Nvidia Faces Criticism Over GeForce RTX 4000 Graphics Cards
Nvidia has been receiving a lot of criticism for its GeForce RTX 4000 graphics cards. While the new cards have significantly improved efficiency, offering higher performance with lower power consumption, they have faced criticism from all sides in the mainstream market. The only aspect that has received some praise is ironically the low power consumption, which initially seemed like it would be the most criticized feature. However, the GeForce RTX 4060 has been particularly poorly received, with all three versions facing unfavorable reviews.
The basic RTX 4060 and RTX 4060 Ti have been criticized for their small 8GB memory, which runs on a narrower 128-bit bus. Although the RTX 4060 Ti with 16GB of memory should address this issue, it comes at a high price and still has low bandwidth. It has now been revealed why these cards are so expensive and why the 8GB version is $100 cheaper.
The issue is not just about the memory chips themselves. A 1GB GDDR6 chip currently costs less than $3.50, so adding an additional 8GB of memory should cost the manufacturer around $25-30 (with an additional margin, the price might not be higher than $40-50). However, the problem lies in the configuration of these memory chips. The RTX 4060 supports a maximum 128-bit memory interface, which means that only 4 positions can be filled with 32-bit GDDR6 chips. This is not a problem for the 8GB variant, which uses 4x2GB chips. However, for a 16GB configuration, 4GB chips would be needed, but they do not exist.
The situation is being addressed with a so-called clamshell configuration, which allows access to 8 chips through a 128-bit interface. However, this also means that each chip communicates using only a 16-bit bus instead of the standard 32-bit bus. This requires more complex circuits that split the communication between two chips, increasing the cost of the card. While it is unlikely that this adds an additional $70 to the price (it will probably be in single-digit USD), it is one of the reasons why the price of the 16GB RTX 4060 Ti is significantly higher. However, the main reason is likely where everyone expects it to be, Nvidia’s desire to make as much money as possible (which is quite understandable).
But what about performance? It is usually claimed that 8GB is not enough for modern cards, so according to this logic, 16GB of memory should bring significantly better results. Unfortunately, reviews show that this is not the case. For example, Techspot measured an average performance increase of less than 5% and 1% Low, where the large memory should have the most impact, was only around 6% in 1080p resolution. However, high differences cannot be expected at low resolutions, but rather at higher ones. In 1440p resolution, the difference was even less than 3% on average, but the differences in 1% Low were much more significant, with the 16GB version bringing a 20% performance increase.
Even more interesting results were obtained by TechPowerUp. There, the differences in 1080p were not noticeable at all, with the average performance being essentially the same. In 1440p, there was a 1% increase in performance, and in 4K, the 16GB version was only 2% faster. As for 1% Low, the 16GB version brought a 1.3% improvement in 1080p, a 1.2% improvement in 1440p, and a 5.3% improvement in 4K. The results areNvidia’s GeForce RTX 4000 graphics cards have been receiving a lot of criticism lately. While the new cards have significantly improved efficiency, offering higher performance with lower power consumption, they have faced criticism from all sides in the mainstream market. The only aspect that can be praised is ironically the low power consumption, which initially seemed like it would be the most criticized feature. However, the GeForce RTX 4060 has received particularly negative reviews, with all three versions facing unfavorable criticism.
The basic RTX 4060 and RTX 4060 Ti have been criticized for their small 8GB memory, which runs on a narrower 128-bit bus. Although the RTX 4060 Ti with 16GB of memory should address this issue, it comes at a very high price and still has low bandwidth. It has now been revealed why these cards are so expensive and why they cost an additional $100 compared to the 8GB version.
The issue is not just about the memory chips themselves. A 1GB GDDR6 chip currently costs less than $3.50, so adding another 8GB of memory should cost the manufacturer around $25-30 (and with an additional margin, the price shouldn’t be higher than $40-50). However, the configuration of these memory chips is also a problem. The RTX 4060 supports a maximum 128-bit memory interface, which means that only 4 positions can be filled with 32-bit GDDR6 chips. This is not a problem for the 8GB variant, which uses 4x2GB chips. However, for a 16GB configuration, 4GB chips would be needed, but they do not exist.
The situation is being addressed with a so-called clamshell configuration, which allows access to 8 chips using a 128-bit memory interface. However, this also means that each chip communicates using only a 16-bit bus instead of the standard 32-bit bus. This requires more complex circuits that split the communication between two chips, increasing the cost of the card. While it may not add an additional $70 to the price (probably just a few dollars), it is one of the reasons why the price of the 16GB RTX 4060 Ti is significantly higher. However, the main reason is likely where everyone expects it to be, Nvidia’s desire to make as much money as possible (which is quite understandable).
But what about performance? It is usually claimed that 8GB is not enough for modern cards, so according to this logic, 16GB of memory should bring significantly better results. Unfortunately, reviews show that this is not true. For example, Techspot measured an average performance increase of less than 5% in 1080p resolution and less than 6% in 1% Low, where the large memory should have the most impact. However, significant differences cannot be expected at low resolutions, but rather at higher ones. In 1440p resolution, the difference was even less than 3% on average, but the differences in 1% Low were much more pronounced, with the 16GB version bringing a 20% performance increase.
TechPowerUp provided even more interesting results. In their tests, there were no differences in 1080p resolution, and the average performance was essentially the same. In 1440p, there was a 1% increase, and in 4K, the 16GB version was only 2% faster. As for 1% Low, the 16GB version brought a 1.3% improvement in 1080p, a 1.2% improvement in 1440p, and a 5.3% improvement in 4K. The results are slightly different from the first test, but as we can see, the quarter-priced card brings increases in performance in very low percentages and almost no increase in performance. And even in the case of the biggest frame rate drops, it doesn’t help as much as we might expect.
How does the decision to use a non-standard 32-bit bus in the 16GB RTX 4060 Ti affect the cost of the card?
Instead of the standard 32-bit bus. This requires more complex circuits, increasing the cost of the card. While it may not add an additional $70 to the price, it is one of the reasons why the 16GB RTX 4060 Ti is significantly more expensive. However, the main reason is likely Nvidia’s desire to maximize profit.
But what about performance? It is often claimed that 8GB is not enough for modern cards, so 16GB of memory should bring better results. However, reviews show that this is not necessarily the case. Techspot measured an average performance increase of less than 5% and only around 6% for the 1% low performance, which is expected to benefit the most from larger memory, in 1080p resolution. In 1440p resolution, the difference was even less than 3% on average, but the 16GB version showed a 20% performance increase in the 1% low category.
TechPowerUp obtained even more interesting results. In 1080p, there were no noticeable differences in performance, with the average performance essentially the same for both versions. In 1440p, there was a 1% increase in performance, and in 4K, the 16GB version was only 2% faster. As for the 1% low category, the 16GB version showed a slight improvement across all resolutions, ranging from 1.2% to 5.3%.
Overall, Nvidia’s RTX 4000 graphics cards have faced criticism for their small memory configuration and high price for the 16GB version. While the low power consumption is praised, the performance increase from the larger memory does not seem significant enough to justify the cost.
Nvidia GeForce RTX 4000: Amazing performance and impressive pricing!
The Nvidia GeForce RTX 4000 seems to be a promising addition to the graphics card market. I’m curious about the performance it offers and how it compares to its predecessors. However, I hope the prices are not too steep, making them out of reach for the average consumer.