How we test
| Storage drive | Crucial RealSSD C300 | Intel X25-M G2 | Corsair Nova V128 | Corsair Extreme X128 | OCZ Vertex | Seagate Barracuda XT 2TB |
|---|---|---|---|---|---|---|
| Drive capacity | 256GB | 160GB | 128GB | 128GB | 128GB | 2,000GB |
| Drive firmware | 1.0 | 02HD | 1.0 | 1.0 | 1.30 | CC12 |
| Approx. price at time of writing | £547 | £352 | £293 | £293 | £295 | £215 |
| Approx price per GB | 2.14 |
2.2 |
2.29 |
2.29 |
2.30 |
0.107 |
| CPU | Intel Core i7 965 Extreme Edition (3.20GHz, 8MB L3 cache, quad-core, LGA1366 - Turbo Boost on) | |||||
| Motherboard | ASUS P6X58D Premium | |||||
| BIOS revision | 0703 | |||||
| Memory | 6GB Crucial DDR3-1,066 | |||||
| Host hard drive | Seagate Barracuda 7200.12 1TB | |||||
| Graphics Card | Sapphire Radeon HD 5850 1,024MB | |||||
| Mainboard software | Intel 9.1.1.1025
and RST 9.6 |
|||||
| Graphics driver | Catalyst 10.2 | |||||
| PSU | Corsair HX1000 | |||||
| Operating System | Windows 7 Ultimate, 64-bit | |||||
Tests
| Benchmarks | CrystalDiskMark 3.0 beta
3 Iometer 2008.06-22 - database and workstation profiles (QD two and 32) HDTach 3.0.1.0 PCMark Vantage 1.0.2.0 File copy test (8.83GB, 544 files) Windows 7 booting test Crysis level-loading test (train) |
|---|
Setup notes
To give some idea of basic performance, we're comparing the Crucial RealSSD C300 against a couple of 128GB SSDs from Corsair - Nova V128 and X128 - along with OCZ's Vertex make it a trio of compared drives featuring the Indilinx controller. We've also thrown in Intel's X25-M G2 and a 2TB mechanical HDD from Seagate for good measure, too.
Firmware 1.0 was used for the Crucial RealSSD C300, together with the latest available for the other drives. We used the garbage collection tool on the Vertex before benchmarking.
We use a Seagate 1TB mechanical hard drive, and it houses the operating system. The benchmarked drives are connected to the motherboard's ICH10R 3Gbps controller should they use SATA2, or the Marvell 6Gbps connector if they're SATA3. All are used as secondary/storage mediums for the first five tests.
Iometer is run by using workstation and database patterns for queue depths (outstanding I/Os) of two and 32, representing very light and moderate loads. Drives with efficient NCQ functions will see performance increase as the queue is deepened.
CrystalDiskMark provides throughput data based on sequential reads and writes, and random (512K/4K) reads and writes. We've used the default 1,000MB file-size for the tests.
PCMark Vantage spits out a bunch of data on the relative speed of the drive(s) when undertaking common tasks.
The file copy test involves copying 8.83GB of data back on to a different folder on the same partition, stressing the read/write ability of the drive.
HD Tach's full variable-zone read/write test is used to write to the drive(s)' blocks and therefore bring the benefits efficient garbage recycling into play. We run this test first to mimic the wear a drive is likely to have during day-to-day operations.
Copying an exact image on to the test drives and using them as the main boot storage, we time how long it takes to load Windows 7 and, once within the operating system, the train level within Crysis, as used in our graphics-card tests.
