Nvidia RTX 3060 vs 3070
Can the newer card bring the fight?
Looking through the low end of Nvidia’s GPU lineup you may consider the RTX 3060 vs 3070. What are the key differences between these two graphics cards and which will compete for the best graphics card?
The 3060 came out four months after the initial 30 series lineup including the 3070. Coming later gives it the advantage of some better specifications. Stemming from the fact it was in reaction to AMD’s lineup, so we take a look at how it might affect the card in its own family.
Both of the RTX 3060 vs 3070 are based on the Ampere architecture but have varying GPUs. The 3060 features the GA106 die, with the 3070 wielding the GA104. The upgraded chip was built to a bigger size with more transistors and cores.
GA106 has a size of 276mm2 compared to the GA104 which has a size of 392mm2. And they fit in 12 billion and 17.4 billion transistors respectively in those dies. Both of them are made with Samsungs 8nm process, the standard across the generation but unlikely for the next.
|RTX 3060||RTX 3070|
|Memory||12GB GDDR6||8GB GDDR6|
|Base clock||1320 MHz||1500 MHz|
|Boost clock||1777 MHz||1725 MHz|
|Memory clock||15 Gbps||14 Gbps|
|Card bus||PCIe 4.0 x16||PCIe 4.0 x16|
|Memory bandwidth||360.0 GB/s||448.0 GB/s|
|TDP||170 W||220 W|
Comparing the graphics cards at their specification level can give us a quick look at what to expect the difference to be. It gives an understanding of the main differences between the two GPUs are. We take a closer look at the core differences are and what they mean.
The video memory of a GPU is used as a frame buffer and texture map storage. Both to decrease latency between the GPU and CPU and allow better running of higher resolution games. As greater screen sizes require more pixels which are stored in the VRAM.
Strangely, in this situation, the 3060 has a higher memory count. With 12GB of GDDR6, it is more focused on trying to compete with the 6600 XT from AMD which has 8GB and may perform better. Whereas, the 3070 only has 8GB of the same VRAM.
However, they don’t run at the same clocks and bandwidths. With a higher memory bus of 256 bit the 3070 can throughput much more data. It has a bandwidth of 448GB/s compared to the 3060s 192 bit and 360GB/s. So performance will come down to speed vs capacity.
The core clock of a GPU determines how fast the die processes graphical instructions. The speed is a direct correlation to the cycle rate of the silicon itself. Although not the sole basis for performance it does contribute and is favorable to be overclocked to extract better performance.
Between these two graphics cards out of the box, the 3070 has the higher base clock. It comes at a speed of 1500MHz, with the 3060 coming in at 1320MHz. Then for the boost clock, it is in fact the 3060 that reaches higher speeds. With a boost clock of 1777MHz whereas the 3070 has a lower boost of only 1725MHz.
With the upgraded GA104 chip the 3070 has an increased amount of available GPU cores. Embedded in its heart the 3070 has 64% more cores than the 3060. Giving it greater power in processing the relevant processes.
In particular, we take a look at shaders, RT cores, and TMUs. These are specialized cores in the die that only work on particular parts of graphics. For these, they are on shadows and shading of scenes, real-time ray tracing, and creating 3d images from bitmaps.
Thermal design power is the maximum amount of power a GPU die will use at any given time. This in turn itself dictates how much heat the card will produce and how well the cooling system must be designed to keep from overheating.
Comparing the 3060 vs 3070 TDPs we find the 3070 power at 220W compared to the 170W of the 3060. Meaning with all the extra power you also require a higher power usage. As well as having a greater heater inside your case that will affect your temps inside and out.
RTX 3060 vs 3070 performance
Using 3dmarks benchmarks we can find the difference between the two GPUs in terms of graphics scores. Unsurprisingly it is the 3070 that takes the lead in this scenario. It gets a score of 13,679 with a price to performance score of 27. While the 3060 gets 8,755 along with a price to performance of 26. This means the 3070 gets a score of 56% higher than the 3060 and 1 point better in price.
Benchmarking offers some better tests to see the performance difference between the two cards. The 3070 offers a higher score 55% higher than the 3060. In gaming, the 3070 achieves on average 31% more FPS across the board at ultra settings and all resolutions. With the 3060 not able to achieve above 60 FPS at 4k even with its higher VRAM count.
Taking a look at the cost of the GPUs will give a better indication of what you can get for your money and to suit your needs. The 3060 has an RRP of $329 (£300), with the 3070 setting an asking price of $500 (£450). For 50% higher cost you don’t get exactly 50% more performance but it’s not far off in context. Especially if you’re looking for 4k performance.
Although in the current market GPUs are expensive and have suffered from a shortage. So it has been hard to find them at MSRP, so we take a look at what the market asking price is. For the 3060 it is $590 (£455) and for the 3070 a staggering $1250 (only £671 in the UK). Which may be a big influence on your choice.
Overall when looking at the RTX 3060 vs 3070 it is clear the 3070 takes it in terms of performance. But it does cost more so only worth it for the 4k ability and if you can get it close to MSRP. However, if you’re still stuck deciding to go in between may be an option with the RTX 3060 Ti. Or if you have time wait for Intel ARC Alchemist or RTX 4000 series for better choice and pricing.
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn more