
Mark Walton
Specifications at a glance: Nvidia GeForce GTX 1070 | |
---|---|
CUDA CORE | 1920 |
TEXTURE UNITS | 120 |
ROPS | 64 |
CORE CLOCK | 1506MHz |
BOOST CLOCK | 1683MHz |
MEMORY BUS WIDTH | 256-bit |
MEMORY BANDWIDTH | 256 GB/s |
MEMORY SIZE | 8GB DDR5 |
Outputs | 3x DisplayPort 1.4, 1x HDMI 2.0b supporting 4K60 10/12b HEVC Decode, 1x dual-link DVI |
Publication date | Founders Edition: June 10 |
PRICE | Founders Edition (as reviewed): £399, €499, $450; Partner cards priced at £329, €419, $379 |
In January, Nvidia’s GTX 970 became the most popular graphics card on Steam. This was a remarkable achievement considering the most popular chip at the time, the HD Graphics 4000, which was not designed for gaming at all and was integrated into Intel CPUs. Today, the GTX 970 still has a five percent share of the Steam audience. Its successor, the GTX 1070 – the second graphics card based on Nvidia’s latest Pascal architecture after the powerful but pricey GTX 1080 – has some big shoes to fill.
[Updated, October 18: Nvidia has just unveiled the GTX 1050 and GTX 1050 Ti, priced very aggressively and released on October 25.]
And it fills them – for the most part. As Nvidia promised, the GTX 1070 is indeed faster than both the GTX 980 Ti and the Titan X, and by some margin: as much as 12 percent in some tests. Just a few months ago, GTX 980ti cards cost over £500 / $650, but the GTX 1070 costs just £399 / $449 at the top end.
In its Founders Edition form (Nvidia’s new reference card nomenclature), the GTX 1070 is also cool and quiet. The smaller, more efficient TSMC 16nm FinFET manufacturing process allows Nvidia to boost performance to Titan-beating levels while keeping the TDP down to a reasonable 150W.
From a price-performance standpoint, the GTX 1070 is undoubtedly a better value than the GTX 1080. It offers about 80 percent of the performance for only 60 percent of the price.
But there are some problems. At £329/$379 for the OEM models (Asus, MSI, Gigabyte, etc), the GTX 1070 is still a good £60/$50 more expensive than the GTX 970 that launched. That’s not to say that those partner cards aren’t really on the market yet. If you want a GTX 1070 when it releases on June 10 (today!) you’ll need to buy a Founders Edition, which costs a hefty £399 / $449.
Still, even at that higher price point, there’s nothing in the same ballpark as the GTX 1070. Its closest competitor, AMD’s similarly priced Fury Nano, comes with just 4GB of memory and is easily beaten by the GTX 1070, while the R9 Fury and R9 Fury X are both significantly more expensive. Nvidia may have moved up the mythical sweet spot for the graphics card a bit, but for now, at least, there’s simply no better option for the gamer who wants high-end performance without the silly expensive price tag.
GTX 1070: Founders Edition 2.0
Like the GTX 1080, the GTX 1070 will initially be released as a Founders Edition, which is essentially a reference card with a higher price tag and available for a year. Yes, early adopters are being stoned again, although judging by the sheer number of GTX 1080 partner cards unveiled at this year’s Computex in Taiwan, GTX 1070 partner cards with custom coolers and factory overclocking should appear shortly after release.
Those who do opt for the Founders Edition get the same beautiful, versatile aluminum case as the GTX 1080, along with a fan-like design that exhausts hot air from a PC case. However, there are some concessions to be made under the shroud, replacing the vapor chamber solution with a simpler aluminum heat sink with three embedded copper heat pipes. The GTX 1080’s acclaimed five-phase power supply also gets a downgrade to a four-phase dual-FET design.


None of these changes have a major effect on performance, especially given the GTX 1070’s lower TDP of 150W. The overclocking power (depending on your chip’s binning) is still impressive. Such a low TDP means there’s just one 8-pin PCIe power connector on top to power the GTX 1070, while there’s three DisplayPort 1.4 ports on the back, one HDMI 2.0b port supporting 4K60 10/12b HEVC decoding, and one dual-link DVI port for those still using older monitors.
Under the hood is the same GP104 GPU as the GTX 1080, which is built on a tiny 314mm² 16nm TSMC chip and the new Pascal architecture. I won’t go into the ins and outs of the architecture here – check out our GTX 1080 review for that – but essentially Pascal is a leaner version of Maxwell, with a focus on the FP32 performance that video games rely on. The main difference between the GTX 1080 and 1070 is that one of the GP104’s graphics processing clusters (GPCs) is disabled, losing five streaming multiprocessors (SMs).
That leaves the GTX 1070 with 15 SMs, 1920 CUDA cores (vs. 2560), and 120 texture units (vs. 160), but the same number of ROPs, which should avoid the GTX 970’s memory snafus (think the 3.5 GB kerfuffle?). It also doesn’t feature the GTX 1080’s fancy GDDR5X memory, instead using standard GDDR5 on a 256-bit bus for 256 GB/s of memory bandwidth. The clock speeds have also been reduced to a 1506 MHz core clock and 1683 MHz boost.
Specifications at a glance: | GTX 1080 | GTX 1070 | GTX TitanX | GTX 980 Ti | GTX 980 | GTX 970 | GTX 780 Ti |
---|---|---|---|---|---|---|---|
CUDA cores | 2560 | 1920 | 3072 | 2816 | 2048 | 1664 | 2880 |
Texture Units | 160 | 120 | 192 | 176 | 128 | 104 | 240 |
ROPs | 64 | 64 | 96 | 96 | 64 | 56 | 48 |
core clock | 1607MHz | 1506MHz | 1000MHz | 1000MHz | 1126MHz | 1050MHz | 875MHz |
Raise the clock | 1733MHz | 1683MHz | 1050MHz | 1050MHz | 1216MHz | 1178MHz | 928MHz |
Memory bus width | 256-bit | 256-bit | 384-bit | 384-bit | 256-bit | 256-bit | 384-bit |
Memory speed | 10GHz | 8GHz | 7GHz | 7GHz | 7GHz | 7GHz | 7GHz |
Memory bandwidth | 320GB/s | 256 GB/s | 336 GB/s | 336 GB/s | 224 GB/sec | 196 GB/s | 336 GB/s |
Memory size | 8GB GDDR5X | 8GB DDR5 | 12GB DDR5 | 6GB GDDR5 | 4GB GDDR5 | 4GB GDDR5 | 3GB GDDR5 |
TDP | 180W | 150W | 250W | 250W | 165W | 145W | 250W |
At the top, there are two SLI interfaces suitable for Nvidia’s new high-bandwidth bridges. Those bridges operate at a higher speed of 650MHz (versus 400MHz) by using the second SLI connector traditionally reserved for three- or four-way SLI configurations. Older bridges also work, but at a lower speed.
Like the GTX 1080, Nvidia only officially supports two-way SLI with GTX 1070. Previously, Nvidia said more cards can be used by downloading an unlock key from its website. It’s now dropped that requirement, but in the process it turned out that three- or four-way SLI setups aren’t supported in games at all and will only work in select benchmarking applications like 3DMark. Two-way SLI has always made the most sense in terms of scalability – and since GTX 1070 has only just been released, no one will have bought more than two cards yet – but it’s a bad show from Nvidia not to be clear about its SLI plans from the start.
Finally, there’s GPU Boost 3.0, Fast Sync, HDR, VR Works Audio, Ansel, and preemption (an alternative approach to asynchronous computation), all of which are excellent additions, but not unique to the GTX 1070. For more on this and the Pascal architecture, check out the GTX 1080 review.