Nvidia’s RTX 4060 could have been a fan favorite and one of the best graphics cards in the mid-range sector, but it looks like it might be a bit of a flop instead.
Reliable hardware leaker kopite7kimi has shared some specs of the card, and if the rumors are to be confirmed, it looks like Nvidia made some strange choices when creating the GPU.
RTX 4060 still uses PG190. AD107-400-A13072FP328G GDDR6 18Gbps115W24M L2
I will try to remain neutral about any future leaks. 😁😁😁
— kopite7kimi (@kopite7kimi) February 13, 2023
In the previous generation of GPUs, the RTX 3060 was always a popular choice. Powerful enough to offer flawless 1080p gaming, but not overpriced like the rest of the range, the RTX 3060 has a lot to offer even now if you’re building a budget PC.
Most of us expected the RTX 4060 to be just as well received, especially when you consider the unexpected success of the RTX 4070 Ti. Simply put, most of us don’t want to spend $2,000 on a GPU, so these more budget-friendly choices are always welcome — unless they turn out to be worse than their last-gen counterparts.
So far, the RTX 4060 seems very unimpressive compared to the RTX 3060. According to kopite7kim, the card will come with just 3,072 CUDA cores (compared to 3,584 on the RTX 3060), 96 Tensor cores (112 for the 3060), and 24 ray tracing cores (28 ). The number of cores is actually 17% less than the last generation RTX 3060. Of course, the jump in architecture – from Ampera to Ada Lovelace – brings some upgrades in itself, such as access to DLSS 3.
The upcoming RTX 4060 also appears to have significantly less VRAM than its Ampere counterpart. It is said to be released with 8GB of 18Gbps GDDR6 memory, while the RTX 3060 had 12GB but ran at 15Gbps. However, the GPU gets a much-needed L2 cache upgrade, and a huge one at that, marking a 700% increase from the Ampere. The RTX 4060 is said to come with 24MB of L2 cache, while the 3060 only has 3MB. However, that’s still half the RTX 4070 Ti’s L2 cache.
Image used with permission of the copyright holder
All of these strange spec choices can be traced back to the chip that is said to contain the RTX 4060. The author on Twitter predicts that it will be based on a PG190 PCB and an AD107-400-A1 die. Nvidia is already using this die for the mobile versions of the RTX 4060 and RTX 4050. This implies that Nvidia is using a mobile chip for the desktop GPU.
It’s hard to say why Nvidia would want to release such a scaled-down mobile version of the chip. Aside from the Ada Lovelace architecture, the RTX 4060 doesn’t seem to have much to recommend it. However, it could be one of the most conservative GPUs in Nvidia’s lineup, as it’s rumored to have a TDP of 115W – 32% lower than the RTX 3060.
Lower power requirements are always good. Chances are the RTX 4060 won’t melt the way the RTX 4090 did. However, it’s probably safe to say that most users would prefer a spec upgrade for their desktop gaming GPU to a lower TDP, since the 3060 had a pretty reasonable TDP to begin with. We’ll just have to see the final specs when the card hits the market.
Editor’s recommendations
Categories: GAMING
Source: newstars.edu.vn