When it’s time to buy a new GPU, buyers are often faced with the dilemma: “should I choose GDDR6 over GDDR5?” and vice-versa. GDDR6 has a higher number, so it must be better. But, by how much? Even if it is better, is it worth the extra bucks? In this article, we will put numbers to these questions.
GDDR6 vs GDDR5: What is the Difference?
The first question is why GDDR6 and GDDR5 have different numbers. Let’s take a look at this.
GDDR6 and GDDR5 are graphics memory standards. GDDR6 is based on DDR4 and GDDR5 is based on DDR3. Do not confuse GDDR5 with DDR5. Check out more about DDR5 memory and whether they are worth it here. The main difference with DDR is that GDDR can perform both read and write operations in one go, and has a much higher read/write speed as compared to DDR, albeit at the cost of latency. This difference is fueled by the different tasks these memory types have to perform. GDDR is mainly used in graphics processing, machine learning, and other such tasks where the need for accuracy is much lower. But, system memory operates much slower with latencies in the order of nanoseconds. This makes DDR much more accurate as compared to GDDR.
GDDR5 has been the industry standard for quite some time. It had a per-pin bandwidth of 8Gb/s. In 2016, graphics card manufacturers launched a revision of GDDR5 called the GDDR5X. This revision doubled the per-pin memory bandwidth to 16Gb/s. This enabled the chips to become much faster, but this came at the cost of high temperatures. Thus, cards rocking GDDR5X, like the GTX 1080 Ti, required beefy heatsinks.
GDDR6 subverted the problems faced by GDDR5. It came with larger modules, so card manufacturers could put in more memory in lesser space. It also increased the memory bandwidth to 16Gb/s, making it the fastest on the market. Also, GDDR6 tends to be much colder than both GDDR5 and GDDR5X. This made card manufacturers hop onto the newer market standard. Cards with GDDR6 memory started rolling out in late 2018, and Nvidia’s RTX 20 Series and AMD’s RX 5000 series were the first products in the market featuring GDDR6 memory.
In 2020, Nvidia announced that they were using GDDR6X for their RTX 3080 and RTX 3090. Later, RTX 3070 Ti, RTX 3080 Ti, and RTX 3090 Ti also joined the GDDR6X club. Much like GDDR5X, GDDR6X is a revision to GDDR6 standards. It increases the per-pin bandwidth to 19-21 Gb/s with a total memory bandwidth of over 1 TB/s.
GDDR5 vs GDDR6: Which One is Better For You?
If you are a normal user, who wants to browse the internet, consume media, and do office work on your computer, graphics cards with GDDR5 memory is going to be fine.
If you are a gamer, the performance of graphics cards does not depend a lot on VRAM generation. The GTX 1080 with 8GB of GDDR5 memory will always give more frames than the GTX 1660 Super with 6GB of GDDR6 memory. Thus, the choice depends on much more than just the VRAM generation your card rocks. Although GDDR6 is much faster and power-efficient than GDDR5, do not let this difference determine your decision. Often between two cards with the same amount of VRAM from the two memory generations, the card rocking the older GDDR5 memory can beat the newer one. Always check multiple reviews and benchmarks for any graphics card you are considering, and land on a conclusive decision.