Computer desktop rendering software can now make use of the GPU, in addition to the CPU, to increase the rendering speed. The AMD Threadripper 5995 WX CPU is a 3D graphics designer’s dream with 64 cores and 128 threads, but it also comes with a big price tag that is beyond the reach of enthusiasts. The ability to utilise the GPU such as the NVIDIA Cuda Cores is a real bonus.
In this article, we keep track of the key specifications of the NVIDIA graphics card that we feel could help us decide when and what to buy, before going into the detailed tests done by other websites.
Series 40, first released October 2022. Architecture: Ada Lovelace
Cuda Capability: 8.9
Ray Tracing Cores: Gen 3
Tensor Cores: Gen 4
Display Resolution:
- Up to 4k 12-bit HDR at 240Hz with DP 1.4a + DSC.
- Up to 8k 12-bit HDR at 60Hz with DP 1.4a + DSC or HDMI 2.1 + DSC.
- With dual DP 1.4a + DSC, up to 8K HDR at 120Hz
Series 40 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. | Connectors |
RTX 4090 | 16384 | 2.52 GHz | 2.23 GHz | 24GB GDDR6X | 384-bit | 450W | 850W | 90C | HDMI, 3xDP |
RTX 4080 | 9728 | 2.51 GHz | 2.21 GHz | 24GB GDDR6X | 256-bit | 320W | 750W | 90C | HDMI, 3xDP |
RTX 4070 TI | 7680 | 2.61 GHz | 2.31 GHz | 12GB GDDR6X | 192-bit | 285W | 700W | 90C | HDMI, 3xDP |
RTX 4070 | 5888 | 2.48 GHz | 1.92 GHz | 12GB GDDR6X | 192-bit | 200W | 650W | 90C | HDMI, 3xDP |
RTX 4060 TI | 4352 | 2.54 GHz | 2.31 GHz | 16/8GB GDDR6 | 128-bit | 165/160W | 550W | 90C | HDMI, 3xDP |
RTX 4060 | 3072 | 2.46 GHz | 1.83 | 8GB GDDR6 | 128-bit | 115W | 550W | 90C | HDMI, 3xDP |
In general, all the models in RTX Series 40 have a higher boost clock, base clock and memory than RTX Series 30.
RTX 4070 TI has 7680 Cuda cores compared to RTX 3070 TI 6144 Cuda cores, but it loses out to the RTX 3080 or higher. Even though it has slightly lower Cuda core count, users may experience better speed due to the significantly higher boost and base clock of the Series 40.
Series 30, released May 2022. Architecture: Ampere
Cuda Capability: 8.6
Ray Tracing Cores: Gen 2
Tensor Cores: Gen 3
Display Resolution:
- Up to 4k 12-bit HDR at 240Hz with DP 1.4a + DSC.
- Up to 8k 12-bit HDR at 60Hz with DP 1.4a + DSC or HDMI 2.1 + DSC.
- With dual DP 1.4a + DSC, up to 8K HDR at 120Hz
Series 30 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. | Connectors |
RTX 3090 TI | 10752 | 1.86 GHz | 1.67 GHz | 24GB GDDR6X | 384-bit | 450W | 850W | 92C | HDMI, 3xDP |
RTX 3090 | 10496 | 1.70 GHz | 1.4 GHz | 24GB GDDR6X | 384-bit | 350W | 750W | 93C | HDMI, 3xDP |
RTX 3080 TI | 10240 | 1.67 GHz | 1.37 GHz | 12GB GDDR6X | 384-bit | 350W | 750W | 93C | HDMI, 3xDP |
RTX 3080 | 8960 | 1.71 GHz | 1.26 GHz | 12GB GDDR6X | 384-bit | 350W | 750W | 93C | HDMI, 3xDP |
RTX 3080 | 8704 | 1.71 GHz | 1.44 GHz | 10GB GDDR6X | 320-bit | 320W | 750W | 93C | HDMI, 3xDP |
RTX 3070 TI | 6144 | 1.77 GHz | 1.58 GHz | 8GB GDDR6X | 256-bit | 290W | 750W | 93C | HDMI, 3xDP |
RTX 3070 | 5888 | 1.73 GHz | 1.50 GHz | 8GB GDDR6 | 256-bit | 220W | 650W | 93C | HDMI, 3xDP |
RTX 3060TI | 4864 | 1.67 GHz | 1.41 GHz | 8GB GDDR6X | 256-bit | 200W | 600W | 93C | HDMI, 3xDP |
RTX 3060TI | 4864 | 1.67 GHz | 1.41 GHz | 8GB GDDR6 | 256-bit | 200W | 600W | 93C | HDMI, 3xDP |
RTX 3060 | 3584 | 1.78 GHz | 1.32 GHz | 12GB GDDR6 | 192-bit | 170W | 550W | 93C | HDMI, 3xDP |
RTX 3060 | 3584 | 1.78 GHz | 1.32 GHz | 8GB GDDR6 | 128-bit | 170W | 550W | 93C | HDMI, 3xDP |
RTX 3050 | 2560 | 1.78 GHz | 1.55 GHz | 8GB GDDR6 | 128-bit | 130W | 550W | 93C | HDMI, 2xDP |
RTX 3050 | 2304 | 1.76 GHz | 1.51 GHz | 8GB GDDR6 | 128-bit | 130W | 550W | 93C | HDMI, 2xDP |
Series 20, released Feb 2019. Architecture: Turing
Cuda Capability: 7.5
Ray Tracing Cores: Gen 1
Tensor Cores: Gen 2
Display Resolution:
- Up to 4k 12-bit HDR at 240Hz with DP 1.4a + DSC.
- Up to 8k 12-bit HDR at 60Hz with DP 1.4a + DSC or HDMI 2.1 + DSC.
- With dual DP 1.4a + DSC, up to 8K HDR at 120Hz
Series 20 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. | Connectors |
RTX 2080 TI | 4352 | 1.64 GHz | 1.35 GHz | 11GB GDDR6 | 352- bit | 260W | 650W | 89C | HDMI,DP,USB-C |
RTX 2080 Super | 3072 | 1.82 GHz | 1.65 GHz | 8GB GDDR6 | 256-bit | 250W | 650W | 89C | HDMI,DP |
RTX 2080 | 2944 | 1.80 GHz | 1.52 GHz | 8GB GDDR6 | 256-bit | 225W | 650W | 88C | HDMI,DP,USB-C |
RTX 2070 Super | 2560 | 1.77 GHz | 1.61 GHz | 8GB GDDR6 | 256-bit | 215W | 650W | 88C | HDMI,DP |
RTX 2070 | 2304 | 1.71 GHz | 1.41 GHz | 8GB GDDR6 | 256-bit | 185W | 550W | 89C | HDMI,DP,USB-C,DVI-DL |
RTX 2060 Super | 2176 | 1.65 GHz | 1.47 GHz | 8GB GDDR6 | 256-bit | 175W | 550W | 89C | HDMI,DP,DVI-DL |
RTX 2060 | 2176 | 1.65 GHz | 1.47 GHz | 12GB GDDR6 | 192-bit | 185W | 550W | 88C | HDMI,DP,DVI-DL |
RTX 2060 | 1920 | 1.68 GHz | 1.37 GHz | 6GB GDDR6 | 192-bit | 160W | 500W | 88C | HDMI,DP,DVI-DL |
Series 16, released Sep 2018. Architecture: Turing
Cuda Capability: 7.5
Ray Tracing Cores: None
Tensor Cores: None
Display Resolution:
- Up to 4k 12-bit HDR at 240Hz with DP 1.4a + DSC.
- Up to 8k 12-bit HDR at 60Hz with DP 1.4a + DSC or HDMI 2.1 + DSC.
- With dual DP 1.4a + DSC, up to 8K HDR at 120Hz
Series 16 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. | Connectors |
GTX 1660 TI | 1536 | 1.77 Ghz | 1.50 Ghz | 6GB GDDR6 | 192-bit | 120W | 450W | 95C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
GTX 1660 Super | 1408 | 1.79 Ghz | 1.53 Ghz | 6GB GDDR6 | 192-bit | 125W | 450W | 93C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
GTX 1660 | 1408 | 1.79 Ghz | 1.53 Ghz | 6GB GDDR5 | 192-bit | 120W | 450W | 95C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
GTX 1650 Super | 1280 | 1.73 Ghz | 1.53 Ghz | 4GB GDDR6 | 128-bit | 100W | 350W | 90C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
GTX 1650 (G5) | 896 | 1.67 Ghz | 1.49 Ghz | 4GB GDDR5 | 128-bit | 75W | 300W | 92C | HDMI 2.0b, DL-DVI-D |
GTX 1650 (G6) | 896 | 1.59 Ghz | 1.41 Ghz | 4GB GDDR6 | 128-bit | 75W | 300W | 90-94C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
GTX 1630 | 512 | 1.79 GHz | 1.74 Ghz | 4GB GDDR6 | 64-bit | 75W | 300W | 90-94C | DP 1.4a, HDMI 2.0b, DL-DVI-D |
Series 10, released Apr 2016. Architecture: Pascal
Cuda Capability: 6.1
Ray Tracing Cores: Gen 1
Tensor Cores: Gen 2
Bus Support: PCIe 3.0
Display Resolution:
- 7680×4320 at 60 Hz RGB 8-bit with dual DisplayPort connectors, or
- 7680×4320 at 60 Hz YUV420 8-bit with one DisplayPort 1.3 connector
Series 10 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. | Connectors |
GTX 1080 TI | 3584 | 1.582 GHz | 1.48 GHz | 11 GB GDDR5X | 352-bit | 250W | 600W | 91C | DP 1.43, HDMI 2.0b |
GTX 1080 | 2560 | 1.733 GHz | 1.607 GHz | 8 GB GDDR5X | 256-bit | 180W | 500W | 94C | DP 1.43, HDMI 2.0b, DL-DVI |
GTX 1070 TI | 2432 | 1.683 GHz | 1.607 GHz | 8 GB GDDR5 | 256-bit | 180W | 500W | 94C | DP 1.43, HDMI 2.0b, DL-DVI |
GTX 1070 | 1920 | 1.683 GHz | 1.506 GHz | 8 GB GDDR5 | 256-bit | 150W | 500W | 94C | DP 1.43, HDMI 2.0b, DL-DVI |
GTX 1060 (6GB) | 1280 | 1.708 GHz | 1.506 GHz | 6 GB GDDR5 | 192-bit | 120W | 400W | 94C | DP 1.43, HDMI 2.0b, Dual Link-DVI |
GTX 1060 (3GB) | 1152 | 1.708 GHz | 1.506 GHz | 3 GB GDDR5 | 192-bit | 120W | 400W | 94C | DP 1.43, HDMI 2.0b, Dual Link-DVI |
GTX 1050 TI | 768 | 1.392 GHz | 1.290 GHz | 4 GB GDDR5 | 128-bit | 75W | 300W | 97C | DP 1.43, HDMI 2.0b, Dual Link-DVI |
GTX 1050 (3GB) | 768 | 1.518 GHz | 1.392 GHz | 3 GB GDDR5 | 96-bit | 75W | 300W | 97C | DP 1.43, HDMI 2.0b, Dual Link-DVI |
GTX 1050 (2GB) | 640 | 1.455 GHz | 1.354 GHz | 2 GB GDDR5 | 128-bit | 75W | 300W | 97C | DP 1.43, HDMI 2.0b, Dual Link-DVI |
Series 9, released Feb 2014. Architecture: Maxwell
Cuda Capability: 5.2
Bus Support: PCIe 3.0
Display Resolution:
- 5120×3200 (Digital)
- 2048×1536 (Analogue)
Series 10 Models | Cuda Cores | Boost Clock | Base Clock | Memory | Bandwidth | Graphics | System | Max. Temp. |
GTX Titan X | 3072 | 1.075 GHz | 1 GHz | 12 GB GDDR5 | 384-bit | 250W | 600W | 91C |
GTX 980 TI | 2816 | 1.075 GHz | 1 GHz | 6 GB GDDR5 | 384-bit | 250W | 600W | 92C |
GTX 980 | 2048 | 1.216 GHz | 1.126 GHz | 4 GB GDDR5 | 256-bit | 165W | 500W | 98C |
GTX 970 | 1664 | 1.178 GHz | 1.05 GHz | 4 GB GDDR5 | 256-bit | 145W | 500W | 98C |
GTX 960 | 1024 | 1.178 GHz | 1.127 GHz | 2 GB GDDR5 | 128-bit | 120W | 400W | 98C |
GTX 950 | 768 | 1.188 GHz | 1.024 GHz | 2 GB GDDR5 | 128-bit | 90W | 350W | 95C |
Useful Reference: