Wednesday, September 14, 2011

HDD Upgrade

As mentioned in a previous post, my current desktop PC at home was (and still is) using storage devices from the previous system. However, the 160GB Seagate drive (a ST3160023AS) where the Windows 7 installation resided had become extremely tight for space.

I had originally intended to save enough money to purchase a large capacity mechanical hard disk alongside a smaller SSD; installing the OS and a select few apps onto the SSD and everything else on the mechanical drive. However, I had decided that the smallest capacity SSD I would be able to work with would be 128GB because of Windows 7's large install footprint, which would only increase following system updates. At the time I was investigating the upgrade, a good 128GB SSD would have cost me approximately GBP 150.00; an expense I couldn't justify, so I decided to try and hold out.

Eventually, the existing hard disk was filled to capacity, and the cost of an SSD hadn't dropped. In fact, at the time I could barely afford a regular hard disk! Fortunately, I had £40 worth of Amazon vouchers so I began searching for an upgrade and in the end I settled on a Seagate ST31000524AS, which has an unformatted capacity of 1TB. My ideal upgrade would have been a 2TB drive, but I could only spare an additional £20, which limited me to models that operated at 5400 RPM. While my primary concern was additional capacity, I was also conscious of drive performance and that a drive that spins at 7200 RPM would give me much better performance, hence the reason I opted for the 1TB Barracuda (which actually spins at 7200.12 RPM).

Before and after migrating my OS over to the new HDD, I did some tests to compare the performance of each drive. A bit unfair on the older drive as it's a SATA 1.0/1.5Gbps model, compared to the new SATA 3.0/6Gbps drive, but I was interested to see the difference. Each test was conducted 5 times and the median result used for comparison. Before performing the tests, I prepared my machine by running through a pre-flight checklist to help ensure I would achieve optimum performance (thanks to Maximum PC for originally publishing their own checklist for me to work from). The steps that I took were:
  1. Turned off the screen saver.
  2. Turned off power saving modes; i.e. I switched the OS to "high performance" mode.
  3. Disabled the network adaptor/disconnected from the internet.
  4. Disabled antivirus apps.
  5. Turned off autoupdate.
  6. Defragged my hard drive.
  7. Disabled System Restore
  8. Rebooted.
  9. Logged in as "Admin"
  10. Waited a few minutes until disk activity had subsided.
  11. Ran ProcessIdleTasks; this involved spawning a “DOS box” with administrative privileges and typing: “Rundll32.exe advapi32.dll,ProcessIdleTasks” This forces Windows to perform all of the tasks it would normally do when the system is idle and once the system activity had died down, I was ready to benchmark.
Synthetic Tests
First, I tested the drives using the "Physical Disks" benchmarks from the SiSoftware SANDRA Lite 2011 SP1 Suite (specifically, version 2011.2.17.36). Obviously, this doesn't demonstrate real world usage, but it gives a good idea of roughly how much faster the new drive is.

The newer drive performed over twice as fast as the older one in the throughput tests, which was expected as my motherboard only supports the SATA 2.0 standard, which would limit the throughput of the newer SATA 3.0 HDD. What's particularly interesting about the results is that, despite the newer hard drive achieving faster write access times, the old drive actually posted slightly better read access time results.

Real World Tests
First, file copying; I had both drives divided into two partitions and created a 10GB file which I timed how long it took to copy from one partition to another. The results were impressive:

As you can see from the graph, the new drive copied the file more than 6 times as fast as the old; a vast improvement!

The next test was to see the difference in boot times. What I did here was time how long it took from the computer being powered on to the Windows log in screen being displayed. Not the most accurate of tests, but the difference between the old and new HDDs was much greater than a couple of milliseconds, I consider the margin of error to be acceptable.

Not as impressive as the copying results, with only around 20 second reduction in boot time. However, it's worth noting that approximately 20 seconds of the total boot time is devoted to the motherboard's BIOS POST, with the remaining time allocated to Windows itself booting. Once this is factored in, the new hard drive actually does live up to the expectations raised following the synthetic tests; that the drive should be twice as fast.

Despite being unable to afford an SSD, I am very pleased with the new hard drive. Not only is it much faster than the original drive Windows was running from, but it's so much quieter as well! Depending how the cost of SSDs changes over the next year, I may choose to upgrade another component in my machine instead, most likely the graphics card.