This feature benchmark dives into one of the top requests we received from our Patreon backers: Undervolt Vega: Frontier Edition and determine its peak power/performance configuration. The test roped us in immediately, yielding performance uplift largely across the board from preliminary settings tuning. As we dug deeper, once past all the anomalous software issues, we managed to improve Vega: FE Air’s power available to the core, reduce power consumption relative to this, and improve performance in non-trivial ways.
Although power target and core voltage are somewhat tied at the hip, both being tools for overclocking, they don’t govern one another. Power target offset dictates how much additional power budget we’re willing to provide the GPU core (from the power supply) in order to stabilize its clock. GPU Vcore governs the voltage supplied, and will generally range from 900 to 1250mv on Vega: FE cards.
Vega’s native DPM configuration runs its final three states at 1440MHz, 1528MHz, and 1600MHz for the P-states, with DPM7 at 1600MHz/1200mv. This configuration is unsustainable in stock settings, as the core is both power-starved and thermally throttled (we’ll show this in a moment). The thermal limiter on Vega: FE is ~85C, at which point the power and clock will fluctuate hard to try and maintain control of the core temperature. The result is (1) spikey frequencies and frametime latencies, worsening perceived performance, and (2) reduced overall performance as frequency struggles to maintain even 1528MHz (let alone the advertised 1600MHz). To resolve for the thermal issue, we can either configure a more intelligent fan curve than AMD’s stock configuration or create a Hybrid card; unfortunately, we’re still left with a new problem – a power limit.
The power limit can be resolved in large part by offsetting power target by +50%. Making this modification is easy and “fixes” the issue of clock-dropping, but introduces (1) new thermal issues – resolvable by configuring a higher fan RPM, of course, and (2) absurdly high power consumption for a non-linear scaling in performance. In order to truly get value out of this approach, undervolting seems the next appropriate measure. AMD’s native core voltage is far higher than necessary for the card to operate at its 1600MHz target, and so lowering voltage improves performance from the out-of-box config. This is for thermal and power reasons alike. We ultimately see significantly reduced power consumption, to the tune of ~90W in some cases, a more stable core clock and thereby higher performance, and lower temperature – and thereby controllable noise.
We can’t get all the way down to the inner workings of the pump on this one, unfortunately, as all of our source images for the Vega: Frontier Edition – Watercooled card are from a reader. The reader was kind enough to remove the shroud from their new WC version of Vega: FE so that we could get an understanding of the basics, leading us to the conclusion that AMD has built one of the most expensive pre-built liquid cooling solutions for a graphics card.
The video tear-down goes into detail on the images we received, but we’ll revisit most of it here. The card uses the same base PCB, same VRM, same GPU/HBM layout and positioning, and same everything as the air-cooled card. The difference is entirely in the cooling solution, where the Delta VRM fan goes away and is replaced with an additional reservoir (more on that in a moment), while the GPU/VRM cooling is handled by liquid plates and a pump. The die-case finstack atop the I/O is also now gone, and the baseplate is simplified to an aluminum plate with no protrusions.
Liquid-cooling the AMD Vega: Frontier Edition card has proven an educational experience for us, yielding new information about power leakage and solidifying beliefs of a power wall. We also learned that overclocking without thermal barriers (or thermal-induced power barriers) grants significant performance uplift in some scenarios, including gaming and production, though is done at the cost of ~33A from the PSU over 12V PSU power.
Our results for the AMD Vega: Frontier Edition liquid-cooling hybrid mod are in, and this review covers the overclocking scalability, power limits, thermal change, and more.
The Hybrid mod was detailed in build log form over in part 1 of the endeavor. This mod wasn’t as straight-forward as most, seeing as we didn’t have any 64x64mm brackets for securing the liquid cooler to the card. Drilling through an Intel mounting plate for an Asetek cooler, we were ultimately able to get an Asetek 570LC onto the card, which we later equipped with a Gentle Typhoon 120mm fan. VRM FET cooling was handled by aluminum finstacks secured by thermal adhesive, cooled with 1-2x Corsair ML120 fans. That said, this VRM cooling solution also wasn’t necessary – we could have operated with just the fans, and did at one point operate with just the heatsinks (and indirect airflow).
This episode of Ask GN returns with our new format, frontloading the episode with some discussion topics before feeding into the user-submitted questions. As always, for consideration in next episode, please leave your comment on the YouTube playback page or in our Ask GN Discord channel for Patreon backers.
The video opens with another “gift” from NZXT, some new power draw testing, AMD Vega naming thoughts (and rushed launches with Intel & AMD), and then addresses user questions. We hop around from liquid metal to CPU and airflow topics, giving a good spread to this episode.
Watch below – timestamps below the embedded video:
We’ve already endured one launch of questionable competence this quarter, looking at X299 and Intel’s KBL-X series, and we nearly escaped Q2 without another. Vega: Frontier Edition has its ups and downs – many of which we’ll discuss in a feature piece next week – but we’re still learning about its quirks. “Gaming Mode” and “Pro Mode” toggling is one of those quirks; leading into this article, it was our understanding – from both AMD representatives and from AMD marketing – that the switch would hold a relevant impact on performance. For this reason, we benchmarked for our review in the “appropriate” mode for each test: Professional applications used pro mode, like SPECviewperf and Blender. Gaming applications used, well, gaming mode. Easy enough, and we figured that was a necessary methodological step to ensure data accuracy to the card’s best abilities.
Turns out, there wasn’t much point.
A quick note, here: The immediate difference when switching to “Gaming Mode” is that WattMan, with all its bugginess, becomes available. Pro Mode does not support WattMan, though you can still overclock through third-party tools – and probably should, anyway, seeing as WattMan presently downclocks memory to Fury X speeds, as it seems to have some leftover code from the Fury X drivers.
That’s the big difference. Aside from WattMan, Gaming Mode technically also offers AMD Chill, something that Pro Mode doesn’t offer a button to use. Other than these interface changes, the implicit, hidden change would be an impact to gaming or to production performance.
Let’s briefly get into that.
Reader and viewer requests piled high after our Vega: Frontier Edition review, so we pulled the most popular one from the stack to benchmark. In today’s feature benchmark, we’re testing Vega: FE vs. the R9 Fury X at equal core clocks, resulting in clock-for-clock testing that could be loosely referred to as an “IPC” test – that’s not exactly the most correct phrasing, but does most quickly convey the intent of the endeavor. We’ll use the phrase “academic exercise” a few times in this piece, as it’s difficult to draw strong conclusions to other Vega products from this test; ultimately, GPUs simply have too many moving parts to simulate easier IPC benchmarks like you’d find on a CPU. As one limitation is resolved, another emerges – and they’re likely different on each architecture.
Regardless, we’re testing the two GPUs clock for clock to see how Vega: FE responds with the Fury X in the ring.
Following our AMD Radeon Vega: Frontier Edition review and preceding tear-down, Buildzoid has now returned to analyze the AMD Vega: Frontier Edition PCB & VRM. This is a 12-phase design (doubled-up 6) that ultimately resembles something similar to a 290X Lightning, making it the hands-down best VRM we've seen on a reference card. Given that Vega: FE is $1000, that sort of makes sense -- but Buildzoid does pose some questions as to what's necessary and how much current is really going through the card.
“Disillusioned and confused” could describe much of the response to initial AMD Vega: Frontier Edition testing and reviews. The card’s market positioning is somewhat confusing, possessing neither the professional-level driver certification nor the gaming-level price positioning. This makes Vega: FE ($1000) a very specifically placed card and, like the Titan Xp, doesn’t exactly look like the best price:performance argument for a large portion of the market. But that’s OK – it doesn’t have to be, and it’s not trying to be. The thing is, though, that AMD’s Vega architecture has been so long hyped, so long overdue, that users in our segment are looking for any sign of competition with nVidia’s high-end. It just so happens that, largely thanks to AMD’s decision to go with “Vega” as the name of its first Vega arch card, the same users saw Vega: FE as an inbound do-all flagship.
But it wasn’t really meant to compete under those expectations, it turns out.
Today, we’re focusing our review efforts most heavily on power, thermals, and noise, with the heaviest focus on power and thermals. Some of this includes power draw vs. time charts, like when Blender is engaged in long render cycles, and other tests include noise-normalized temperature testing. We’ve also got gaming benchmarks, synthetics (FireStrike, TimeSpy), and production benchmarks (Maya, 3DS Max, Blender, Creo, Catia), but those all receive less focus than our primary thermal/power analysis. This focus is because the thermal and power behavior can be extrapolated most linearly to Vega’s future supplements, and we figure it’s a way to offer a unique set of data for a review.
Following our first battery of tests, we dismantled our AMD Vega: Frontier Edition card (which we purchased retail) to get a closer look at the VRM & power design, thermal design, card assembly, and sizes for everything on the board. The tear-down process is the first step to our inevitable hybrid mod of AMD Vega, which should determine the card’s headroom with the thermal limitation removed. We’re also using this as an opportunity to report rough die size measurements, HBM stack measurements, and mounting distances for the community.
Full review testing is still forthcoming, as we didn’t have the usual pre-release embargo period to look things over, but this will serve as our first official Vega: FE coverage. Our next round of coverage will likely be a VRM analysis by Buildzoid, which will be accompanied shortly by thermal/power testing and overclock/gaming testing. Production tests will land in there somewhere – those are already half done – we just need to figure out where they fit best, based on content scheduling.
As we do each week, we’ve rounded-up the major hardware news topics for the past week of industry announcements, all headlined in today’s video. The show notes are also posted in this article, if that format is preferred.
Major news topics seem to pertain to Vega: Frontier Edition – which has had fresh hype attached to it, following a newly published preview – and special edition mining cards, with additional news covering industry topics and launches. SSD and RAM prices remain a mainstay discussion topic for us, though we’ve also got some information on a Blender update, LGA2066 cooler compatibility, SilverStone’s Strider PSUs, and G.Skill’s new memory. More below.