Updated: Nvidia GeForce GTX 1070 8GB Pascal Review

How We Tested GeForce GTX 1070

All of the testing we did for our GeForce GTX 1080 review is just as relevant now as it was two weeks ago. Thus, we’re using the same hardware, software and benchmarking methodology, with one notable exception. Rather than testing The Witcher 3 with v-sync off and its in-game limiter set to 60 frames per second, we re-tested all cards with the artificial ceiling lifted. Many gamers stabilize The Witcher’s frame rate using a 60 FPS cap; however, enough readers asked to see how high top-end cards could go at 2560x1440 that we wanted to show those numbers.

In the same vein, we had a lot of requests for Otoy’s OctaneBench. Just so you know, we tried to include results in today’s piece, but were stymied by a lack of support for GeForce GTX 1080 and 1070 in the demo version of OctaneRender. Nvidia acknowledges that it's sometimes necessary to work with developers before an application will recognize unreleased hardware, and in this case, Otoy will likely need to update its software to detect Pascal-based GPUs.

From Nvidia GeForce GTX 1080 Pascal Review:

“Instead of our Haswell-E-based Core i7-5930K at 3.5GHz, we’re using a Skylake-based Core i7-6700K at 4GHz, giving us two generations worth of IPC improvements and an extra 500MHz base clock rate to alleviate host processing bottlenecks wherever they may surface. Of course, the CPU’s LGA 1151 interface also calls for a different motherboard—we tapped MSI’s Z170A Gaming M7 for all of our game benchmarks, and dropped in G.Skill’s F4-3000C15Q-16GRR memory kit composed of four 4GB modules at DDR4-3000. Crucial’s MX200 SSD remains, as does the Noctua NH-12S cooler and be quiet! Dark Power Pro 10 850W power supply.

Gone is Windows 8.1, though. Prior to benchmarking, we installed a clean version of Windows 10 Professional and a new suite of games representing popular AAA titles, some DirectX 12-specific selections and a mix of genres.

Because the GeForce GTX 1080 is a flagship, its competition is limited to the top end from AMD and Nvidia. We chose a GeForce GTX Titan X, 980 Ti, and 980 to go against it, along with AMD’s Radeon R9 Fury X, Fury, and 390X. All of the cards are reference except for Sapphire’s Nitro Radeon R9 Fury and MSI’s R9 390X Gaming 8G.”

Naturally, to this line-up we tack on Nvidia’s new GeForce GTX 1070.

Drivers And Benchmarks

Again, from our GeForce GTX 1080 review:

“The Maxwell-based cards employ Nvidia’s newest driver at the time of testing, GeForce Game Ready Driver 365.10. For the GeForce GTX 1080, we had to use the company’s press driver, 368.13. All three cards based on AMD GPUs use Radeon Software Crimson Edition 16.5.2 Hotfix, released on May 10.”

Driver version 368.13 does not support the GeForce GTX 1070. Rather, Nvidia pushed out a 368.19 update ahead of today’s story.

“Our benchmark suite also reflects some changes. A few old favorites remain—notably Battlefield 4, Grand Theft Auto V and The Witcher 3. But we’re also adding Hitman, Project CARS, Rise of the Tomb Raider, The Division and Ashes of the Singularity.

The Ashes charts represent DirectX 12 performance using the game’s built-in benchmark/logging tool. Hitman and Tomb Raider are presented using DirectX 11. However, we have results from DirectX 12 using those games as well, which we’ll mention in the analysis (spoiler: in most cases, performance drops with DirectX 12). Everything else is DirectX-based, recorded with Fraps.”

Swipe to scroll horizontally
Ashes of the Singularity: DirectX 12, Extreme quality preset, built-in benchmark
Battlefield 4
DirectX 11, Ultra quality preset, custom Tom’s Hardware benchmark (Tashgar jeep ride), 100-second Fraps recording
Grand Theft Auto V
DirectX 11, Very High quality settings, 4x MSAA, built-in benchmark (test five), 110-second Fraps recording
Hitman (2016)
DirectX 11, Ultra level of detail, FXAA, High texture quality, built-in benchmark, 100-second Fraps recording
Metro Last Light Redux
DirectX 11, Very High detail preset, SSAA off, 16x AF, Normal Motion Blur, Normal Tessellation, built-in benchmark, 145-second Fraps recording
Project CARS
DirectX 11, Ultra quality settings, High anti-aliasing, High texture resolution, Nürburgring Sprint, 100-second Fraps recording
Rise of the Tomb Raider
DirectX 11, Custom quality preset, Very High quality settings, built-in benchmark, 80-second Fraps recording
Tom Clancy's The Division
DirectX 11, Custom quality preset, Ultra quality settings, Supersampling temporal AA, built-in benchmark, 90-second Fraps recording
The Witcher 3: Wild Hunt
DirectX 11, Highest quality settings, HairWorks disabled, custom Tom’s Hardware benchmark, 100-second Fraps recording

MORE: Best Graphics Cards

MORE: All Graphics Content

Chris Angelini
Chris Angelini is an Editor Emeritus at Tom's Hardware US. He edits hardware reviews and covers high-profile CPU and GPU launches.
  • adamovera
    Archived comments are found here: http://www.tomshardware.com/forum/id-3073584/nvidia-geforce-gtx-1070-8gb-pascal-performance-review.html
  • gromann
    Fury X is going for $399-449 as of yesterday on Newegg.
  • nitrium
    So given the simultaneously lower price and higher performance of the partner boards, only an actual idiot would buy the "Founder's Edition" GTX 1070?
  • George Phillips
    I feel that I should regret getting MSI 1070 FE. MSI's custom designs perform superior then FE cards in every way. Very impressive. Asus and Gigabyte's custom designs must also do better than FE cards.
  • Krushe
    When you're talking about the heat on FE cards. I think the default fan speed is 45-50% at 83c. Make it 80% and the card never reaches 70c even with boost clock up to 1900+. What speeds are the MSI fans running at during your temp measurements?
  • DookieDraws
    Edit: The article has been updated, so I deleted my original comment about the MSI GPU.
  • Tony Casagrande
    "This means that the lowest possible GPU Boost clock rate step gets eliminated from the bottom of the BIOS’ table. So, if you want an additional space at the top, you need to make room for it by getting rid of the very bottom one."

    If it were me, I would have removed a low to middle clock rate instead of the very lowest to get both the low idle power consumption and the OC speed.
  • straatkat
    Still annoyed that they posted this review without including the 970 as a baseline.
  • neblogai
    Regarding the possible audible noise because of power spikes on PEG: it is not really about cheap MB, but about using analog audio out of MB, and not anything digital, right?
    Also, about overclocking: I think reviews of all these new generation nVidia and AMD cards should include average clock that cards operated when doing all game benchmarks. Official boost clock numbers are a bit useless, because AMD cards run games at below boost clocks, and average for nVidia GTX1070 is above boost clocks. Having just official boost clock numbers make it difficult to evaluate overclocking potential and make real gains look much bigger or smaller than expected.
  • Calculatron
    Am I the only one that noticed that the Founder's Edition cards managed to pull over 75 watts from the motherboard PCIe slot and that no one went bonkers over it?