In October 2022, Nvidia unveiled their updated 40 series graphics cards. The star of the series, RTX4090, has top-notch performance, grabbing the attention of many gaming enthusiasts. But, it’s important to note that superior performance comes with a high cost – with a hefty price tag of $1499, it could be a deal-breaker for some.
However, the less advanced RTX 4080 and 4070 have drawn some negative feedback from customers. The criticism often focused on the severe parameters reduction and high price, which for a considerable time made me lose interest in the 40 series of graphics cards altogether. I even contemplated waiting for the launch of the 50 series before making any further considerations.
On January 9th, at CES 2024, Nvidia officially announced the RTX 40 SUPER series graphics cards. The release included the RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER, with respective prices of $999, $799, and $599. This launch truly delivered more value for a lower price. Interestingly, in less than 10 months, we should expect the launch of the RTX 50 series graphics cards. This brings up the question: Is it worth buying a 40 series card at this time?
Before deciding if a 40 series card purchase is a loss or a gain, we should get a basic understanding of their specifications and gauge what upgrades have been made.
Starting with the RTX 4080 SUPER, it progressed from a core aspect of AD103-300 to AD103-400, the CUDA count increased from 9728 to 10240, and the core frequency saw respective increases of 90MHz and 50MHz for the base and acceleration.
The RTX 4080 SUPER graphics card presents a significant upgrade in its memory specifications. Its effective frequency has increased from 22.4GHz to 23GHz, and despite the constants of the bandwidth and capacity of its 256-bit 16GB GDDR6X memory, its bandwidth has reached 736GB/s, surpassing the RTX 4090. As per official claims, the 4080’s performance doubles that of the RTX 3080 Ti with an approximately 40% increase, suggesting a measurable improvement.
Additionally, the RTX 4070Ti SUPER stands out. It’s the first NVIDIA graphics card to simultaneously carry both the Ti and SUPER suffixes. This card, unique among its triple release, has a replaced core chip, upgraded from AD104-400 to AD103-275. It also appreciates in CUDA core count from 7680 to 8448 and memory from 192-bit 12GB GDDR6X to 256-bit 16GB GDDR6X. Consequently, its bandwidth soared to 672GB/s, indicating a substantial enhancement.
Official figures indicate that the RTX 4070Ti Super’s performance is around 1.6 times faster than the RTX 3070Ti. However, various reviews reveal it to be closer to a 1.4 times increase.
The advancements in the RTX 4070Ti SUPER indicate notable progress that exceed expectations. It arguably presents superior cost-effectiveness compared to the currently available RTX 4070Ti, given its average market cost of around $800. A key upgrade is its video memory, which has moved from 12GB 192bit to 16GB 256bit, greatly enhancing its capability to run 4K games. Achieving high or medium settings in some games should no longer be an issue.
The RTX 4070 SUPER is the most accessible option among the new releases and also boasts the most substantial improvements. Key upgrades include a shift in core chip from AD104-250 to AD104-350, along with a surge in CUDA cores from 5888 to a striking 7168, nearly on par with the 4070 Ti. The core base frequency has been boosted by 60MHz, while the acceleration frequency holds steady at 2475MHz.
Interestingly, claims have been made that its performance overshadows that of the RTX 3090. However, online evaluations suggest the RTX 4070’s performance is merely 20% faster than the RTX 3070. This explains the initial skepticism and jests from the online community, as the RTX 4070 SUPER’s performance rating falls significantly short when compared to the leading RTX 3090.
In summary, from the three new graphics card releases, the RTX 4070 SUPER represents the best value for money. It excels in delivering enhanced capacity at a reduced price. The other two cards provide increased capacity at no additional cost. These three graphics cards are certainly worthwhile considerations for anyone planning imminent upgrades.
After evaluating leaks from @mooreslawisdead, it appears that the advancements in the 50 series graphics cards are relatively modest, even drawing comparisons of squeezing out the last drop. The 50 series, hosting the Blackwell architecture GPU, primarily targets enhancements in energy efficiency rather than sheer performance, therefore, it’s unlikely to see the nearly 100% performance increase from the transition of the 30 series to 40 series.
Taking into account the market positioning of the RTX40 series, a reasonable assumption is that upgrades to the budget-friendly RTX 5050, 5060, and 5070 graphics cards may be minimal, leaving the primary focus on the high-end RTX 5080 and RTX 5090. Put simply, acquiring a 40 series SUPER graphics card now may offer equivalent performance as the 50 series while potentially costing less.
However, Nvidia’s current market situation allows for this strategy. Firstly, revenue from the data center has significantly surpassed that of the gaming graphics cards, indicating a priority shift from the gaming experience to the data center. Secondly, with AMD’s graphics cards not being compelling, and rumors suggesting their next-generation RDNA 4 architecture may not have top models to compete with Nvidia, let alone Intel having any impact, the business focus migration and decreasing competitive pressure mean that Nvidia doesn’t have to push aggressively in R&D and upgrades as prior. After all, everyone prefers making money the easier way.
Certainly, this updated focus may not sit well with us as gaming enthusiasts. Typically, Nvidia and AMD have been among the go-to options for gamers. However, as AMD shifts attention away from the high-end venture and Nvidia plunges determinedly into the AI sphere, it’s an understandable source of discontent among gamers, particularly the more budget-conscious ones.
Nevertheless, it’s relieving that Nvidia still keeps the gaming community in mind. Numerous AI technologies disclosed at their recent conference have applicable value to gaming, promising an elevated gaming experience for users. Yet, the question remains how attentive to the gaming market Nvidia will be, especially in the new age of 60 series, 70 series, and 80 series graphics cards – it is an answer yet to be known.