One of the most common pieces of advice when it comes to making your computer faster is “install an SSD”. And sure, this is an useful thing to do: SSDs are blazing fast, compared to mechanical hard drives and help with loading times, opening applications, and accessing your files more quickly. However, I wanted to know more. Specifically, I wanted to know how much an SSD will impact boot times on a brand-new Windows install. To do this, I decided to devise a simple HDD vs SSD showdown.
Since I didn’t want to nuke my SSD just for the sake of running this test, I created a new partition on my SSD and another on a WD Blue 1 TB hard drive that I use for storing documents. On those, I installed Windows 10 and updated both installations to the latest version without installing any additional software. The only exception is an Asus utility called AI Suite that allows to control fan speeds, because my BIOS is stupid and won’t allow me to specify a decent fan profile and the noise was really grating on my nerves. AI Suite doesn’t show up in Task Manager’s Start-up tab, but it’s started via a scheduled task
After having updated Windows, I automated 10 consecutive reboots and logons into Windows, disabling password authentication and creating a scheduled task that was going to take care of rebooting the computer after five minutes. At the end of the sequences of reboots, I extracted data from the Diagnostics-Performance log in the Event Viewer and processed that data.
The second batch of testing entailed seeing how much our average boot times on an empty PC compared against one with common utilities installed. These include an antivirus, a web browser, a copy of Office with Outlook set to run at startup, Skype, and Dropbox. It’s not an unreasonable amount of applications that I know many professionals will install on their computers and that are quite taxing on boot times.
All these tests were carried out on my main rig, with the following specifications:
CPU: Intel Xeon E3 1231-v3 @ 3.40 GHz
Motherboard: Asus B85M-G
RAM: 16 GB DDR3 @ 1600 MHz
Storage: Crucial MX100 256 GB SSD + 2 x WD Blue 1TB hard drives
The partitions I installed Windows into had the following size:
SSD: 31.2 GB (33,554,427,904 bytes)
Hard drive: 87.8 GB (97,371,835,904 bytes)
Our tests highlighted how the SSD fared better than the mechanical hard drive in both scenarios. On average, an SSD with a minimal install of Windows was 2.11% faster than a “cluttered” install on an SSD, 13.81% faster than a minimal install on an HDD and 100.23% faster than a cluttered HDD install.
Interestingly, the minimal HDD install was relatively close to both installs on an SSD, as shown in the graph below:
At the same time, the cluttered install on an HDD proved to be slower than all other installs, from as little as 14% slower to as high as 100.23% slower.
Here’s how the two HDD installs compare:
The really interesting chart, however, is that of SSD performance. The SSD install took a very small penalty when loaded up with software that launches at startup, being 2.11% slower than the minimal install.
The whole data set is available in this Excel spreadsheet (100 KB).
Conclusion and personal observations
The above charts prove that a higher cost per gigabyte of an SSD is greatly counterbalanced by the performance increase in boot times it provides. While it’s true that a 1 TB SSD is too expensive to justify its addition to an office computer, pairing a relatively small SSD (128 to 256 GB in size) for the operating system and applications with a generously-sized mechanical hard drive for data storage is a sensible upgrade to an existing machine and a no-brainer for a newly-built computer.
It should also be noted that, in addition to making the computer boot quicker, the SSD makes the operating system more fluid and responsive across the board, even when it’s still loading start-up applications. By contrast, opening any application (even File Explorer) on an HDD install made the computer freeze for about ten seconds. I didn’t test those slow-downs rigorously because they were out of the scope of this article, but the slowdowns were definitely there.
I would’ve liked to include results for hybrid drives and 2.5″ hard drives as well, but I currently do not have any of those. I would expect 3.5″ hybrid drives to performs somewhere in between an SSD and an HDD, 2.5″ hybrid drives to be slightly worse than a 3.5″ traditional hard drive, and 2.5″ hard drives finishing last by a large margin. This, of course, isn’t the gospel, but nothing more than what I know based on experience and reviews I read. I don’t think I’m very far off, though.