Picking Up The Pieces
Our research into the performance problems of recent T-Force DDR4 on MSI motherboards led us nowhere, as neither MSI nor Team Group could figure out the problem. Team Group grew increasingly worried as our Nighthawk Legend RGB review sat in the publishing queue with a “Hold” tag, and we became even more restless after finding that builders within the enthusiast community hadn’t even noticed the problem. Team Group suggested we retest with an Asus board, but we couldn’t hang the problem on MSI alone: We had to find out if this was merely an MSI-only problem, or an Asus-only solution. And for that, we needed more boards.
Our first set of tests was to make sure the performance problem wasn’t specific to the Z370 Godlike Gaming, so we threw a few benchmarks at the MPG Z390 Gaming Pro Carbon and found no noticeable improvement. We then went step by step through every available memory setting to find any solution. With yet another day of testing down the tank, we reverted to the original board’s results so that we could tie the original benchmark results to today’s retest, where the Maximus XI Hero takes on the Z390 Designare and Z390 Taichi. That retest begins with an overclocking comparison.
The Z370 Godlike Gaming had a slight edge in memory overclocking and had operated for months without any performance problems, so you can see why it became our choice for memory testing before the arrival of Teamgroup’s recent kits. But can we hang Team Group’s performance problems on MSI?
More bandwidth is better. Less latency is better. Sandra shows the T-Force Night Hawk Legend RGB at or near the top of the performance scale on the Godlike Gaming as well as the Maximus Hero, but we remember that F1 2015 and 7-Zip were the weak spots.
Shocker: The Asus Maximus XI Hero extracts the best performance from Team Group’s T-Force Night Hawk Legend RGB. Double-shocker: Gigabyte’s modules also perform nearly 25% better in F1 2015 when installed on an Asus board than on a Gigabyte board. Ouch! And the most motherboard-agnostic kit came from Adata!
Since we’re using a fixed CPU clock, there’s not much to see in the control benchmarks. Moving on…
Lower is better for a timed test where the Night Hawk Legend RGB performed well only when paired with an Asus motherboard. Now notice the Adata results again: MSI, Gigabyte, and ASRock all beat Asus there. If we worked for Asus, we’d start telling review sites to test with Team Group memory…
You’ll notice that the Adata kit takes second place despite being the most consistent across motherboard brands, but there’s a reason for this: Its DDR4-3200 data rate is chocked by 16-18-18 timings. Aorus DDR4-3200 has those same timings and, when paired to the Asus board, gets the same score. The G.Skill and Team Group kits have straight 14-14-14 timings, and at least the Asus board shows these like timings offering an identical performance advantage.
Where do we go from here?
Had we been using an Asus board from the outset, we might never have known about the performance problems T-Force Night Hawk Legend RGB (or even the previous XCalibur RGB) has with other brands of board. Those differences can be extreme in certain software suites, particularly F1 2015 (which is typically indicative of everything that uses the EGO engine). And it’s not like 7-Zip was easy on the T-Force kits, either.
It’s shocking to think that we may have issued either of the previous two Team Group kits an award, or even a four-to-five star rating, based solely on Asus-based benchmarks. Doing so would have left readers with ASRock, Gigabyte, and MSI with a false impression. But, given our research into T-Force users, they might not have noticed all that lost performance.
Thus, we turn to you: We can’t go on testing every memory kit on every brand of board, particularly when we’re hoping to continue including performance at a variety of data rates. Running a benchmark series that takes a month to conclude would be a disservice to both our leadership and our readers as the number of products reviewed would decrease drastically. Perhaps a combination of half as many settings with twice as many boards is required? We look forward to your thoughts, but until then, the T-Force Night Hawk Legend RGB gets a bump from the expected 2/5 stars to 3/5 stars, with an exclusivity warning in its verdict.