After completing all of our gaming, power, thermal, and other benchmarks for the new nVidia Titan V graphics card, we took the unit apart for cooler, PCB, and VRM analysis. We’ll be joined by overclocker ‘Buildzoid’ in the next few days for the advanced overclocking analysis of the PCB and VRM, but have some immediate information on the assembly of the Titan V and its cooler.
The card follows the same screw pattern as all previous nVidia Founders Edition cards, including the Titan Xp and GTX 1080, primarily isolating its cooler and shroud into a single, separable unit. Build materials are all the same, assembly is the same, but the underlying GPU, HBM2, VRM, and heatsink are different.
Vega 64 may consume more power than a GTX 1080, but until now, we haven’t known if that impact is relevant to room temperature. That’s what we wanted to know, and we eventually expanded that concept to include how much a 900W+ mining machine increases room temperature, a 600W machine, and so on. We were able to effectively replace any need of a heater for the past week, and right when it started to get colder.
In this test, we’re looking at the room ambient impact of various PC builds. This helps to conceptualize the real-world impact of all those power and thermal tests you see us (and others) publish, as it puts real numbers to the user experience outside of the case. Although this concept has about a million variables and “what ifs,” we controlled to the best of our abilities, are laying-out all the major variables, and can present an academic experiment that demonstrates room temperature increase from computer equipment. All watts are basically created equal, for the purposes of this test: A 940W mining rig will output just as much heat into the room as a 940W gaming rig, or a 940W rendering machine, and so forth; as long as the power load is equal between all of these (read: constant), watts are watts, and you can extrapolate room temperature for each type of machine.
The testing originally was concepted after our Vega 56 Hybrid mod, which used power mods and other mods to push the card up towards 400W of power consumption. We wanted to test a straight Vega 56 versus GTX 1070 for room ambient impact, but shifted that up a tier (to Vega 64 and a GTX 1080) for some parts that are more likely to show a difference. After that, we shifted up to a 940W mining machine, then picked a middle-ground ~600W machine (which could also represent SLI gaming or HEDT render systems).
NVidia introduced its new Titan V GPU, which the company heralds as the “world’s most powerful GPU for the PC.” The Titan V graphics card is targeted at scientific calculations and simulation, and very clearly drops any and all “GTX” or “gaming” branding.
The Titan V hosts 21.1B transistors (perspective: the 1080 Ti has 12B, P100 has 15.3B), is capable of driving 110TFLOPS of Tensor compute, and uses the Volta GPU architecture. We are uncertain of the lower level specs, and do not presently have a block diagram for the card. We have asked for both sets of data.
We need some clarity on this issue, it seems.
TLDR: Some AMD RX 560 graphics cards are selling with 2 CUs disabled, resulting in 896 streaming processors to the initially advertised 1024 (64 SPs per CU). Here’s the deal: That card already exists, and it’s called an RX 460; in fact, the first two lines of our initial RX 560 review explicitly states that the driving differentiator between the 460 and 560, aside from the boosted clocks, was a pre-enabled set of 2CUs. The AMD RX 460s could already be unlocked to have 16 CUs, and the RX 560 was a card that offered that stock, rather than forcing a VBIOS flash and driver signature.
The RX 560 with 2CUs disabled, then, is not a new graphics card. It is an RX 460. We keep getting requests to test the “new” RX 560 versus the “old” RX 560 with 1024 SPs. We already did: The RX 560 review contains numbers versus the RX 460, which is (literally) an RX 560 14CU card. It is a rebrand, and that’s likely an attempt to dump stock for EOY.
Jon Peddie Research reports that the AIB market is likely returning to normal seasonal trends, meaning the market will be flat or moderately down from Q4 2017 through Q1 2018.
In a typical year, the AIB market is flat/down in Q1, down in Q2, up in Q3, and flat/up in Q4. The most dramatic change is usually from Q2 to Q3, on average a 14.4% increase (over the past 10 years). Q3 2016 was roughly twice that average with more than 15 million AIBs shipped, 29.1% more than Q2 and a 21.5% increase year-over-year.
After a year of non-stop GPU and CPU launches, a GPU round-up is much needed to recap all the data for each price-point. We’ll be looking at strict head-to-head comparisons for each price category, including cards priced at $100-$140, $180-$250, $400-$500, and then the Ti in its own category, of course. As noted in the video, a graphics card round-up is particularly difficult this year: Chaos in the market has thrown-off easy price comparisons, making it difficult to determine the best choice between cards. Historically, we’ve been able to rely on MSRP to get a price (+/-$20, generally) for comparison between both AMD and nVidia; the partners hadn’t strayed too far from that recommendation, nor the retailers, until the joint mining & gaming booms of this year. Fortunately, much of that pandemonium has slowed down, and cards are slowly returning to prices where they sat about 6-8 months ago.
Another point of difficulty, as always, is that price-matched video cards will often outperform one another in different types of workloads. A good example would be Vega vs. Pascal architecture: Generally speaking – and part of this is drivers – Pascal ends up favored in DirectX 11 games, while Vega ends up favored in asynchronous compute workload games (DOOM with Vulkan, Sniper with Dx12). That’s not necessarily always going to be true, but for the heavyweight Vulkan/Dx12 titles, it seems to be. You’ll have to exercise some thought and consider the advantages of each architecture, then look at the types of games you expect to be playing. Another fortunate note is that, even if you choose “wrong” (you anticipated Vulkan adoption, but got Dx11), a lot of the cards are still within a couple percentage points of their direct-price competition. It’s hard to go too wrong, short of buying bad partner cooler designs, but that’s another story.
Almost as painfully as for our DDR4 RAM sales article, we trudged through video card sales and “sales” alike in attempt to find gold in a strapped market. Video card sales weren’t as exciting as previous years, with some settling down to simply MSRP – in other words, half off – and others seeing $10-$20 drops. We found a couple of good ones, nonetheless, including GTX 1080, RX 570, RX 560, and GTX 1060 sales. Find those below.
Having gone over the best CPUs, cases, some motherboards, and soon coolers, we’re now looking at the best GTX 1080 Tis of the year. Contrary to popular belief, the model of cooler does actually matter for video cards. We’ll be going through thermal and noise data for a few of the 1080 Tis we’ve tested this year, including MOSFET, VRAM, and GPU temperatures, noise-normalized performance at 40dBA, and the PCB and VRM quality. As always with these guides, you can find links to all products discussed in the description below.
Rounding-up the GTX 1080 Tis means that we’re primarily going to be focused on cooler and PCB build quality: Noise, noise-normalized thermals, thermals, and VRM design are the forefront of competition among same-GPU parts. Ultimately, as far as gaming and overclocking performance, much of that is going to be dictated by silicon-level quality variance, and that’s nearly random. For that reason, we must differentiate board partner GPUs with thermals, noise, and potential for low-thermal overclocking (quality VRMs).
Today, we’re rounding-up the best GTX 1080 Ti graphics cards that we’ve reviewed this year, including categories of Best Overall, Best for Modding, Best Value, Best Technology, and Best PCB. Gaming performance is functionally the same on all of them, as silicon variance is the larger dictator of performance, with thermals being the next governor of performance; after all, a Pascal GPU under 60C is a higher-clocked, happier Pascal GPU, and that’ll lead framerate more than advertised clocks will.
NVIDIA’s Battlefront II Game Ready driver version 388.31 shipped this week in preparation for the game’s worldwide launch. In possibly more positive news for the vast number of redditors enraged by EA’s defense of grinding, the driver is also updated for Injustice 2 compatibility and boasts double-digit % performance increases in Destiny 2 at higher resolutions.
Battlefront 2 is the headliner for this driver release, but this chart is about all NVIDIA has to say on the subject for now:
Early reports surrounding Vega GPU packaging indicated minimally two different package processes, though later revealed a potential third. For the two primary forms of Vega GPU packaging, we’re looking at clear, obvious differences in assembly: The silicon (GPU + HBM) is either encased in an epoxy resin (“molded”) or is not encased at all (“resinless”). There is another type of resinless package that has been shown online, but we haven’t yet encountered this third type.
The initial concern indicated that packaging process could impact HBM2 contact to cooler coldplates – something for which, after working on this content, we later discovered new data – and we wanted to test that mounting pressure. Just last night, days after we finalized this content piece, we found another data point that deserves a separate article, so be sure to check back for the follow-up to this piece.
In the meantime, we’re using a chemically reactive contact paper to test various Vega GPUs and vapor chambers or coolers, then swapping coolers between those various GPUs to try and understand if and when differences emerge. Some brief thermal testing also helps us validate whether those differences, which would theoretically be spurred-on by packaging variance, are actually relevant to thermal performance. Today, we’re testing to see the mounting pressure and thermal impact from AMD’s various Vega 56 & 64 GPU packages, with a brief resurrection of the Frontier Edition.
Note: We used torque drivers for the assembly, so that process was controlled for.
We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.