NVIDIA GeForce GTX 1070 review: A Titan X at less than half the price

Note: This article was first published on May 30.

A golden age for PC gaming

This is a glorious time to be a PC gamer. When NVIDIA first unveiled the GeForce GTX 1080 and 1070 in early May, it wasted no time in talking up the new cards as being faster than last year’s single-GPU king, the GeForce GTX Titan X. And as we saw in our review, the GeForce GTX 1080 more than lived up to those claims, a fairly impressive feat in and of itself, considering that the new Pascal-based consumer flagship costs only US$699 compared to the Titan X’s US$999.

But how about the GeForce GTX 1070? NVIDIA ended up walking back its claims about the card a little, clarifying that the latter card would offer Titan X levels of performance, instead of being outright faster. Be that as it may, it’s still quite a tantalizing prospect, given that the card costs just US$449 for the Founders Edition. Ultimately, it looked like NVIDIA had just made a GeForce GTX Titan X at less than half the price, complete with all the goodies and new features – Simultaneous Multi-Projection, GPU Boost 3.0, and Fast Sync, to just name a few – that came with the Pascal architecture and 16nm FinFET manufacturing process.

We mentioned in our GeForce GTX 1080 review that NVIDIA might just have defined a new performance class with that card, where ultra-enthusiast performance was trickling down to so-called “high-end” segment. NVIDIA looks to continue this trend with the GeForce GTX 1070 – the company’s GTX x70 cards have traditionally straddled the line between the mainstream and enthusiast markets, but the GeForce GTX 1070 looks to be doing this in a whole new way.

For the first time, gamers who buy a GTX x70 card can look forward to – at the very least – similar performance to last generation’s undisputed performance flagship. If that’s not the very definition of progress, we don’t know what is. Once again, the high-end gaming segment is showing why it continues to be bright spot in an otherwise gloomy PC market.

Meet the GeForce GTX 1070

The GeForce GTX 1070 is based on the same GP104 GPU as the 1080, but with certain functional units disabled. While the latter card boasted a total of 20 Streaming Multiprocessors (SMs), the GeForce GTX 1070 has just 15 SMs. Each SM houses 128 CUDA cores, so this brings the number of CUDA cores to 1,920, 25 percent fewer than the flagship Pascal card.

NVIDIA didn’t provide us with a block diagram for the GeForce GTX 1070, but we’ve reproduced that of the 1080’s GeForce GTX 1080 here for easier reference.

Each Graphics Processing Cluster (GPC) features 16 render output units (ROPs) and 40 texture mapping units (eight in each SM). But because the GeForce GTX 1070 actually still has 64 render output units (ROPs) – the same as its bigger brother – and 120 texture mapping units (TMUs), it looks like NVIDIA hasn’t disabled a single GPC entirely on the 1070. Instead, it appears to have selectively put certain SMs out of action in each GPC, resulting in the same number of ROPs as the GeForce GTX 1080, but 25 percent fewer TMUs.

And thanks to the new Pascal architecture, the GeForce GTX 1070 is equipped with fairly aggressive clock speeds as well, to the tune of a 1,506MHz base clock and 1,683MHz boost clock.

Unlike the GeForce GTX 1080 however, the 1070 doesn’t use GDDR5X memory, relying instead on the more traditional GDDR5 variant. Its 8GB of GDDR5 memory is clocked at 8,000MHz, but both Pascal cards share the same 256-bit memory bus. For the GeForce GTX 1070, this translates into a total available bandwidth figure of 256GB/s, a fair bit behind the 1080’s 320GB/s.

That aside, NVIDIA says that the 8Gbps data rate is the highest memory speed of any GDDR5 GPU on the market. Furthermore, because of Pascal’s improved memory compression engine that offers far more efficient color compression, the GeForce GTX 1070 can avail itself of around 20 percent more effective memory bandwidth.

More impressive is the card’s Thermal Design Power (TDP), which is the maximum amount of power the cooling system is expected to dissipate. This sits at 150 watts for the GeForce GTX 1070, a mere five watts more than GeForce GTX 970. And considering how much more powerful the 1070 is than its Maxwell-based predecessor, we begin to see the real benefits of moving to the smaller and more efficient 16nm process node. We’d also like to point out that the GeForce GTX Titan X has a TDP of 250 watts, and considering that the 1070 delivers similar levels of performance, the huge stride that Pascal has taken in terms of efficiency gains becomes truly apparent.

Here’s a table comparing the specifications of both Pascal cards.

[hwzcompare]

[products=556120,554644]

[width=200]

[caption=]

[showprices=1]

[/hwzcompare]

NVIDIA GeForce GTX 1070 Founders Edition

Save for the differences under the hood, the 1070 Founders Edition is practically a carbon copy of the GeForce GTX 1080. This means the same die cast aluminum body, a sleek metal backplate, and the same “faceted” industrial design.

There’s also the blower-style fan that will exhaust heat outside the case, a design that has actually proven itself quite effective at keeping temperatures in check. As we noted when we reviewed the GeForce GTX 1080, owners of small form factor systems may favor the Founders Edition cards because heat isn’t dumped back into their system’s chassis.

NVIDIA says the card also houses three separate copper heatpipes that channel heat away from the GPU to an aluminum heatsink. However, the metal backplate actually doesn’t provide any passive cooling benefits, and NVIDIA recommends that you remove it in dual-GPU setups for better airflow.

The GeForce GTX 1070 Founders Edition has also been designed with a low-impedance power delivery network, custom voltage regulators for more consistent voltages, and a four-phase dual-FET power supply for cleaner power delivery. Taken together, this theoretically translates into better power efficiency, reliability, and more overclocking headroom.

In the way of display connectors, you’ll have the usual selection of three DisplayPort 1.4, one HDMI 2.0b, and one dual-link DVI port. It also requires just a single 8-pin PCIe connector for power.

For even more photos of the card and some info on the new HBI SLI bridges that are recommended for multi-card configurations, check it out at GameAxis.

Test Setup

The detailed specifications of our current graphics card testbed system are as follow:-

• Intel Core i7-5960X
• ASUS X99-Pro (Intel X99 chipset) motherboard
• 2 x 4GB Corsair Vengeance LPX DDR4-2133 (Auto timings: CAS 15-15-15-36)
• Samsung SSD 840 Pro 256GB SATA 6Gbps solid state drive (OS + benchmark + games)
• Western Digital Caviar Black 1TB SATA 6Gbps hard drive (general storage)
• Windows 10 Pro 64-bit
• Intel INF 10.1.1.14

NVIDIA supplied us with beta driver version 368.19 for testing, and that’s what we used for both benchmarking and overclocking. For our comparison cards, we included NVIDIA’s top performers from the last generation, and a couple of AMD’s Fiji cards, the Radeon R9 Fury and Nano.

The full line-up of graphics cards and their driver versions are listed below:

• NVIDIA GeForce GTX 1070 (ForceWare 368.19)
• NVIDIA GeForce GTX 1080 (ForceWare 368.13)
• NVIDIA GeForce GTX Titan X (ForceWare 365.19)
• ASUS ROG Matrix GTX 980 Ti Platinum (ForceWare 365.19)
• MSI GeForce GTX 980 Gaming 4G (ForceWare 365.19)
• NVIDIA GeForce GTX 970 (ForceWare 368.22)
• Sapphire Radeon R9 Nano (Crimson Edition 16.3.1)
• ASUS Strix Radeon R9 Fury (Crimson Edition 16.3.1)

[hwzcompare]

[products=556120,554644]

[width=200]

[caption=NVIDIA GeForce GTX 1070 and 1080 compared]

[showprices=1]

[/hwzcompare]

[hwzcompare]

[products=498700,550554,479126,474324,547151,515698]

[width=200]

[caption=Test cards compared]

[showprices=1]

[/hwzcompare]

Benchmarks

We’ve updated our performance benchmarks to include a handful of the latest titles. Two benchmark games, Hitman and Ashes of the Singularity, also take advantage of DirectX 12, so we were also able to get a look at how the GeForce GTX 1070 performed using Microsoft’s latest gaming API.

Here’s the list of the benchmarks we used:

• Futuremark 3DMark (2013)
• Middle-earth: Shadow of Mordor
• Crysis 3
• Tom Clancy’s The Division
• Ashes of the Singularity
• Hitman

We used the Fire Strike Extreme test in 3DMark (2013) for our power and temperature tests.

Futuremark 3DMark (2013)

In 3DMark, the NVIDIA GeForce GTX displayed figures in line with what the GeForce GTX Titan X served up. There were minuscule – bordering on insignificant – differences across the board, so there’s really little point in calculating percentage differences. However, compared to the GeForce GTX 1080, the latter was around 15 to 20 percent faster than the 1070, depending on the tested resolution.

The GeForce GTX 1070 was also over 50 percent faster than the GeForce GTX 970 in Fire Strike Extreme and Ultra, which bodes well for gamers looking for an upgrade to a relatively affordable card that is better able to handle high resolution gaming.

The card also handily beat out the two AMD Fiji-based cards. At US$499, the Nano even costs more than the GeForce GTX 1070. Time and time again, things continue to work out in NVIDIA’s favor, as the green camp consistently shows that it can beat AMD at nearly every performance segment and price point (at least until AMD's next-gen comes out).

Middle-earth: Shadow of Mordor

But now for the actual gaming benchmarks. In Middle-earth: Shadow of Mordor, the GeForce GTX 1070 continued to serve up numbers that were almost uncannily close to that of the GeForce GTX Titan X. Truly, NVIDIA just made a Titan X at less than half the price.

We also weren’t too surprised to see that the GeForce GTX 1080 was still quite far ahead of the pack, coming in at around 28 percent faster than the 1070. With that said, the flagship Pascal card, and the factory overclocked GeForce GTX 980 Ti, were about all that managed to beat the GeForce GTX 1070. It cruised to a commanding lead over all the other cards, the GeForce GTX Titan X excepted.

Crysis 3

There was a similar pattern in Crysis 3. The GeForce GTX 1070 and Titan X were once again neck-and-neck with each other, while the 1080 maintained its lead at the head of the group. We’ll focus on the higher resolution settings here as the Pascal cards are intended for more demanding scenarios like these, and would really be wasted on regular 1080p gaming.

At Ultra settings with a resolution of 1600p, the GeForce GTX 1070 was a whopping 55 percent faster than the GeForce GTX 970. In turn, the GeForce GTX 1080 was about 22 percent faster than the 1070. That may not seem like an especially staggering lead, but it is still very significant, especially at higher resolutions with all the eye candy turned up. Even the most powerful cards continue to be taxed with settings like these, and a 20 percent lead can often make the difference between a choppy and playable gaming experience.

Tom Clancy’s The Division

The GeForce GTX 1070 fared less well in The Division, falling behind the GeForce GTX Titan X more than it did in the other benchmarks. That aside, it was still ahead of the majority of the other 'high-end' tier cards.

However, the GeForce GTX 1080 really showed that it isn’t just the single-GPU king in name only. At a resolution of 1600p and Ultra settings, it was around 33 percent faster than the 1070. As it turns out, while NVIDIA may be pitching both cards in a similar vein as each other, that is, for VR and other demanding use cases, it’s clear that you’ll still want to get the GeForce GTX 1080 if you game at quad HD resolutions and like to crank all the settings up.

Ashes of the Singularity

Thanks to Pascal’s improved asynchronous compute capabilities, the GeForce GTX 1070 was able to edge out GeForce GTX Titan X in DirectX 12 mode. However, the less powerful Pascal card didn’t benefit as much from the switch to DirectX 12 as the GeForce GTX 1080, and the most significant boost was observed at High settings and 1080p, where there was approximately a 14 percent jump in performance. Having said that. it’s worth noting that the GeForce GTX 1080 also evinced decreasing gains in performance as the settings and display resolution were ramped up.

In the DirectX 12 benchmark that ran at a 1600p resolution and Crazy settings, the GeForce GTX 1080 was a good 24 percent faster. Once again, the top Pascal card justifies why it commands a US$250 price premium over the GeForce GTX 1070.

Hitman

Our results in Hitman mirrored what we observed in Ashes of the Singularity. Although both Pascal cards displayed decreasing gains as the benchmark became more taxing, the GeForce GTX 1080 continued to hold a strong lead over the 1070. At the most demanding settings (1600p, Ultra settings) in DirectX 12 mode, the former card had a hefty 27 percent advantage.

Still, while these numbers might appear to case the GeForce GTX 1070 is a slightly unflattering light, it’s important to bear in mind that they are relative advantages. A look at the raw numbers show that the GeForce GTX 1070 is still an extremely capable card that represents a huge leap over previous generation cards like the GeForce GTX 970 and 980. What’s more, it even edges out the GeForce GTX Titan X here as well, and that, to us at least, is more than enough reason to embrace what the card delivers.

Temperature and Power Consumption

In our temperature and power consumption tests, the GeForce GTX 1070 expectedly ran cooler and consumed less power than the GeForce GTX 1080. However, while it looks like it actually runs quite a bit hotter than the previous generation NVIDIA cards and the AMD Fiji GPUs, do note that these cards all sport sport custom coolers. In fact, it is only the GeForce GTX 1070, 1080, and Titan X that utilize NVIDIA’s reference design and cooler. The NVIDIA GeForce GTX 970 that we tested is actually a Palit GeForce GTX 970 JetStream that has been downclocked to reference speeds. As a result, we expect the 1070’s peak temperature to drop quite a bit once we’ve had the chance to test out some of the cards with aftermarket coolers.

Power consumption figures were more impressive, as the GeForce GTX 1070 consumed less power than the GeForce GTX 970, which it absolutely bats out of the park in terms of performance. We’ve said it before but we’ll say it again. This is truly what happens when you move to a smaller process node.

Overclocking

The GeForce GTX 1070 also proved that there’s plenty of overclocking headroom to squeeze out of it. We were able to boost the base clock by 210MHz over the stock speed, and the memory clock by 110MHz. This ultimately meant that the card was able to hit a boost speed of around 2,063MHz while we were running 3DMark, although it was running just under 2,000MHz most of the time and only hit the top speed in the most demanding scenes.

This translated into around a five to 10 percent gain in 3DMark (the GeForce GTX 1080 had similar proportional gains). The biggest benefit was seen in Fire Strike Ultra, so we’d extrapolate and say that you might possibly see the most benefit from overclocking when gaming at 4K resolutions. 

Is this what mainstream performance looks like now?

For all the beastly performance figures put out by the GeForce GTX 1080, the 1070 might actually be the more impressive card if you discount the raw performance numbers. At US$449 for the Founders Edition, it serves up performance on par with the GeForce GTX Titan X, and even manages to edge it out in certain DirectX 12 games. Remember, the Titan X is a US$999 card, and we’re still beginning to wrap our heads around the fact that NVIDIA has brought previously ultra-enthusiast levels of performance to a GeForce GTX x70 card in just one generation.

Like its predecessors, the GeForce GTX 1070 sits on the line between the mainstream and performance segments, and all of a sudden, one generation’s enthusiast card is the next generation’s mainstream GPU. It seems almost an injustice to characterize the GeForce GTX 1070 as a mainstream card, but there’s really no avoiding the fact that at US$449 (for a Founders Edition card), the card is within the reach of mainstream consumers who want more performance to keep up with the latest games and VR applications.

The GeForce GTX 970 launched at US$349 back in late 2014, so you might also say that there are countervailing forces – in the form increased prices and the Founders Edition premium – that are working against the democratization of bleeding edge performance. As we noted in our review of the GeForce GTX 1080, NVIDIA wants to turn the cards it produces in-house into their own line of profit, and AMD’s lack of a proper response has allowed it to do so with impunity. After all, we don’t live in a perfect world, and we’d be kidding ourselves if we expected businesses to automatically put consumers’ interests first without some stiff competition.

But we don’t want to take anything away from what NVIDIA has achieved with the GeForce GTX 1070. The Founders Edition price premium aside, the fact is still that a whole lot of gamers are now going to be able to enjoy a level of performance that used to be prohibitively expensive. Maybe the card isn’t as cheap as it could have been had AMD been more competitive, but progress is still progress, and enthusiast levels of performance are now far more accessible than before. Let's not forget that NVIDIA also announced that non-reference (Founders Edition) cards will be available from US$379, so there's definitely going to be more affordable options later this year.

There’s also reason to hope that as this sort of performance becomes the norm at these price points, we might see prices drop in succeeding generations. And if you’re currently rocking a GeForce GTX 980 or 970, running out to pick up a GeForce GTX 1070 when it launches on 10 June will net you a very significant performance upgrade. Truly, this is a great time to be a PC gamer.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.