All articles tagged with #gpu benchmark
A 35-GPU benchmark of Control Resonant evaluates performance across presets, DirectX 12, ray tracing, path tracing, and upscaling (DLSS/FSR), while analyzing VRAM usage to inform hardware requirements for an optimal experience on Remedy's Northlight engine.
Geekerwan’s early efficiency tests of Google’s Tensor G6 (G6) and Tensor G5 (G5) on a 3nm process show both chips trail Qualcomm’s Snapdragon 8 Elite and Apple’s A-series in performance-per-watt for CPU and GPU workloads, with Kirin 9030 Pro nearly matching them despite an older process. In Geekbench 6 CPU tests, the G6 peaks around 8,000 points at ~11W; the Steel Nomad Light GPU test shows ~1,200 points at ~6W. Overall, Tensor G6/G5 sit behind current flagship silicon in efficiency, though Kirin 9030 Pro remains a close competitor on power.
A 52-game benchmark with varied presets and ray tracing shows the RTX 5070 Ti and RX 9070 XT are nearly identical in overall performance at 1440p and 4K (RTX ahead by about 1–4%). The RX leads in many rasterization-heavy titles and uses about 20% less power, but the RTX shines in ray tracing and path tracing workloads. Price is the deciding factor for most buyers, with the RX 9070 XT around $700 vs the RTX 5070 Ti around $1,000, making the Radeon the better value despite Nvidia’s RT advantages.
Apple's M3 Max, the flagship of the M3 lineup, outperforms its predecessor, the M2 Max, by 18% in GFXBench's Aztec Ruins test, solidifying Apple Silicon's dominance in the laptop space. While it falls slightly behind Nvidia's RTX 4080 Laptop GPU by 7%, the M3 Max's performance is commendable considering its lower power consumption. The M3 Max-powered MacBook Pros are expected to reach customers by late November.