Introduction: Drive Specifications and the Test Bed
We all know by now, or should know that SSDs are fast. And they are getting faster according to the specifications. One example is the Corsair Force Series SSDs which are SATA2/3Gb SSDs based on the SandForce 1200 series controller.
I was browsing a popular online retailer recently and I noticed the Crucial C300 128GB SATA3/6Gb SSD listed. Retail at that time was $289. I had also noticed that our Corsair Force Series F120 SATA2/3Gb drive was retailing for $229. Naturally, I wondered what that extra $60 was supposed to buy other than an additional 8GB of unformatted capacity..
The Corsair F120 is spec’d at 285MB/s read speed and 275MB/s write speed. The Crucial drive is rated at 355MB/s read speed and 140MB/s write speed. Benchmarking SSDs can be a tricky business. Pure speed and technical specifications are not the only indicators of the SSD experience. So, I decided to take a look at a variety of indicators to see what these 2 drives offered other then their raw spec numbers.
For my test setup I used the following:
Gigabyte X58AS-UD7 Rev 1.0
Intel Core i7 950 CPU, stock settings, no overclock and cooled with a Corsair H70
Corsair CMP12GX3M3A1600C9 Memory Kit
Corsair AX850 PSU
ATI HD5970 GPU
Corsair Force Series F40 SSD for the system drive
The OS is Windows 7 64bit with all current updates I used the 10.0.0.1046 driver for the Intel ICH10-R, the most current driver at the time of testing. The board was flashed with the most recent F8e beta BIOS which, according to Gigabyte provides improved SSD support and a new firmware for the Marvell 9128 SATA3/6Gb drive controller. All Force series drives used have firmware 2.0. The Crucial drive carried the most recent firmware available at the time of testing.
I have bit of skeptic in me and it emerges each time a new computer related specification is released. As we all know, specifications are important but do not always show the true measure of the actual performance of a computer component.
I remember a lot of people getting excited back in the time when the ATA133 standard was certified and released. Interestingly enough, the first SATA drives appeared around this same time. ATA133 drives just had to be faster than ATA100 drives…...right? Wrong! The ATA133 drives listed a higher specification but did not perform any better than the ATA100 drives. So of course I began to wonder if that would hold true again.
Here are the specifications for each drive from their respective product pages.
- Maximum sequential read speed 285 MB/second
- Maximum sequential write speed 275 MB/second
- Random 4K write performance of 50,000 IOPS (4K aligned)
- Latest generation SandForce controller and MLC NAND flash for fast performance
Crucial C300 128GB SSD
- Read speeds up to 355MB/s
- Sequential Access - Write up to 140MB/s
- SATA 6Gb/s interface
- High-speed Synchronous NAND
- Random 4K Read/Write: 60K / 30K IOPS
For more information on the SATA2/3Gb or SATA3/6Gb specification, check with the Serial ATA International Organization.
Testing and Benchmarking the Drives
Corsair has used ATTO to validate SSD performance since our initial entry into the SSD market 2 years ago. According to Crucial, they have validated their drive speeds using the HD Suite from Futuremark’s PCMark Vantage. I decided to test both drives with each benchmark. Here are the screenshots from ATTO. ATTO and all benchmarks used in this testing were run 3 times with the high and low run dropped. The machine was rebooted between every benchmark run.
As you can see, both drives maxed out the READ measurement. However, the Force drive is clearly faster in the WRITE measurement.
Due to the Crucial drive not meeting specification on READs, I decided to check both drives again with ATTO using the Marvell 9128 SATA3/6Gb controller. The maximum write speed with this controller for the Crucial drive was 322638MB/sec transfer rate. However, the WRITE speed dropped almost 10% using Crucial drive on the Marvell controller. The Corsair Force drive results were within 2%-3% on both controllers with a slight overall drop on the Marvell controller.
PC Mark Vantage - PC Marks
I tested both drives using PCMark Vantage. From the Futuremark website: A PCMark score is a measure of your computer’s performance across a variety of common tasks such as viewing and editing photos, video, music and other media, gaming, communications, productivity and security. I ran the benchmark using the Marvell controller and the Intel ICH10-R. In this test, the Intel ICH10-R was the superior performer over the Marvell controller. The Force drive performed approximately 25% better on the Intel controller and the C300 drive performed 16% better on the Intel controller. The Intel controller results are shown here. The Force drive is best in both the overall PC Marks score and the HD Suite
Crystal Disk Mark
PC Mark Vantage—HD Suite
Here are the actual raw numbers PCMark Vantage compiles and then uses to determine a score for the drive suite. In this simulated actual usage, the Force drive is clearly the better performer. According to Futuremark, this should be a good indicator of real world usage. Results are shown using the Intel ICH10-R controller.
Crystal Disk Mark
Crystal Disk Mark uses a different default testing methodology from ATTO. The default test is a random mix of 0 and 1 data. You can also select a 0 Fill or 1 Fill test. The C300 is a strong performer here. We’ll need to do some more testing to see if this translates to real world performance or in other benchmarks that use different methodologies and data types. F120 results are on the left, C300 results are on the right.
Corsair F120 Crucial C300
File Transfer Test
For the file transfer test, I added a 2-drive RAID-0 stripe using 3 Force Series F40 SSDs to the ICH10-R. I then created a 5GB test file from my server using a combination of Word documents, Excel documents, uncompressed photos, MP3s, AVIs, and Matroska movie files. I also included .zip files comprised of these files to be sure to include a variety of compressed and uncompressed data types. I transferred them to and from the Force 120 and the C300 using the RAID-0 stripe. For the test, I copied and pasted the individual files, not the folder in which they were contained.
The transfer times for both the F120 and the C300 were within 2 seconds of each other in both directions and averaged from 31 seconds to 33 seconds reading from and 33 seconds to 35 seconds writing to. I hand timed each transfer using a stopwatch. The time difference in each was insignificant and well within the margin of error of my thumb on the stopwatch. .
So, what does that extra $60 for the SATA3/6Gb drive get a person aside from an extra 8GB of unformatted space? It appears that it simply buys a newer specification.
The first thing I discovered is that while the Crucial drive does test faster on the Marvell controller in WRITE speed, the READ speed suffered approximately an equal amount. To date, I have only tested this single motherboard using a SATA3/6Gb but I tried it with 3 BIOS revisions and 2 different driver versions. This may be an indicator that the current SATA3/6Gb controllers are not ready for prime time. If this is the case, any implied advantage of current SATA3/6GB drives cannot be fully utilized.
ATTO and Crystal Disk Mark are both very valid benchmark tools for pure speed. However, pure speed is not the only measure of SSD performance. ATTO uses only 0 Fill data which works well with the SandForce Durawrite Technology. This technology involves a patented form of data compression. According to SandForce, DuraWrite technology extends the endurance of MLC memory providing at least 5 year lifecycles w/ 3-5K cycle MLC Flash. Over the long term, his technology should easily offset the slower benchmark results we see in the Random and 1 Fill tests using Crystal Disk Mark and many other benchmark utilities tested but not shown here.
In the PCMark Vantage tests, the SandForce based Force 120 was the clear winner in both overall PC Marks and in the HD Suite. Purely numeric benchmarks are only theoretical indicators of actual performance as we noted at the beginning of this article. However, the PCMark Vantage benchmark was chosen for its ability to emulate real world usage. The file transfer test was a general illustration of the blazing speed of SSDs. Moving 5GB of data in 30 seconds is spectacular on both drives and especially so when compared to standard hard disk drives. The SandForce based Force 120 drive showed itself to be the equal of the Crucial 300 in spite of the Crucial bearing the newer specification. Additionally, the Force Series drives will offer superior performance for users not yet owning motherboards with SATA3/6Gb controllers or a PCI-e SATA3/6Gb controller.