GeForce GTX 1070 Arrives Guns Blazing, With 1920 Cores And 1683MHz Boost Clock On June 10
We're all lusting over NVIDIA's GeForce GTX 1080 graphics card and dreaming of having two of those beasts configured in SLI, but for the more frugal enthusiast, the GeForce GTX 1070 can't be ignored. It's the other graphics card NVIDIA introduced as part of its impressive Pascal launch, and to followup with that, we now have specifications to share.
The more affordable GeForce GTX 1070 is a monster card in its own right, one that wields a 16nm GP104 GPU with 7.2 billion transistors. It also boasts 1,920 CUDA cores with base and boost clockspeeds of 1,506MHz and 1,683MHz, respectively, along with a heaping 8GB of GDDR5 memory. The large memory buffer should help tremendously with VR gaming and running titles at a 4K resolution.
NVIDIA's GeForce GTX 1080 uses Micron's fancier (and faster) GDDR5X memory, but both cards boast the same capacity and 256-bit memory interface. With an effective memory speed of 8Gbps on the GeForce GTX 1070, the less expensive sibling is able to deliver 256GB/s of memory bandwidth.
Source: NVIDIA
How does that compare to a GeForce GTX 970 based on Maxwell? Pretty darn well—the GeForce GTX 970 has half the amount of memory, and of course there was that whole brouhaha over the last 500MB of memory being disjointed from the other 3.5GB. In terms of performance, NVIDIA claims the GeForce GTX 1070 is up to 3X faster than a GeForce GTX 970, depending on the scenario.
If sticking to the reference design, the GeForce GTX 1070 is a two-slot card measuring 10.5 inches in length and 4.376 inches in height. Connectivity consists of DisplayPort 1.4, HDMI 2.0b, and Dual Link-DVI connectors, with support for display resolutions up to 8K (7860x4320).
Outside of raw performance, the other highlight of Pascal is power efficiency. Like its bigger sibling, the GeForce GTX 1070 only requires a single 8-pin PCI Express power connector. It draws 150W and has a maximum GPU temperature of 94C—not at all shabby for a card of this caliber.
NVIDIA set the MSRP on the GeForce GTX 1070 at $379, and $449 for the Founders Edition (reference card). The card will be available to purchase starting June 10.
The more affordable GeForce GTX 1070 is a monster card in its own right, one that wields a 16nm GP104 GPU with 7.2 billion transistors. It also boasts 1,920 CUDA cores with base and boost clockspeeds of 1,506MHz and 1,683MHz, respectively, along with a heaping 8GB of GDDR5 memory. The large memory buffer should help tremendously with VR gaming and running titles at a 4K resolution.
NVIDIA's GeForce GTX 1080 uses Micron's fancier (and faster) GDDR5X memory, but both cards boast the same capacity and 256-bit memory interface. With an effective memory speed of 8Gbps on the GeForce GTX 1070, the less expensive sibling is able to deliver 256GB/s of memory bandwidth.
Source: NVIDIA
How does that compare to a GeForce GTX 970 based on Maxwell? Pretty darn well—the GeForce GTX 970 has half the amount of memory, and of course there was that whole brouhaha over the last 500MB of memory being disjointed from the other 3.5GB. In terms of performance, NVIDIA claims the GeForce GTX 1070 is up to 3X faster than a GeForce GTX 970, depending on the scenario.
If sticking to the reference design, the GeForce GTX 1070 is a two-slot card measuring 10.5 inches in length and 4.376 inches in height. Connectivity consists of DisplayPort 1.4, HDMI 2.0b, and Dual Link-DVI connectors, with support for display resolutions up to 8K (7860x4320).
Outside of raw performance, the other highlight of Pascal is power efficiency. Like its bigger sibling, the GeForce GTX 1070 only requires a single 8-pin PCI Express power connector. It draws 150W and has a maximum GPU temperature of 94C—not at all shabby for a card of this caliber.
NVIDIA set the MSRP on the GeForce GTX 1070 at $379, and $449 for the Founders Edition (reference card). The card will be available to purchase starting June 10.
