Gigabyte GeForce RTX 3060 GAMING OC 12GB Graphics Card

Product Information

The full GA106 chip has 12 billion transistors, down from 17.4 billion in GA104. That shrinks the die size from 393mm square to just 276mm square, which not only helps to reduce the cost of the chip, but also increases the number of chips Nvidia can get from a single wafer — and if you're wondering, GA106 is less than half the size of GA102, which measures 628.4mm square and has 28.3 billion transistors. At an estimate, Nvidia can get around 130 dies per wafer with GA104 (some of which are defective, most of which end up as partially disabled chips), while the smaller size of GA106 allows for around 200 dies per wafer. More dies mean better yields and more graphics cards to go around. That's the hope.

The biggest change in Ampere is in the Streaming Multiprocessor’s (SM) datapath. Where previous generations had separate pipelines for Integer and Floating Point 32 (FP32) workloads, Nvidia has effectively doubled the amount of cores that can handle standard FP32 work – which just happens to be incredibly important for gaming performance. Because of this, the amount of CUDA cores per SM has doubled to 128, rather than the 64 found in something like card from our Nvidia GeForce RTX 2060 review. Our benchmark suite is pretty taxing, with lots of ray traced titles represented. This should prove beneficial to the RTX 4060, but the A750's ray tracing performance is actually quite good. DLSS 3 is Nvidia's golden goose this generation, but Intel's XeSS is maturing nicely and if you play games that support it, it can be a tremendous value add. Which one should you go for? The RTX 4060 is cheaper, offers much better performance per watt and has the advantage of support for DLSS 3 and Frame Generation. Both have the same amount of VRAM. But the fact that we're even considering the RTX 3060 Ti over the RTX 4060 is damning in and of itself.

Board Design

We're going to look at the 1080p and 1440p data for half a dozen of the 14 games tested before we jump into our usual performance breakdown graphs. The scaling here isn’t perfect, however, but thanks to this massive difference in the amount of shaders available to the RTX 3060, it is able to achieve double the performance of the GTX 1060 in many games. The RTX 4060 is generally a match for the RX 7600 in traditionally rendered titles, but once you add ray tracing into the mix, the RTX 4060 pulls ahead, and that's before the gains it will achieve if you enable DLSS 3 and Frame Generation. AMD has an answer to that in the works, but until we see FSR 3 out in the real world, Nvidia holds the upscaling technology advantage. Leaving the RX 6600 XT and RX 6650 XT's runout pricing aside, the RTX 4060 is superior by every important measure. Its performance is better, particularly when factoring in DLSS 3. The differences aren't as dramatic when looking at raster performance, but the RTX 4060 is still ahead there too.

The 1440p data is mostly more of the same: 10% slower than the 5700 XT and now 20% slower than the 3060 Ti. It matches the RX 5700 again, so it was 9% faster than the RTX 2070. In Watch Dogs Legion at 1440p, the RTX 3060 was good for just 62 fps on average at the native resolution. Enabling quality DLSS boosted the frame rate by 30% to 81 fps, a very nice performance upgrade. It’s powered by a single 8-pin PCIe connector, which means you don’t need to worry about an unsightly 12-pin adapter. Again, another benefit for users that want to build in a smaller case. I/O is also incredibly simple, with three DisplayPorts and one HDMI 2.1 port. For testing we're still using the Ryzen 9 3950X test system as we progress on updating everything to be tested under the R9 5950X. Expect all that data once we get through this latest round of product releases. The test system is configured with 32GB of dual-rank, dual-channel DDR4-3200 CL14 memory.

And you really can’t talk about Nvidia Ampere without talking about ray tracing and Deep Learning Super Sampling (DLSS). Both of these technologies are supported by the RTX 3060, though the smaller GPU means that there is less dedicated hardware for RT and DLSS workloads. It's a tie when it comes to feature sets. Both the RTX 4060 and Ti are Ada Lovelace generation cards, which mean both come with support for DLSS 3, 3rd generation RT cores, 4th generation Tensor cores and an improved video engine. The GeForce RTX™ 3060 is powered by Ampere—NVIDIA’s 2nd gen RTX architecture. Built with enhanced Ray Tracing Cores and Tensor Cores, new streaming multiprocessors, and high-speed G6 memory up to 8GB, giving you the power you need to rip through the most demanding games.