Nvidia has launched its latest generation of graphics cards, which encompasses both desktop and laptop models. These GPUs are commonly known as Nvidia Ada Lovelace. The RTX 40-series made its debut in 2022 and is expected to receive further updates with new cards throughout 2023.
This article covers the Nvidia RTX 40-series’ features, cost, architecture, performance, DLSS 3, and more.
RTX 40-Series Specs
Nvidia’s RTX 40-series comprises four powerful GPUs, each with its own unique specifications:
RTX 4090 | RTX 4080 | RTX 4070 Ti | RTX 4070 | |
Architecture | Ada Lovelace | Ada Lovelace | Ada Lovelace | Ada Lovelace |
Process node | TSMC N4 | TSMC N4 | TSMC N4 | TSMC N4 |
CUDA cores | 16,384 | 9,728 | 7,680 | 5,888 |
Ray tracing cores | 144 3rd-gen | 76 3rd-gen | 60 3rd-gen | 46 3rd-gen |
Tensor cores | 576 4th-gen | 304 4th-gen | 240 4th-gen | 184 4th-gen |
Boost clock speed | 2,235MHz | 2,205MHz | 2,610MHz | 2,475MHz |
VRAM | 24GB GDDR6X | 16GB GDDR6X | 12GB GDDR6X | 12GB GDDR6X |
Memory clock speed | 10,500MHz | 10,500MHz | 10,500MHz | 10,500MHz |
Memory data rate | 21Gbps | 21Gbps | 21Gbps | 21Gbps |
Bus width | 384-bit | 256-bit | 192-bit | 192-bit |
Total Graphics Power (TGP) | 450W | 320W | 285W | 200W |
These GPUs are built on TSMC’s advanced 4nm process node, offering significant improvements in performance and efficiency over their predecessors.
RTX 40-series: Pricing and release date
In this generation, Nvidia made a powerful entrance with the introduction of its flagship RTX 4090. The first announcement of this top-tier graphics card was made on September 20, 2022, and became available for purchase on October 12 of the same year.
Right from the beginning, the RTX 4090 Founders Edition cost $1,600. However, it’s worth noting that many of the cards produced by Nvidia’s board partners carried even higher price tags.
Much like the challenging days of the GPU shortage, the RTX 4090 often faced stock shortages. It was often resold on eBay for more than twice its original retail price in the days after its launch, and scalpers are still making money off of it today.
Following the RTX 4090, Nvidia introduced the RTX 4080 with both 16GB and 12GB variants. The 16GB version became available in stores on November 16, priced at $1,200. Interestingly, the RTX 4080 12GB faced some controversy regarding its price and expected performance, to the extent that it was temporarily “unlaunched” by Nvidia. However, it later re-emerged at CES 2023 and was officially released on January 5, 2023, under the name RTX 4070 Ti.
On April 13, Nvidia expanded its lineup with the addition of the more budget-friendly RTX 4070, priced at $600.
The RTX 4080 also found itself entangled in price-related discussions, as it was only $400 cheaper than the RTX 4090 flagship. This pricing led many to consider the RTX 4090 as a more attractive option, boosting its sales.
As for the desktop GPU lineup, there’s still no sign of the RTX 4060 and RTX 4050. Although details about these cards are scarce at the moment, it’s reasonable to assume that Nvidia may introduce them later in the year.
During its CES 2023 keynote, Nvidia made a significant announcement regarding the RTX 40-series mobile range. What’s noteworthy is that this lineup offers a more extensive selection of GPUs compared to the current desktop range. It’s a marked expansion from its predecessor, particularly in terms of laptop GPUs. In the previous Ampere generation, the most powerful laptop GPU available was the RTX 3080 Ti. However, the Ada Lovelace lineup takes laptop gaming performance to a whole new level with options reaching up to the RTX 4090M.
For those eager to experience high-performance gaming on laptops, you won’t have to wait too long. Powerful gaming laptops equipped with the RTX 4090M and RTX 4080M are slated for release on February 8, starting at a price point of $2,000. Later in the same month, specifically on February 22, we can anticipate the arrival of the RTX 4070M, RTX 4060M, and RTX 4050M models. These laptops will be priced starting at $1,000, with the final cost dependent on the specific card model and additional components you choose.
RTX 40-series: Architecture
Nvidia’s Ada Lovelace architecture is a notable advancement, primarily due to its reliance on TSMC’s N4 manufacturing process, characterized by a 4nm process node. This marks a significant reduction in size compared to the previous Ampere generation, which was based on an 8nm node. While not all the GPUs within this generation have been released, we know that the GPU dies span from the AD102, which represents the flagship, to the AD106, catering to budget-conscious consumers.
The ability to fit more cores onto the same GPU die is one of the significant benefits of the switch to a smaller manufacturing process. This is evident when examining the top-tier AD102 GPU in its fully unlocked state, boasting an impressive 18,432 CUDA cores. This marks a substantial increase in core count compared to the previous-generation RTX 3090 Ti, which had 10,752 CUDA cores. It’s worth noting that Nvidia has not yet released the complete AD102 GPU, suggesting the possibility of future releases, such as the RTX 4090 Ti or even a revival of the RTX Titan, both of which would harness the full potential of the AD102 GPU.
In its current state, the RTX 4090 stands as a formidable graphics card, leveraging the advantages of the 4nm process node and pushing the clock speed to a remarkable 2.52GHz. This results in a substantial generational leap, surpassing not only the RTX 3090 Ti but also the RTX 3090 in terms of performance and capabilities.
It’s not surprising that Nvidia has reintroduced Tensor cores and ray tracing cores in the RTX 40-series, and they’ve been improved compared to previous iterations. This latest generation boasts fourth-generation Tensor cores and third-generation ray tracing (RT) cores. These advancements place Nvidia at a distinct advantage, especially in the realm of ray tracing, solidifying the RTX 4090 as a powerhouse for ray tracing tasks. Even while running games at ultra settings with ray tracing enabled, at 4K resolution, it manages to maintain impressively high frame rates.
Additionally, this generation introduces DLSS 3, a technology that goes beyond pixel prediction and can predict entire frames in games. This capability can significantly boost your frames per second (fps), enhancing the gaming experience. Additionally, by rearranging GPU workloads for greater efficiency, Shader Execution Reordering (SER) is another breakthrough that helps to improve ray tracing performance and can raise frames per second.
It’s worth noting that Nvidia adopts a monolithic design for its RTX 40-series cards, in contrast to AMD’s Radeon RX 7000 cards, which utilize a multi-chip module (MCM) design. Although there were reports that Nvidia was shifting to an MCM design for its Hopper data centre GPUs, the RTX 40-series GPUs that are now on the market are built on monolithic dies.
RTX 40-series: Performance
The flagship RTX 4090 stands as the undisputed powerhouse among graphics cards on the market. It offers a significant generational leap over the RTX 3090 and RTX 3090 Ti. In our 4K gaming evaluations, the RTX 4090 outperformed the RTX 3090 Ti by an impressive 68% and left the RTX 3090 trailing behind by a staggering 89%. Even if most games don’t seem to need it, this GPU can easily handle 4K gaming at high settings and even shines in demanding games like Cyberpunk 2077.
Furthermore, the RTX 4090 shines in ray tracing performance, surpassing any previous GPU in this regard. Once again, in Cyberpunk 2077, it establishes a commanding lead over its predecessor, outpacing the RTX 3090 Ti by a substantial 71%.
But as we proceed through the product list, things seem a little less promising.The Nvidia RTX 4080, which drew controversy due to perceived overpricing from the outset, couldn’t steal the limelight from the RTX 4090. While it undeniably packs power, it’s overshadowed by the RTX 4090. Nevertheless, it remains a robust card, particularly suited for 4K gaming. Across our testing suite, we observed it to be approximately 30% slower than the RTX 4090, a notable difference considering it’s only 25% cheaper. Positively, it performs much better than its predecessors, with a 42% better performance than the RTX 3080 Ti and even a comfortable 22% advantage over the flagship RTX 3090 Ti from the previous generation.
Now, turning our attention to the RTX 4070 Ti, it’s evident that 4K gaming ambitions must take a backseat, with a focus on delivering an exceptional 1440p gaming experience. While the GPU is certainly capable of running games at 4K, Nvidia primarily positions it as a 1440p GPU, a classification confirmed by our testing results. The RTX 4070 Ti maintained a respectable 77.1 fps on average at 4K, but it remains around 20% slower than the RTX 4080 and a substantial 41% behind the RTX 4090.
At 1440p, however, it shines. It exhibits approximately 13% better performance than the last-gen RTX 3090 and a remarkable 40% improvement over its predecessor, the RTX 3070 Ti. Nevertheless, it still trails behind both of AMD’s graphics cards, with the RX 7900 XT consistently outperforming it in most scenarios, excluding ray tracing.
Nvidia’s mobile GPUs represent potent additions to the product lineup, even if they don’t quite match the raw power of their desktop counterparts. The RTX 4090 inside laptops like the Asus Scar 17 and MSI GT77 Titan ranks at the top, while the Lenovo Legion Pro 7i demonstrates that even the RTX 4080 can deliver flagship-level performance in the right laptop configuration.
DLSS 3
Another noteworthy development that sets Nvidia apart from earlier generations is the arrival of DLSS 3. Currently, this technology is exclusively available on RTX 40-series cards. Nvidia has placed considerable emphasis on this tech, highlighting its capacity to generate distinct frames rather than just individual pixels.
We conducted tests of DLSS 3 on each of the available Ada GPUs, and the results are indeed impressive. As demonstrated in the screenshot above, we ran the 3DMark Port Royal DLSS 3 benchmark on an RTX 4090 at 8K resolution with frame generation enabled. The outcome is remarkable, showcasing a remarkable 578% increase in frame rates. While real-world gaming scenarios may not deliver such extreme gains, we still observed improvements of up to 128% in certain titles.
DLSS 3, in combination with frame generation, is complemented by Nvidia’s Reflex technology, which focuses on reducing latency and minimizing rendering lag.
DLSS 3 is undoubtedly impressive, but it’s not without its growing pains. Presently, there are noticeable visual artifacts when using frame generation. While this does impact the overall experience, we’ve generally found that the performance gains can justify it. As Nvidia continues to refine DLSS 3 and expand its support to more games, we anticipate that the visual quality will improve accordingly.
Power Draw and Efficiency
A substantial increase in performance often corresponds to higher power requirements. Prior to the release of the RTX 40-series cards, there was much speculation in the rumor mill regarding the potential power draw of the RTX 4090, with some sources suggesting a total board power (TBP) as high as 600 watts. Fortunately, this did not come to pass.
The flagship RTX 4090 is capped at 450W, which is still considerable but identical to the last-gen RTX 3090 Ti. The RTX 4080 takes a slight step down, maintaining a TBP of 320 watts, while the RTX 4070 Ti reduces it further to 285W.
On paper, 450W may seem high, but in practice, the RTX 4090 rarely reaches that level. While specific tests designed to push the card to its limits may hit the 450W mark, everyday gaming usage keeps it below that threshold while delivering exceptional performance.
The other Ada cards in the lineup have shown us comparable efficiency, which is a positive sign for the whole GPU family. However, Nvidia has not been able to avoid controversy in this area. Despite the exaggerated power requirements not materializing, reports emerged after the launch of the RTX 4090 regarding instances of the card melting.
The issue appears to be related to the 16-pin 12VHPWR power connector used by the RTX 4090. For some users, after extended periods of use, this connector began to melt. As a result, affected cards were sent in for replacement, and Nvidia collected these models to investigate the problem.
Most sources indicate that this issue may have been caused by the connector not being securely seated. Given that the RTX 4090 is a large graphics card, it may not fit comfortably in many PC cases without some degree of power cable bending. Angling the connector could contribute to increased temperatures and, for some users, melting.
Although some time has passed, Nvidia has not issued a clear statement regarding the cause. Fortunately, reports of such incidents are relatively rare, suggesting that it is a limited-scale problem.
Nvidia GeForce RTX 40 series
For PC gamers, the latest Nvidia GeForce RTX 40 series cards bring a lot of excitement. Currently, these RTX cards are more accessible compared to the RTX 30 series, although there have been instances of price increases at specific retailers. Fortunately, the prices are not as outrageously high as they were in the previous generation.
In general, if you’re looking to purchase a new RTX 40 series card, you should be able to find one at its market value.