Page 1:A Lot More SandForce
Page 2:How Can Seven SandForce-Based SSDs Differ?
Page 3:Test Setup And Firmware Notes
Page 4:What's Important: Steady-State Performance
Page 5:Benchmark Results: Storage Bench v1.0 And Real-World Analysis
Page 6:Benchmark Results: 4 KB Random Performance (Throughput)
Page 7:Benchmark Results: 4 KB Random Performance (Response Time)
Page 8:Benchmark Results: 128 KB Sequential Performance
Page 9:Sequential Performance Versus Transfer Size
Page 10:PCMark 7: Storage Suite
Page 11:Final Words
Benchmark Results: Storage Bench v1.0 And Real-World Analysis
As always, we're going to kick off our analysis using Storage Bench v1.0 because it allows us to examine the real performance of an SSD within the first two weeks of use.
IPEAK introduces two time-oriented metrics in addition to speed: busy and service time.
- Busy time is the total time a drive has an operation in progress.
- Service time is a little different. The total service time for the three operations in the diagram is 1785 s, which is longer than the busy time, because the operations are queued three-deep.
Service time gives more weight to periods of high queue depth (QD), while busy time gives more weight to periods of low queue depth. If you're a desktop user, busy time is a more relevant measure because you're less likely to be running workloads that impose a high queue depth.
How do we know that you should be looking at numbers that represent a low queue depth, rather than high-depth numbers most SSD vendors use to reflect peak performance? Look at the distributions of QDs in Storage Bench v1.0 in the chart below.
If you're using a hard drive or an entry-level SSD, you're more likely to encounter queue depths between two and five. But if you're using a mainstream or performance-oriented SSD, you'll mostly see a queue depth of one. This is because operations complete so much faster on a higher-end SSD that they don't have an opportunity to stack up.
Busy time and average data rate are directly tied to one another, but looking at both allows us to examine performance in a slightly different way.
Of the SandForce-based SSDs in the lab, our 240 GB Vertex 3 reigns supreme. Though, if you're comparing the 120 GB drives, Patriot's Wildfire and Mushkin's Chronos Deluxe lead the pack. Both of these drives operate 18% faster than the 120 GB Vertex 3. Interestingly (though perhaps not, given their similar memory technologies), Adata's S511 ties with the Vertex 3 here.
The second-gen SandForce SSDs using asynchronous memory (Force 3, Agility 3, and Solid 3) all fall very close to one another, offering performance similar to Intel's SSD 320. That's almost 45% slower than the Wildfire and Chronos Deluxe. Compared to the ONFi 2.0-based Vertex 3, the SandForce drives equipped with asynchronous NAND are about 20% slower.
Clearly, there's a wide range here. The 120 GB Force 3 turns out to be the slowest of the SF-2200-based SSDs with a data rate of 119 MB/s. At the other end of the spectrum, the 120 GB Wildfire achieves 174 MB/s.
Many of you have requested that we use a 7200 RPM desktop SATA hard drive as an anchor point to put these drives in perspective. If you're still restricted to conventional magnetic storage, take a look at our new numbers. The 2 TB Barracuda XT still runs 66% slower than Crucial's 64 GB m4.
- A Lot More SandForce
- How Can Seven SandForce-Based SSDs Differ?
- Test Setup And Firmware Notes
- What's Important: Steady-State Performance
- Benchmark Results: Storage Bench v1.0 And Real-World Analysis
- Benchmark Results: 4 KB Random Performance (Throughput)
- Benchmark Results: 4 KB Random Performance (Response Time)
- Benchmark Results: 128 KB Sequential Performance
- Sequential Performance Versus Transfer Size
- PCMark 7: Storage Suite
- Final Words