By “limiting” itself to the announcement of the RTX 3060, NVIDIA, therefore, keeps a close watch for weeks to come.
The rumors were therefore true: NVIDIA took advantage of CES in Las Vegas to formalize the latest in the RTX 3000 series range. If the manufacturer probably has a refresh – did you say “Ti”? – 3080 in its boxes, so it will be for later.
RTX 3060: 3,584 CUDA cores / 12 GB GDDR6
Initially, NVIDIA, therefore, decided to lift the veil on the fourth version of the Ampere generation intended for desktop computers. After the RTX 3080, RTX 3090, RTX 3070, and RTX 3060 Ti at the very end of last year, NVIDIA will therefore launch the RTX 3060 by the end of February.
From the outset, we will ask ourselves the question of the actual availability of the cards on this date, but we will note that NVIDIA still has more or less six weeks to try to build up stocks with its partners who should all respond to the call. The official price is $ 329, with the company not yet providing a price in euros.
NVIDIA focused on the 3,584 CUDA cores backed by 12 GB of GDDR6 memory. Regulars have noticed that this is 50% more than on the RTX 3060 Ti, but note that the memory interface is only 192-bit while it is 256-bit on the Ti. Note also that we have no precision on the frequency of this memory.
On the other hand, we know that the GPU sees its frequencies evolve. Basically, we are at 1.41 GHz and 1.32 GHz respectively on the RTX 3060 Ti and on the RTX 3060. The latter however takes the advantage in boost frequency with 1.78 GHz when the RTX 3060 Ti cannot reach “That” 1.67 GHz… at least on standard models.
During its conference, NVIDIA logically over-sold its new card by emphasizing in particular the gains brought about by the activation of DLSS. It would become possible to play with RTX effects in 60 fps at Watch Dogs Legions or Cyberpunk 2077. These are probably averages, but we are waiting to be able to verify this.
Already 3 RTX30 “lapstop series”
The second part of NVIDIA’s CES Las Vegas 2021 conference focused on another part of the PC video game industry: the laptop. More and more gamers are playing their favorite hobby on a laptop, and we were all waiting to see what the Ampere architecture could do with consumer planed versions.
Not surprisingly, NVIDIA presented a range available in three versions: there is, therefore, the RTX 3060 Laptop, the RTX 3070 Laptop, and the RTX 3080 Laptop. Each time, it is of course a question of integrating the Max-Q technology designed – and quite effective – to optimize all the subsystems of the machine.
The RTX 3060 Laptop will power laptops starting at $ 999. Its 3,840 CUDA cores should allow it to exceed the performance obtained on laptops equipped with RTX 2080 Super. The frequencies go from 1,283 MHz to 1,703 MHz ( boost ) and it is a question of integrating a maximum of 6 GB of GDDR6 to reach 90 fps on all the latest games in 1080p, details at the maximum.
A cut above, the RTX 3070 Laptop is – like its desktop counterpart – intended for 1440p gaming. NVIDIA talks about 5,120 CUDA cores, operating frequencies between 1290 and 1620 MHz as well as a total of 8 GB of GDDR6. We’re still talking about 90 fps with maximum detail, but at 1440p and NVIDIA is talking about 50% gains over the RTX 2070 Mobile. Laptops and equipment should start at $ 1,299.
Finally, the highlight of the show, the RTX 3080 Laptop is not however targeting the game in 4K. NVIDIA prefers to look at over 100 fps at 1440p with maximum detail. The first gaming laptops with such a card should be billed $ 1,999 knowing that they can come with 8 or 16 GB of video memory to support a GPU equipped with 6,144 CUDA cores with frequencies between 1,245 and 1,710 MHz.
Let us conclude by pointing out that during this conference, NVIDIA also emphasized techniques such as Dynamic Boost 2.0 and Whisper Mode 2.0 before insisting on the Resizable BAR: basically, at NVIDIA too, the CPU access to all GPU memory at once for better performance.
Source: NVIDIA conference, press releases