Streaming Benchmarks
Note: This streaming section is heavily scripted for the video portion of the review, as we play back clips side-by-side for the benchmark comparison. We will not be removing these references in the article version, so you’ll want to check the video for playback comparisons.
Streaming benchmarks are next. We define this testing in the article linked below. We’ll play some side-by-side clips without revealing the CPUs while explaining this testing. For the basics, we are testing with OBS for capturing gameplay while streaming at various quality settings. Generally, Faster is a good enough H264 quality setting, and is typically what we use in our own streams. Fast and Medium improve quality at great performance cost, but quickly descend into placebo territory at medium and beyond. Still, it offers a good synthetic workload to find leaders beyond the most practical use cases. We are testing with Fortnite and DOTA2 on the 9900K and 2700X when both are stock. Fortnite is set to ‘high,’ as is DOTA2, and both are 1080p and streamed at 60FPS. We also measure baseline performance without any active streams to better understand performance loss from streaming. Streaming is heavily multithreaded, so for people who want a “multi-tasking” workload, this is it.
Streamer-Side FPS in Fortnite
Starting with streamer-side FPS in Fortnite, we observed a baseline framerate of 258FPS AVG when not streaming at all. We were hitting GPU constraints at the top-end in this benchmark. The lows remain moderately timed, at 161FPS 1% and 117FPS 0.1% low. The R7 2700X is next, and operates at 200FPS AVG for its baseline performance. In terms of frametimes, we’re looking at about 3.9ms for the 9900K versus 5ms for the 2700X. With the 9900K streaming at 10Mbps and using Fast H264 encoding, we lose about 27% off the baseline and land at 188FPS AVG. Lows remain spaced proportionately to the average. The 2700X drops to 124FPS AVG, losing about 38% off its baseline. That doesn’t mean it’s worse, necessarily, because we also need to look at viewer-side frame throughput. Remember, there are two pieces to this, and that’ll come up next. At 12Mbps and Medium encode, which increases quality at a huge performance cost, the 9900K falls to 170FPS AVG, with lows now entering worse territory at 43FPS 0.1% low. The 2700X falls to 118FPS AVG.
Here’s a frametime plot of the 9900K at 10Mbps Fast versus 12Mbps Medium and baseline. For these plots, the most important metric is consistency of the plot, alongside lower overall values. Lower is better, but more consistent is better than lower. We are looking at the interval frame-to-frame, with 16ms equating 60FPS. The 12Mbps/Medium result occasionally spikes close to 80ms frametimes, which can create a noticeable stutter or latency difference between surrounding frames.
And here’s the 2700X plot. Performance is overall lower as a result of the lower frequency, but the baseline and 10Mbps Fast results are still relatively consistent. Although we don’t see a spike up to 80ms this time, we must also note that a single frame at 80ms is not the norm on the 9900K, either. We are seeing consistent frame intervals of 40ms on both products when streaming at the torture load.
Let’s take a look at some side-by-side footage of Fortnite at 10Mbps and Fast (see the video for this part). We won’t reveal which CPU is which until the end of the clip. For the streamer, the take-away is that the streamer gets a good experience on both platforms. Intel is technically superior here, but cost is a consideration as well – we’ll talk about that in the conclusion. For now, either platform looks playable to the streamer. What we need to know is if the output is any good to the viewer, as they are the most important part of the equation here. That’s what you’ve been looking at here.
Viewer-Side FPS in Fortnite
Viewer experience is boiled-down to percent of frames delivered at 1080p 60FPS via YouTube. In our testing, both the 2700X and 9900K were able to deliver 100% of frames at 10Mbps and Fast encoding, which is perfectly adequate for any streamer, really. The 12Mbps Medium quality setting is entering placebo territory, though still has some benefits. With Fast encoding the 9900K delivered 87% of its frames within 16.67ms, or 60FPS. The 2700X delivered 96.5% of its frames within the same window. We have seen this behavior before, and have found that Intel stabilizes its delivery when manually managing process priority. This has something to do with task scheduling on each device. With Medium settings, Intel’s 9900K is impressive in its ability to still deliver a consistently good viewer experience at 98% of all frames encoded. The 2700X delivers just 68.4% of its frames in the same test, a result of its lower frequency. Let’s get a side-by-side of these up. Again, with this quality setting, it’s leaning into placebo territory. We want to emphasize that the 2700X is still perfectly good for streaming and gaming simultaneously, you’d just want to keep it to 10Mbps and Fast. Intel can maintain higher quality settings, but it may be an unnecessary level of quality – it just depends on how serious you are about streaming. Ultimately, a secondary system would still improve low-end frametimes.
Streamer-Side FPS in DOTA2
Moving on to DOTA2, we start again with streamer-side FPS. This is what the host of the stream sees. DOTA2 tends to favor Intel CPUs hard for the frequency dependence, following Amdahl’s Law well, and positions the 9900K at the top of the chart. The CPU outputs 191FPS AVG, with lows mixed at 124FPS and 55FPS 0.1%. The 2700X manages 144FPS AVG baseline. When we start streaming, those numbers drop to 151FPS for the 9900K with Fast settings, or 134FPS when under Medium settings. The 2700X drops to 92FPS and 84FPS, respectively. Here’s some side-by-side of the 2700X and 9900K when using 10Mbps/Fast settings. Overall, both are reasonable performers for the streamer. The 12Mbps/Medium frametimes get a bit jumpy, particularly on the 2700X with a 0.1% low of 20FPS, but it isn’t terrible.
Viewer-Side FPS in DOTA2
Viewer-side performance is where it matters. For this one, the 9900K and 2700X both manage to encode 100% of frames at 10Mbps/Fast. This means that viewers will see all 60 frames per second when viewing the stream, and so will not be able to perceive a quality difference between the 9900K and 2700X. They are functionally the same, lending favor to the 2700X for its value proposition. If you did want more quality, the 12Mbps/Medium playback retains a lead for the 9900K at functionally 100% of frames encoded, with the 2700X falling to 92% of frames encoded. You’re looking at side-by-side footage of the 2700X and 9900K at 12Mbps and Medium. We’ll reveal them at the end of the clip. This isn’t terrible, and can be compensated for with a permanent overclock or with a slight reduction in settings. The 2700X does better here than what we saw in Fortnite. Part of this comes down to resource allocation. Frame-capping either game on the streamer-side would also help, although that may be inadvisable for some ultra-competitive players. Let’s reveal those CPUs before moving on.
Streaming Power Consumption
Power consumption while streaming is an interesting topic. This chart is for Fortnite power consumption. With the ASUS Maximus XI Hero that we used for our 9900K, a stricter adherence to Intel’s stock policies means we see a sharp drop-off in power consumption when testing under full stock conditions. The CPU falls to 100W load and stays there, leaving more performance available if we were to remove power targets and limits. Some of the other motherboards shipping today will exit these Intel power specs and draw more power. The 2700X pushes closer to about 120W-125W draw. Intel manages to achieve better overall combined throughput for the player and viewer while maintaining a lower power consumption, for which the 9900K deserves acclaim. It is, of course, significantly more expensive, and that is a big factor to play into our conclusion.
