closer look
- Test Bench and Methodology
- AS SSD Benchmark (Pass 3)
- ATTO (Pass 3) Benchmark)
- AIDA 64 Access Test
- Boot Load Test
- TRIM Test
- CrystalDiskMark Benchmark
- HD Tune Pro Benchmark
- PC Mark Vantage x64 HDD Benchmark
- PC Mark 7 HDD Benchmark
- Anvil SSD Benchmark
The front casing is plastic and the base is of brushed aluminium finish. Like a lot of SSD drives, this is made in Taiwan.
After removing the metal base from the rest of the SSD, there’s a thermal pad over the Indilinx controller to aid in heat dissipation.
The controller supports SATA 6Gb/s interface, with TRIM and dynamic/Static Wear levelling protection. Indilinx uses a feature called "Ndurance 2.0" which helps in write data and increases the life-span of the Flash NAND memory. One advantage that Indilinx has over LSI/Sandforce controller is that the controller does not have any limitation- so technically the performance numbers should be same in both compressible and incompressible data.
On the front PCB (left), there’s Indilinx IDX400M00-BC Everest 2 controller surrounded by 8x Intel 25nm BASED 29F64G08ACME3 Flash NAND and Hynix H5TQ2G63BFR- A 128MB DDR3 Cache module as mentioned in their spec. On the other side of the PCB (right) comes with another set of 8x Intel Flash NAND and another 128MB Cache Module. Do note that this Indilinx controller supports from 8 to 16 channels.
When I received the drive, this drive was used for 17 hours. Secure Erase was done before starting with the stress testing and array of benchmarking. Do note that “SSD Life Left” is shown as 64 out of 100.
CDI didn’t give a lot of information, but I would like to point out that, with only 17 hours of usage, the raw value of remaining life shows as 64 when it should have been 100. I am curious to know much of actual host write did he put through to reduce that count before I received it, but the drive’s S.M.A.R.T doesn’t have that option. Do note that the percentage of the Health falls down below 100% when value of “remaining like” shows as 63. In any case keeping this in mind, keep a note on the remaining life.
Before I did any of the stress testing, I did a secure erase using partition magic and started with the endurance testing followed by multiple passes of benchmarks. In the end after constantly testing it for 43 hours, the remaining life count reduced to 62- and health status showing as 98%. I’ve used partition magic again (when I was testing TRIM) and it didn’t bring back the health status. After benchmarks, the drive reduced life down to 60- yikes!
Remember, when I do these series of tests on SSDs, it easily crosses 10-12TB of host write data, even though OCZ drive’ S.M.A.R.T doesn’t show how much write data it has (which also would have reflected how much write data was during the 17 hours of Power On time).
But this isn’t the first 3K P/E cycle drive I’ve tested recently. Kingston HyperX 3K works just fine the way it should- even now and SMART doesn’t show degrade in performance even after stress testing or doing series of tests. Its a speculation from my part, but looking at things there maybe a possibility that garbage collection is aggressive more than it should be.
When I asked OCZ about it, this is the reply I’ve received:
Hardware BBQ’s question is about the NAND Flash P/E count, the original wafer from INTEL or Micron are both provide 3,000 times P/E, after sorting we can have 3,000 ~ 5,000 times P/E to use on our product (Vertex 4)
Which means we’re using the premium level NAND.
There’s also point to note down- and this is clear during HD Tune testing with 64KB and 1MB block size testing.
Although read isn’t affected at all, write is a different story. Refer to HD Tune shown below:
File Size= 64KB Sequential Write (After Secure Erase)
File Size= 64KB Sequential Write (Dirty Drive after 1 hour Idle mode before system restart to test TRIM)
File Size= 1MB Sequential Write (Same condition as above but with larger file size)
The Write performance of the drive, with 64KB and 1MB File size takes a hit at some point after a particular point- irrespective if its a clean or Dirty for waited for a while for TRIM to kick in and the file size that is used, there’s something going on over there.
This time, I’ve used AIDA64’s Linear Write test with 1MB file size and OCZ Toolbox for secure erase (after system restart):
After crossing 50%, the drive’s synthetic benchmarks slump down. Then again, we’re talking about synthetic benchmarks.
One of the best ways to see if this is the issue in real life is by doing a TRIM test- and one of the best ways to test the difference of speed in real life is how RWLabs does it- the Vantage TRIM test.
You do this by keep a track on HDD’s overall score in PC Mark Vantage HDD suite test using the following test condition:
Clean drive (after secure erase)
Leave the SSD idle for 1 hour and then retake HDD test.
In this case, you’re not only testing if TRIM is working and how effective it is since vantage scores reflects real world conditions (almost- but its close enough for you to give you a much better idea) but see if this degrade in graph is reflecting in such tests and if it does, how bad is it. Check the benchmark pages for Trim testing.