NVIDIA GeForce RTX 4070

The RTX 4070's strongest measured trait is frames per watt at 0.49 fps/W, which puts it in the 92nd percentile — that metric describes how much of its power budget turns into frames. Its weakest is bandwidth per pixel at 1,398 B/px (22nd percentile). That is a wide spread for a single card: it is specialised rather than balanced, and whether it suits you depends on which of the two you actually load.

The RTX 4070 turns a 200W board power budget into 97.8 fps at 1440p, or 0.49 fps/W against a pool median of 0.32 fps/W. Power, not transistor count, is the binding constraint on a modern GPU — Dennard scaling stopped holding around 2005, and Esmaeilzadeh and colleagues showed at ISCA 2011 that the fraction of a chip that can be powered at full speed shrinks with every node (“dark silicon”). Frames per watt is the shortest honest summary of what an architecture is doing with that budget.

For scale: the closest NVIDIA card from an earlier generation is the NVIDIA GeForce RTX 3080 (2020), at 96.4 fps and 320W. The RTX 4070 is 1% faster while drawing 120W less.

Every rendered pixel has to be fed from memory. At 1440p the RTX 4070 has 1,398 B/px of memory bandwidth available per pixel it draws, against a median of 1,808 B/px — a tight ratio, which is why its frame rate falls off harder than its shader count suggests as resolution and texture quality climb. This is the roofline model applied to a game frame: Williams, Waterman and Patterson (Communications of the ACM, 2009) showed that any workload runs into either a compute ceiling or a memory-bandwidth ceiling, and that the ratio of the two is what decides which one you hit.

On memory capacity, the RTX 4070 carries 12GB against a median of 16GB across the 11 other cards within 10% of its 1440p frame rate. It has less memory than the cards it competes with on speed, which is the constraint to watch: a card fast enough to run high texture settings but short on VRAM to hold them stutters rather than slows.

Method: percentiles are computed over the 77 non-legacy cards with a 1440p figure. Frame rates come from Tom's Hardware GPU hierarchy; specifications from manufacturer datasheets. Theoretical FP32 TFLOPS is deliberately not used as a cross-vendor comparison — RDNA 3 and 4 dual-issue FP32 doubles the paper figure without doubling game performance, and AMD itself declined to count it as extra shaders. The measured Steel Nomad score is used as the compute proxy instead. GPURanker does not run its own benchmarks.