Custom GeForce RTX 2070 cards will have 2304 CUDA cores with base and boost GPU clocks coming in at 1410MHz and 160MHz respectively. Individual OEMs, however, will be able to set their own speeds, so we will likely see wide variation between companies like MSI, ASUS, Gigabyte and EVGA. Stepping up to the Founders Edition will see those base and boost clocks rise to 1410MHz and 1710MHz respectively. 8GB of GDDR6 memory will be onboard GeForce RTX 2070 cards using a 256-bit memory interface for effective bandwidth of 448GB/sec.
NVIDIA says that the GeForce RTX 2070 can perform 45 trillion RTX operations per second (compared to 60 trillion and 75 trillion respectively for the GeForce RTX 2080 and GeForce RTX 2080 Ti). The graphics cards will be fully capable of delivering stunning visual using its real-time ray tracing engine, but you’ll be limited to resolutions of 1080p and below to achieve acceptable frame rates.
Beautiful any way you look at it.
The GeForce RTX 2070 will be available on October 17th. #GraphicsReinvented
Shop starting at $499 ($599 Founders Edition) → https://t.co/ammFWibyFy pic.twitter.com/IsScoXm5rZ
— NVIDIA GeForce (@NVIDIAGeForce) September 25, 2018
In addition, the GeForce RTX 2070 won’t support the NVLink SLI connector like its more powerful siblings, likely in an effort to reduce board costs and provide some additional product differentiation.
We were highly impressed with the performance put forth by both the GeForce RTX 2080 and GeForce RTX 2080 Ti in our extensive review, so stay tuned to see what our thoughts are on this pared down entry.