Frame Rates Examined
We've also heard postulations that SSDs help improve frame rates, which simply doesn't seem to be the case. Proving that, however, is often challenging. There's variation inherent to every sequence, and of course we can't use pre-rendered cutscenes. The goal, then, becomes finding a consistent metric able to yield the most precise results possible. In our first exploration of graphics performance in Battlefield 3, for example, we found that a 90-second capture from the opening of Going Hunting delivered repeatable scores.
For this comparison, we're switching to the intro for the Thunder Run mission because it's slightly more intensive.
These cursory benchmarks confirm that, when it comes to frame rates, storage is not your bottleneck. Speeding it up won't yield bigger numbers. Almost always, graphics is what holds you back. Less often, processor performance gets in the way. The improvements facilitated by SSDs are seen in game and level loading.
Why, then, do gamers love SSDs beyond those two benefits? Well, solid-state storage does empower your machine with the ability to maintain extremely high throughput as queue depth increases. As a result, you can run into performance-degrading situations when multiple apps are running concurrently, reflected as choppiness when you're playing a game. It's not a constant phenomenon; rather, you see periodic hiccups that interrupt the suspension of disbelief made possible by really great games. You don't see it with an SSD installed. The best (and perhaps most typical) example of this is when an anti-virus app launches a full scan in the background.
This point is largely a non-issue nowadays. Most security suites are smart about the way they start up, postponing activity or throttling their demands if they detect a system load. We decided to force the issue anyway, compelling Norton Internet Security 2012 to scan our drives with the same Battlefield 3 benchmark running.
We end up with slight choppiness when the system is forced to wait on the disk to finish read or write operations issued by the other application. However, between each stutter event, frame rates are going to be the same for SSDs and hard drives. It's the choppiness that causes the average frame rate to drop. With that full virus scan going on in the background and periodic stutters occurring as a result, average performance drops to 53.5 FPS on our hard drive-equipped platform with a GeForce GTX 580 rendering Ultra Quality settings. Compare that to 59.6 FPS on the same machine, unmolested.
Good read. Thanks for being so thorough.
I'm a little confused why you would want to see that comparison. We established there was no diff between a slow HDD and a fast SSD. And you expect a difference between a fast HDD and fast SSD?
He probably want to see that comparison on the "Hard Drive Performance Comparison" page, where is a big difference between SSD and HDD performance (the HDD up to 477% slower)