Nvidia is preparing to release another powerful graphics card, the GeForce GTX 1070 Ti. According to numerous sources, this GPU will be formally announced on Oct. 26, 2017.
In connection with the upcoming release, said graphics card recently surfaced on benchmark website Ashes of the Singularity (AOTS). This upcoming NVIDIA product was given a score of 6,200 which is pretty much within striking distance of the latest AMD graphics cards, and, just recently, two more benchmark test results surfaced.
Benchmark results
According to tech website WCCFTech.com, Time Spy, and Fire Strike Extreme, benchmarking websites got a hold of the upcoming Ti version of the GeForce GTX 1070 graphics card.
Compared to the leading GPUs from both sides of the aisle, the Ti version has shown consistent figures on each of the two benchmark websites.
Based on the Time Spy benchmark, the GTX 1070 Ti graphics card scored 6,777 which is higher compared to an overclocked non-Ti model (clock rate at 1998MHz) which only scored around 6,000. This score sits somewhere in the vicinity of AMD Radeon RX Vega 56 which many tech analysts believe is the true target of NVIDIA.
The Fire Strike Extreme benchmark produced a lower score of 9,449 points for the GTX 1070 Ti – exactly 500 points lower than the RX Vega 56 which scored 9,949 points. However, once this NVIDIA graphics card is launched, there’s a good chance that the score will be changed as the testing was done using only a generic driver installed – not the correct driver.
Rumored GPU specifications
The leaked specifications of GTX 1070 Ti suggest that it sits somewhere between the non-Ti model and Geforce Gtx 1080. While most hardware specifications are patterned with the non-Ti model, a few specifications match that of the higher GTX 1080.
In particular, the two graphics cards share the same 1607MHz base clock and 180W TDP.
The GeForce GTX 1070 Ti also has a 2432 Stream Processor (CUDA Cores) – only 128 less than GTX 1080. The upcoming Ti model also boasts an 8.1TFLOPs (FP32 Compute) compared to the non-Ti model with only 6.5 TFLOPs.
The remaining hardware specifications of both GTX 1070 GPUs are pretty much the same: 1683MHz (boost clock), 8 GB GDDR5 VRAM, 8Gbps memory speed, and bandwidth.
Meanwhile, the three NVIDIA graphics cards share the same build specifications: GP104, 16nm FinFET process node, 314mm2 die size, 7.2 billion transistors, 256-bit bus interface, and display outputs (3x Display Port 1.4, 1x HDMI 2.0b, and 1x DVI).
