Nvidia has been the company on the lips of all enthusiasts lately. After the company recently unveiled its new graphics architecture “Turing”, there has been a lot of talk about the new graphics cards from the graphics giant that use it. Some information has been revealed in advance and now Nvidia is officially unveiling the models.
As expected, Nvidia is moving away from the name scheme with the Geforce GTX and the new models will instead be part of the Geforce RTX 2000 series. The letters “RTX” refer to the company’s technology of the same name that will enable hardware-accelerated ray tracing in real time for more realistic graphics in games, among other things.
Specifications: Nvidia Geforce RTX 2080 Ti, RTX 2080 and RTX 2070
RTX 2080 Ti | GTX 1080 Ti | RTX 2080 | RTX 2070 | |
---|---|---|---|---|
Technical | 12 nm TSMC | 16 nm TSMC | 12 nm TSMC | 12 nm TSMC |
Circuit | TU102-300A | GP102-350 | TU104-400A | TU104 |
Circuit surface | 754 mm2 | 471 mm2 | ? | ? |
Transistors | 18.6 billion | 12 billion | ? | ? |
Architecture | Turing | Pascal | Turing | Turing |
CUDA cores | 4 352 st. | 3 584 st. | 2 944 st. | 2 304 st. |
Texture units | 272 st. | 224 st. | 184 st. | 144 st. |
Raster units | 88 st. | 88 st. | 64 st. | 64 st. |
Clock frequency | 1 350 MHz | 1 480 MHz | 1 515 MHz | 1 410 MHz |
GPU Boost | 1 545 MHz | 1 582 MHz | 1 710 MHz | 1 620 MHz |
Computational power | 13 448 GFLOPS | 11 334 GFLOPS | 10 068 GFLOPS | 7 465 GFLOPS |
Memory bus | 352-bit | 352-bit | 256-bit | 256-bit |
Memory amount | 11 GB GDDR6 | 11 GB GDDR5X | 8 GB GDDR6 | 8 GB GDDR6 |
Memory frequency | 14 000 MHz | 11 000 MHz | 14 000 MHz | 14 000 MHz |
Memory bandwidth | 616 GB/s | 484 GB/s | 448 GB/s | 448 GB/s |
Power supply | 8+8-pin | 8+6-pin | 8+6-pin | 8-pin |
Outputs | 1 st. HDMI 2.0b | 1 st. DL-DVI | 1 st. HDMI 2.0b | 1 st. HDMI 2.0b |
TDP | 250 W | 250 W | 215 W | 175 W |
Rec. Award | SEK 13,999 | SEK 7,500 | SEK 9,199 | SEK 6,899 |
First out in the new Geforce RTX 2000 series are RTX 2080, RTX 2080 Ti and RTX 2070 which replace the previous Geforce GTX 1080, GTX 1080 Ti and GTX 1070. In terms of specifications, it is about 4,352 CUDA cores and a clock frequency of 1,350 MHz and 11 GB GDDR6 memory for Geforce RTX 2080 Ti.
For Geforce RTX 2080, 2,944 CUDA cores, a base frequency of 1,515 MHz and 8 GB of GDDR6 memory apply. However, it gets a slightly higher turbo frequency of 1,710 MHz, compared to 1,545 MHz for RTX 2080 Ti.
Compared to previous models, a higher memory bandwidth of 616 GB / s and 448 GB / s is also offered for RTX 2080 Ti and RTX 2080, respectively. Otherwise, it is a TDP value of 250 W for RTX 2080 Ti and 215 W for RTX 2080, and the power supply is managed by two 8-pin PCI Express for the top model and 6 + 8-pin PCI Express for the smaller sibling.
Like the big sibling Geforce RTX 2080, scaled-down RTX 2070 comes with 8 GB of GDDR6 memory accompanied by a 256-bit memory bus. The number of CUDA cores amounts to 2,304, and the clock frequencies in base and turbo mode are specified to 1,410 and 1,620 MHz.
It is worth mentioning, however, that the company’s “Founders Edition” graphics card is delivered with a turbo frequency of 1,710 MHz and a specified TDP value of 185 watts. Power is supplied through a single 8-pin ATX connection and the cooler solution is again equipped with two fans.
Nvidia has also chosen to make some updates when it comes to image connections compared to the previous generation. Gone is DVI, which has instead been replaced by a USB Type-C according to the cable standard Virtuallink. In addition to this, there are three Displayports and an HDMI available.
To connect several graphics cards, Nvidia has long used an SLI connection, which now strokes the foot for the two more powerful cards. This connection has instead been replaced with the high-speed interface NVLink, which was previously found in the company’s company-oriented products. Whether the Geforce RTX 2070 can be connected to other graphics cards at all is unknown.
During the presentation, Nvidia showed a Star Wars demo, which shows the visual effects of radiation tracking. The Geforce RTX 2080 Ti can render just over 22 frames per second, which is marginally higher than for the company’s pre-built system DGX-2, equipped with four Volta-based graphics cards.
The performance life of the Geforce GTX 1080 Ti is all the greater, as the Pascal card can only present a little more than three frames per second. The large increase in the specific scenario is attributed to the RT cores that are new to the Turing architecture. These can, among other things, increase performance by working with AI to find and correct errors during the rendering process.
Nvidia Geforce RTX 2080 Ti, RTX 2080 and RTX 2070 are now available for pre-booking via Nvidia’s website, and the first two will be launched sharply on 20 September. Geforce RTX 2070 is expected to launch later in October. The prices for the graphics cards in the Swedish online store are SEK 13,999, 9,199 and SEK 6,899, including VAT.