NVIDIA GeForce RTX 4070 SUPER vs NVIDIA GeForce RTX 4080 SUPER

Every spec side by side, from each maker's published figures.

Better figure, where one is clearly better

GeForce RTX 4070 SUPER compared with GeForce RTX 4080 SUPER
NVIDIAGeForce RTX 4070 SUPERNVIDIAGeForce RTX 4080 SUPER
Overview
ReleasedJanuary 2024January 2024
Performance
Chip makerNVIDIANVIDIA
ArchitectureAda LovelaceAda Lovelace
Shader cores7,16810,240
Boost clock2,480 MHz2,550 MHz (better)
Ray tracingThird-generation RT Cores; 82 RT TFLOPSThird-generation RT Cores; 121 RT TFLOPS
AI coresFourth-generation Tensor Cores; 568 AI TOPSFourth-generation Tensor Cores; 836 AI TOPS
Upscaling and frame generationDLSS 3.5 with Super Resolution, DLAA, Ray Reconstruction and Frame GenerationDLSS 3.5 with Super Resolution, DLAA, Ray Reconstruction and Frame Generation
Base clock1.98 GHz2.29 GHz
Cache48 MB L264 MB L2
Memory
Video memory12 GB GDDR6X16 GB GDDR6X (better)
Memory bus192-bit256-bit (better)
Memory bandwidth504 GB/s736 GB/s (better)
Memory detail–GDDR6X at 23 Gbps
Connections and power
Board power220 W (better)320 W
InterfacePCIe 4.0PCIe 4.0
Display outputs3 × DisplayPort 1.4a, 1 × HDMI 2.1a; up to 4 displays; HDCP 2.33 × DisplayPort 1.4a, 1 × HDMI 2.1a; up to 4 displays; HDCP 2.3
Maximum resolution4K at 240 Hz or 8K at 60 Hz with DSC, HDR4K at 240 Hz or 8K at 60 Hz with DSC, HDR
Power connector1 × 300 W or greater PCIe Gen 5 cable, or 2 × PCIe 8-pin cables with the adapter in the box; some partner cards use 1 × 8-pin1 × 450 W or greater PCIe Gen 5 cable, or 3 × PCIe 8-pin cables with the adapter in the box
Recommended power supply650 W750 W
Slot width2-slot (Founders Edition)3-slot (Founders Edition)
Size (reference card)244 × 112 mm, 2-slot (Founders Edition)304 × 137 mm, 3-slot (Founders Edition)
Video encode and decode1 × eighth-generation NVENC, 1 × fifth-generation NVDEC; AV1 encode and decode2 × eighth-generation NVENC, 1 × fifth-generation NVDEC; AV1 encode and decode

More comparisons