Benchmark Results and Final Analysis
Our standard benchmarks and power tests are performed using the CPU’s stock frequencies (including any default boost/turbo), with all power-saving features enabled. We set optimized defaults in the BIOS and the memory by enabling the XMP profile. For this baseline testing, the Windows power scheme is set to Balanced (default), so the PC idles appropriately.
Synthetic Benchmarks
Synthetics provide a great way to determine how a board runs, as identical settings should produce similar performance results. Turbo boost wattage and advanced memory timings are places where motherboard makers can still optimize for either stability or performance, though, and those settings can impact some testing.
Results using our updated platform and Ryzen 9 7950X showed what Paul’s article proved already, in that Zen 4 is a performant beast. Compared to the 12900K and 5950X before it, in most tests, the new system easily bests the old generation (as it should). Compared to its peers we’ve tested (2 others in addition to this one) the Taichi tends to be around the average of all the boards. Most results are close at this time, and you’d hardly notice a difference between the boards.
Timed Applications
In our timed applications, the X670E Taichi was again average or slightly slower. Here again, the results from our tests are grouped tightly together, so while it was the slowest in a couple of these tests, the difference is at or close to the margin of error.
3D Games and 3DMark
Starting with the launch of Zen 4, we’ve updated one of our games, F1 21 to F1 22, while keeping Far Cry 6. We run the games at 1920x1080 resolution using the Ultra preset (details listed above). As the resolution goes up, the CPU tends to have less impact. The goal with these settings is to determine if there are differences in performance at the most commonly used (and CPU/system bound) resolution with settings most people use or strive for (Ultra). We expect the difference between boards in these tests to be minor, with most falling within the margin of error differences. We’ve also added a minimum FPS value, which can affect your gameplay and immersion experience.
In F1 2022, the X670E and 7950X combo proved to be just a bit faster than the 5950X, though the improvement is less than we’d like to see in that title. Compared to the other boards, the 93/104 frames per second (minimum and average, respectively) is the slowest of the group, but only by a couple of percent. That said, you would likely only be able to see it when you’re looking at a frames-per-second counter on the screen.
In Far Cry 6, this system reached 109(minimum)/124(average) frames per second. This is a much better result than the 5950X and only a frame or two behind the other boards. The Time Spy and Fire Strike results were on par with the group. In short, this board is a performant gamer, too. No worries on this front.
Power Consumption / VRM Temperatures
We used AIDA64’s System Stability Test with Stress CPU, FPU, Cache and Memory enabled for power testing, using the peak power consumption value. The wattage reading is from the wall via a Kill-A-Watt meter to capture the entire PC (minus the monitor). The only variable that changes is the motherboard; all other parts remain the same.
At idle, the X670E Taichi used 104W, and during load testing the system peaked at 293W. Both results are on the high side compared to their peers, but the load wattage is very close to the similarly performing MEG Ace. The idle power use seems a bit high, however. So far, we’ve seen relatively high idle power use on the platform compared to X570 and Z690. These premium boards do tend to use more power at idle as there are more features to power than budget-class boards with fewer extras. We’ll see if that trend continues or if it was just an anomaly from early testing and a few data sets.
VRM temperatures on our Taichi were well within specification during our testing and were nothing to worry about. That said, even though the X670E Taichi has actively cooled VRMs, it ran the warmest in our testing, which was a surprise. After overclocking the CPU to 5.4 GHz on all cores/threads, power use increased marginally and so did the thermals. In this case, our board peaked at 52 degrees Celsius, and the MEG Ace (the only other board we could overclock at this time) ran around 10 degrees Celsius cooler. Still, there’s nothing to be concerned about when it comes to temps here.
Overclocking
Over the last few CPU generations, overclocking headroom has been shrinking while the out-of-box potential has increased. For overclockers, this means there’s less fun to have. For the average consumer, it means you’re getting the most out of the processor without manual tweaking. Our goal in this section is to simply increase the load on the VRMs and see if they can handle the additional stress. Overclocking AMD CPUs can be done in a couple of ways (all-core or adjust PBO values). But for simplicity’s sake, we just went with an all-core overclock of 5.4 GHz with 1.30V to increase the power output.
Since our approach is to add power through all cores, we simply raised the CPU multiplier to 54x, manually set the voltage to 1.30V, and adjusted LLC to minimize vdroop. On the memory side, AMD states the sweet spot is around DDR5-6000, so we dropped our HyperX kit on the board and set the new AMD EXPO profile, and checked for stability.
Overclocking on the X670E Taichi was painless. All of the options you need (outside of LLC adjustment) for this simple overclock are on the same page in the BIOS. We simply set the 1.30V the CPU needs to be stable for our clocks, set LLC to Level 2 to mitigate vdroop, and off we went. VRM temperatures were in order, and the system was stable! This board has no issues pushing the flagship processor to your thermal limits.
On the memory side, we dropped our Kingston Fury Beast DDR5-6000 kit in, enabled EXPO and tested it without issue. There’s still some headroom left on the platform, but as always, your mileage may vary. For the best chance of success, stick to the Memory QVL list.
Bottom Line
The ASRock X670E Taichi proved to be a worthy motherboard compared to the three others we’ve tested so far. AMD’s Ryzen 9 79750X is a stellar performer and the results from our tests landed right around the average, with no seriously off results in either direction. The board comes chock full of features, including the robust 24-phase 105A MOSFETs, dual USB4 Type-C (40 Gbps) ports, a flagship-class audio solution, and plenty of storage options–one of them a Blazing M.2 PCIe 5.0 x4 socket. But it’s not all about the features. The board also looks the part of a premium platform base, with the signature Taichi appearance and a tasteful RGB implementation. For $499.99, it’s a well-rounded option in the premium mid-range space.
If we compare like boards, it competes by price with the Gigabyte X670E Aorus Master ($499.99), and the MSI MPG X670E Carbon WIFI ($479.99). All of these boards include capable power delivery and at least one PCIe 5.0 slot and M.2 socket. The difference comes down to M.2 socket count/type, SATA port count (Taichi has the most here with eight), and audio (Taichi sports the best codec), as the rest of the specifications/features are similar. Among these three boards, the Taichi is the most inclusive option and arguably has the best signature appearance. If you’re not a fan of the mostly black look, the Taichi Carrara with its white marble accents is another option and currently sells for $30 more.
About the only drawback of this board is the price. But to be fair, the $499 price point is $90 less than the Z690 Taichi, which shocked many reviewers and consumers, launching at nearly $600. These new boards are more expensive to make. And sadly, the high prices are likely here to stay as there are few hints the price trends will change anytime soon. If your budget for a new Zen4-based system allows for such a board, the X670E Taichi is a good option among its peers and is my AMD AM5 board of for around $500.
MORE: Best Motherboards
MORE: How To Choose A Motherboard
MORE: All Motherboard Content
