Why You Should Buy the 16GB RTX 4060 Ti if You are a Content Creator

 If you're a PC gamer looking to buy a mid-tier GPU, then I'll be straight with you: the newly released 16GB version of the RTX 4060 Ti priced at $499 isnt the one for you. However, if you happen to be a content creator, a skilled video editor, or even a fellow YouTuber like myself, this GPU becomes an entirely different story. In fact, the 16GB RTX 4060 Ti could be the most compelling sub $500 GPU option available for those involved in content creation. So let's talk about it.

Before talking about RTX 4060 Ti, let's start with RTX 4060 which is priced at $299. In all honesty, one could argue that this card should be named the RTX 4050, considering Nvidia's decision to decrease its VRAM from 12GB, as seen in the RTX 3060, to a mere 8GB. But they decreased its price. In fact the cheapest RTX 3060 card that I can find in the market today cost around $40 more than the RTX 4060. So I think RTX 4060 will be a great entry level GPU for folks who are just starting out with basic 4K video editing. What sets it apart from similarly priced AMD and Intel GPUs is the inclusion of the 8th Generation NVIDIA NVENC Encoder. This is the same encoder that you will find in higher tier Nvidia GPUs and it supports the latest AV1 codec. This means you can work with cutting-edge video formats while experiencing improved encoding and decoding performance.

But when it comes to gaming, there are some uncertainties surrounding the performance of the RTX 4060. It might be wise to hold off and consider waiting for the AMD Radeon RX 7600, which is expected to be similarly priced but potentially offer better rasterization performance than Nvidia's offering. Nvidia claims that the RTX 4060 is approximately 20 percent faster than the RTX 3060. But I will take these claims with a grain of salt. Nvidia's own benchmarks appear somewhat misleading because in their comparison charts, they showcase RTX 4060 frame rates with DLSS 3.0 enabled. The actual performance difference between the RTX 4060 and the RTX 3060 is likely to be in the single-digit percentage range. Moreover, the reduction in video RAM from 12GB to 8GB is a big disappointment. With many new gaming titles requiring increasingly higher amounts of video RAM, especially when aiming to crank up the resolution and graphics settings, this reduction could pose limitations.

Now, let's talk about RTX 4060 Ti. Personally, I believe it's best to overlook the 8GB model of this GPU, as it simply doesn't justify its $400 price tag. One of the main concerns with this variant is its narrow 128-bit memory bus and the limited 8GB RAM. As a result, it appears like a modest upgrade over RTX 3060 Ti. Here are the benchmark results that Nvidia shared. And you can clearly see that the titles where RTX 4060 Ti appears to be ahead are actually DLSS 3.0 enabled benchmarks. In my opinion these charts are a bit misleading when evaluating pure rasterization performance. While there are some improvements, I don't think a groundbreaking generation leap

Interestingly, Nvidia is now largely positioning the 8GB version of the RTX 4060 Ti as a 1080p gaming card. This implicitly acknowledges that 8GB VRAM is not sufficient for playing most modern titles at maximum settings on a 1440p resolution. However, to give credit where it's due, Nvidia did increase the L2 cache from 4MB to 32MB in the RTX 4060 Ti. A larger cache allows for storing frequently used instructions and textures, which reduces the number of times the GPU needs to access the VRAM. This improvement should help mitigate some of the disadvantages of the narrower 128-bit memory bus. According to Nvidia's, this additional cache boosts the effective memory bandwidth from 288 GB/s to 554 GB/s. But I will take this claim with a grain of salt. Honestly, I would skip this card and wait for the arrival of 16GB version of RTX 4060 Ti

Now as a content creator and video editor, I am really looking forward to the 16GB version of RTX 4060 Ti. The standout feature of this card is obviously its 16GB VRAM. Up until now, if you wanted a 40 series RTX GPU with 16GB VRAM, you had to shell out around $1200 for the RTX 4080. Even the RTX 4070 Ti and RTX 4070 come with only 12GB of VRAM. So if you are a content creator, the 16GB version of RTX 4060 Ti looks like a perfect $500 video card.

In fact the 16GB version of the RTX 4060 Ti effectively renders RTX 4080, RTX 4070 Ti, and RTX 4070 obsolete for the majority of content creators. Of course folks doing advanced 3D animation in Blender will go with RTX 4090, because it's a significantly faster GPU and offers a substantial 24GB of VRAM. But despite boasting double the VRAM compared to RTX 3060 Ti, RTX 4060 Ti is still using the same 128-bit memory bus. Remember, the last generation RTX 3060 Ti had a memory bandwidth of 448 GB/s despite coming with only 8 GB VRAM. This was because it came equipped with a 256 bit memory bus.  To understand how badly this affects the performance of RTX 4060 Ti, we first need to understand the significance of memory bus size. Picture it as a bustling highway with multiple lanes. The more lanes dedicated to traffic, the smoother the flow. So RTX 4060 Ti with 128 bit memory will typically use 4 memory chips since each memory chip has a 32-bit wide bus. Since each memory chip offers 2GB VRAM, four memory chips will only give you a total of 8GB VRAM.

So how did Nvidia manage to put 16 GB VRAM in this GPU? Nvidia is using something called "memory Clamshelling". Basically Nvidia put 4 2GB memory chips on each side of the PCB. So effectively there are 4 pairs of 2GB memory chips sharing memory controllers and 128 bit memory bus. Nvidia has employed this technique in the past to double the VRAM in RTX 3090 from 12GB to 24GB. Take a look at this image showcasing the front side of the RTX 3090 PCB. Right in the center, you'll spot the GA102 graphics chip, surrounded by 12 memory chips, each with a 1GB capacity. Flip the PCB over, and you see an additional 12 memory chips on the back. Effectively, there are 12 pairs of memory chips, all sharing the 12 memory controllers and the 384-bit memory bus. This implementation had allowed Nvidia to achieve the impressive 24GB VRAM capacity without needing to resort to major design overhauls or additional memory controllers. Now the Memory Clamshelling worked on the RTX 3090 because it had a wide 384-bit memory bus which allowed for efficient data transfer and did not result in any bottlenecks. But with a narrow 128-bit memory bus on the RTX 4060 Ti, there is a legitimate concern that the memory Clamshelling arrangement may not perform as efficiently. The constrained bandwidth could potentially limit the overall effectiveness of sharing memory across the eight chips. Nvidia has countered this argument by saying that 32MB of L2 cache would more than compensate for narrow memory bus. In fact in their charts Nvidia claims that RTX 4060 Ti would offer an effective bandwidth of 554 GB/s which is even higher than 448 GB/s that RTX 3060 Ti offered with its 256 bit memory bus. How much of it is true and how much is Nvidia marketing bullshit, we will find out when the cards gets released.

Of course, alternatives like the Radeon RX 6800 or even the RX 6800 XT can be found for under $500, and they also offer 16GB of VRAM. Additionally, Intel's Arc A770 offers 16GB VRAM for only $350. However, when it comes to performance and reliability, these cards simply cannot compete with Nvidia. A quick glance at the PugetBench Davinci Resolve score paints a compelling picture. In content creation workloads, Nvidia clearly dominates AMD and Intel. Personally, as a user of Davinci Resolve for video editing, I can attest to the superiority of Nvidia's CUDA cores over those of AMD and Intel. The difference in performance is stark, and Nvidia's CUDA cores outshine the competition by a significant margin.

Another crucial factor that sets Nvidia apart is the stability of their drivers. While both AMD and Intel drivers can be prone to flakiness and have a tendency to crash Adobe Premiere and Davinci Resolve, Nvidia's studio drivers remain rock solid. In fact, I have experienced zero crashes while using Nvidia GPUs for video editing over the past three years. This level of reliability allows me to focus on my work without interruptions or frustrations.

Overall, I feel like the 16GB version of the RTX 4060 Ti presents an exceptional value proposition for content creators and video editors. With its ample VRAM, it bridges the gap between high-end offerings and more affordable options, making it an attractive choice at the $500 price point. Nvidia's dominance in performance, especially in applications like Davinci Resolve, coupled with their reliable and stable drivers, solidify their position as the go-to brand for content creators seeking a seamless and powerful editing experience.

So what do you think of RTX 4060 Ti? I'd love to hear your thoughts on whether the 16GB version of the RTX 4060 Ti is a worthwhile purchase. Let me know if you mostly game on your GPU, or you use it to livestream, edit videos and do content creation. What price point would you consider to be the maximum you'd be willing to pay for RTX 4060 Ti? Feel free to share your opinions in the comments section below.

By the way, if you're in the market for a GPU specifically for video editing and budget is not a concern, I highly recommend checking out this video over here. In this video, I dive deep into the latest PugetBench results, providing an in-depth analysis of the video editing performance comparison between the RTX 4090 and RX 7900 XTX. So go ahead and watch that video, and I will see you guys over there. Bye!