Frametimes represent the time taken to render individual frames through the course of the benchmark. While framerates are an indication of overall performance, frametimes are used to gauge the frame-to-frame smoothness of the experience. It’s useful for diagnosing stutters and impromptu drops in framerates.
At 4K RT, the RTX 4070 and 4080 deliver smooth frametimes but the RTX 3070 Ti runs into multiple spikes through the course of the benchmark.
Disabling raytracing reduces the load on the RTX 3070 Ti, resulting in smoother frametimes, nearly identical to the RTX 4070. Unsurprisingly, the RTX 4080 offers the best performance with little to no bumps. Cyberpunk 2077 literally stalls the render pipeline of the RTX 3070 Ti at times with frame times varying from 250 to 380 milliseconds. The RTX 4070 seems minor spikes while the RTX 4080 is smooth as butter.
Disabling raytracing allows the RTX 3070 Ti to deliver playable frametimes nearly on par with the RTX 4070, while the 4080 continues its winning streak.
At 1080p (with raytracing), the RTX 4070 and 4080 deliver similar frametimes while the 3070 Ti runs into a spiky patch in the first half of the test. We see similar spikes in Dying Light 2 at 4K with raytracing. The RTX 4070 and 4080 offer relatively smoother frametimes.
Disabling raytracing eliminates most of the frametime spikes for the RTX 3070 Ti. It is still notably rougher than the RTX 4070 and 4080.
The Callisto Protocol is a stuttery mess at 4K with raytracing enabled. The RTX 4070 makes for a relatively smoother experience than the 3070 Ti while the RTX 4080 delivers the best performance.
At 4K (with RT disabled), the RTX 3070 Ti and 4070 offer similar frametimes with the RTX 4080 once again distinguishing itself as the superior GPU. Shifting gears to 1080p Ultra, we see a repeat of the 4K chart, albeit with smaller spikes.
Power, Thermals and Clocks
Continued on the next page…