Hitachi 7K250 120GB Performance - HDD Performance in Real-World Applications

HD Performance In The Real World:

Synthetic tests always show a tremendous boost in peformance when running a RAID 0 array.  In this next series of tests we are going to take a look at the performace as it translates to the real world.  We look at loading times of games as a measure of true Hard Drive performance.  Granted there are other factors such as level caching, decompression and the like, but this is primarily the arena of RAID 0 - the gaming community - so we'll see how much benefit we see from RAID 0.  A faster processor and more RAM may make a impact on these scores, so that is why we listed our test setup on the previous page.  You may reference your system in comparison.

Battlefield 1942: Desert Combat Final:

This game is still one of the more popular online games today.  This mod was developed by Trauma Studios and is made public and free to the owners of Battlefield 1942.  While the graphics engine is quite dated, the levels are huge and take a lot of time to load.  This is one game where we should see some improvement using fast drives.

When testing this game, I had set all the detail to the maximum settings and loaded the level 5 times.  Between each test I rebooted the computer to avoid caching the level in memory.  I then threw out the fastest and the slowest times and average the remaining three runs.  Please keep in mind that although the graph looks quite extreme, it is only scaling from 21 to 29 seconds.  There is only a maximum of 4.5 seconds between the fastest setup and the slowest setup.

When loading Operation Battleaxe, the Western Digital drive actually performs the slowest.  Next comes the Hitachi RAID 1 array followed by the single Hitachi drive.  The RAID 0 array takes the fastest time, but the gap is very close - less than 1 second.  Welcome to the real world.

Doom 3:

The next game in our level loading test is Doom 3.  This game has large levels and lots of textures and spends a fair bit of time loading levels as well.  For the Doom 3 test we loaded up the demo1 timedemo using the command "timedemo demo1.demo usecache".  The "usecache" switch loads all the textures of the whole level into memory before running so that there are no pauses as the game loads more information later on.  In this test each configuration was tested 5 times with the fastest and slowest times thrown out and the remaining three runs averaged.  The test rig was rebooted between each run to ensure that nothing was remaining cached.

The Doom 3 tests show virtually the same results as the Desert Combat tests.  The Western Digital drives is the slowest followed by the Hitachi RAID 1 array.  Next in line is the single Hitachi followed by the RAID 1 array.  Although the performace scales logically, there is not much difference in the single Hitachi drive and the RAID 0 array.  In fact the difference is just a little over half a second.  Hard Drive speed is not the only factor when loading Doom 3.

Half-Life 2:

For our last test we'll take a look at Valve's Half-Life 2.  In my opinion this is one of the best games ever made.  The graphics are amazing, the gameplay is excellent, and it scales well on a lot of different machines.  Today we are just taking a look at it from a level loading perspective.  In these tests, all the detail was set at maximum and the resolution was 1024x768.  Again, the level was loaded 5 times per configuration and the system was rebooted between each run.  The fastest and slowest times were thrown out and the remaining three were averaged.

I chose a save-game from the WaterHazard level as it has a bunch of outside textures as well as some buildings, tunnels, and vegitation.  Those of you that have played HL2 can identify that there is a lot of level loading in this game, so this is a game where some fast drives would come in handy.

I was quite surprised at how this test played out.  Again, the Western Digital drive came in last, but this time the single Hitachi came in next.  Both the RAID 1 and RAID 0 arrays outperformed single drives.  The RAID 0 array ran away with this test, and showed very noticeable performance improvement in game.  The first time I ran this test on the RAID 0 array I didn't even get the time because I hadn't expected it to load so quickly.  With a 36% improvement in loading speed over it's nearest competitor and 44% more than a Western Digital drive, we can easily say that the Hitachi RAID 0 array takes the cake in Half-Life 2.

There are some interesting results in the real world tests as well as the synthetic tests.  Head on over to the final page while we recap and draw some conclusions.