Introduction

MSI GeForce RTX 4070 Gaming X Trio leads the company’s custom-design RTX 4070 graphics card series from the front. The Gaming X line of graphics card propelled MSI to the top of the custom graphics market. The company succeeded in fusing the best performance, noise, and product aesthetics into a well-rounded product The RTX 4070 Gaming X Trio offers factory-overclocked speeds of an impressive 2610 MHz, compared to 2475 MHz reference. It serves up a well-executed RGB LED setup, the company’s latest Tri Frozr triple-slot, triple-fan cooling solution, with its latest TorX fans that come with webbed, axial-flow impellers. Supporting its factory-overclock is a powerful VRM solution that uses the modern 16-pin 12VHPWR power connector (this is the only custom-design we’ve tested that uses it). MSI is pricing the Gaming X Trio at $650, a nifty premium over the NVIDIA MSRP, which is the only reason it missed out on review coverage on April 12.

The GeForce Ada graphics architecture that the RTX 4070 is based on, debuts the third generation of RTX, NVIDIA’s ground-breaking technology that steps-up realism in games by combining real time ray traced elements with classic raster 3D graphics. Even this bit of ray tracing requires enormous compute power, and so the company created dedicated hardware inside the GPU that takes care of these workloads. Ada debuts the 3rd generation RT core with a generational uplift in ray tracing intersection performance; and 4th generation Tensor cores, which accelerate AI deep-learning neural nets, by tapping into even newer capabilities. The Ada CUDA core, plus higher GPU clock-speeds, and a completely redesigned memory sub-system with larger on-die caches, make up the new architecture.

The RTX 4070 is based on the same “AD104” silicon that the RTX 4070 Ti maxes out, but is heavily cut down. It features 5,888 CUDA cores, 46 RT cores, 184 Tensor cores, 64 ROPs, and 184 TMUs. The memory setup is unchanged from the RTX 4070 Ti—you get 12 GB of 21 Gbps GDDR6X memory across a 192-bit wide memory bus, yielding 504 GB/s of memory bandwidth, which is still higher than the 448 GB/s the previous-generation RTX 3070 manages over its 256-bit memory bus. The core-configuration, particularly the shader count, is identical to that of the RTX 3070, but over a brand new architecture.