Benchmark Results & Final Analysis
Starting with the memory benchmarks, we can see that the B660M Steel Legend is second only to the Z690 Extreme WiFi 6E in most metrics.
While all three motherboards are capable of producing similar performance throughout 3DMark, the problem for the competing Asus board is that its default 1:2 (aka Gear 2) memory controller ratio is…the default. That won’t matter to anyone running slower memory, but those who want the fast stuff will want to remember to use manual configuration.
PCMark favors ASRock slightly, regardless of chipset. That could be a big deal to anyone who desires the cost savings of the cheaper board.
Compression under 7.Zip was always a big win for DDR5, but we see hardly a difference between various models of DDR4 motherboard.
Cinebench and Corona Benchmark favor the newer B660M Steel Legend for unknown reasons that might have something to do with its newer firmware revision.
The B660M Steel Legend had the shortest completion times in several real-world tests, though the difference between all three boards is marginal.
The stripped-down B660M Steel Legend starts off with lower idle power than ASRock’s own Z690 Extreme WiFi 6E, but ends up consuming more power at full load. We attribute the lower idle power to its smaller onboard component group and higher full load power to a deficiency in voltage regulator efficiency.
The B660M Steel Legend’s voltage regulator thermal protection kicked in after around ten minutes of Prime95 small FFTs (the power-hungriest CPU test) when loaded with our DDR4-4000 and Core i9-12900KF. Believing that this was an edge case, we successfully ran our full test duration first with our memory’s XMP enabled, and then with a set of DDR4-3600 at full XMP settings.
The highest temperature we measured was nearly 105° (ambient temperature is subtracted in the above chart), and that might not even be the hottest part of the board. The great news is that the board survived by simply limiting ourselves to DDR4-3600, which is probably faster memory than any low-cost motherboard buyer would use, but the terrible news is that our fix barely put it below the reset threshold.
The B660M Steel Legend started off with higher power consumption than the Z690 Extreme WiFi 6E, and the latter’s slight performance lead enlarges its efficiency lead.
A low price gives the B660M Steel Legend a big performance per dollar win, but these wins are always diminished when the cost is combined with other system components. Most users will likely pair the B660M Steel Legend with a cheaper processor and RAM, as well they should since it couldn’t consistently run our Core i9-12900KF/DDR4-4000 combo under Prime 95 small-FFTs for more than 10 minutes.
|ASRock B660M Steel Legend|
|With great performance and a low price, the B660M Steel Legend appears to be a great value for builders who will not be stress testing the fastest-available hardware.|
Pushed to its limit by our Core i9-12900KF, the B660M Steel Legend could offer ideal value to anyone who’s using lower model LGA-1700 processors from the humble Celeron G6900 to the lofty Core i7-12700 in a properly vented case. Its ability to repeatedly protect its circuitry undamaged from overloads means that we can’t recommend against its use by budget builders who understand its limitations and never plan on exceeding those.
TLDR On Cooling: Our motherboard test system is designed and tested to mimic the voltage regulator temperatures of a properly vented mid-tower case with single-120mm-fan cross-flow CPU cooler. Anyone who says that we’ve greatly understated the B660M Steel Legend’s VR thermal issues probably tested theirs without a real-world level of airflow.