After almost two years of carrying the ray Tracing to the mainstream with their cards , Turing, Nvidia has been working hard on the successor of its popular flagship RTX 2080. Series Nvidia GeForce RTX 3080 will be the new architecture Ampere who made his debut in the supercomputers DGX A100 from Nvidia, marking the first time the company used a manufacturing process of a 7 nm for your GPU.
Like Turing before, it is expected that the card RTX 3000 technology Ampere provides a graphics performance even better for gamers and workstations, but the update is most notable for the enthusiasts of the computer will be the ability of Ray tracing without bottlenecks. Here is everything we know about it until now.
Nvidia has remained silent on the launch of their graphics cards RTX 3080, and the current pandemic is adding additional uncertainty. Without an official announcement or guidance from Nvidia, the industry observers speculate that the GPU RTX 3000 could be launched in the fall of 2020 (northern hemisphere, September-November).
Historically, Nvidia cut the launch of a new family of GPU consumption, and we hope that again be the case this time. This means that the card and badge of the first level, such as the RTX-3080 and the RTX 3080 You, that could be renamed under the line RTX Titan, debutarían first, with more economic options, including the RTX 3070 and 3060, arriving at a later date that may be extended until the beginning of 2021.
TweakTown reported that the RTX 3080 entered production and could be an announcement as soon as August, with the possibility that the card is available for purchase in September. This timeline would coincide with what Nvidia used to launch RTX 2080, which was announced in August, two years before.
The new line of RTX Nvidia represents the premium segment of graphics cards for gaming and workstations, and it is expected that the series RTX-3000 has a higher price. As a reference, the RTX-2080 Ti Nvidia sells for $1,199 dollars, and the RTX-2080 Super costs $699 dollars. The RTX 3080 will probably have a price in that range on the higher end of the spectrum.
In the past, there was a big jump in price between the generations of Nvidia graphics: card Turing were more expensive than the Pascal that they replaced, and this could be the case again.
There are rumors that Nvidia may be moving the manufacturing from TSMC to Samsung. This movement will give Nvidia access to the process extreme ultraviolet or EUV of 7 nm of Samsung, which would reduce costs. However, the accuracy of that rumor is not clear.
Players without a budget large enough for an upgrade should not be discouraged. The price of the RTX 2060 currently starts at $299 dollars, and the RTX-3060 up to date you will probably have a similar price. The RTX 2070 mid-range has a starting price of $499.
These cards lower price will help Nvidia compete against the next line of graphics RDNA 2 AMD. In addition, a point of reference leaked earlier this year showed that a GPU-Radeon unknown exceeded the RTX-2080 Ti Nvidia by 17 per cent. The performance more competitive than your opponent you can force Nvidia to participate in a price war with the launch of the RTX 3080.
The architecture Ampere Nvidia will appear in multiple versions of graphics cards. The first card on debut, was presented in may at the Conference of computer Graphics Technology line of Nvidia in the form of A100 for data Centers and Supercomputer DGX A100. The company rated Ampere as the chip of 7 nm for the largest ever built, and was designed for a better performance of artificial intelligence.
The series RTX-3000 will probably use the same architecture of 7 nm that debuted in the version for the data Center, which would provide a better energy efficiency. A previous report from Taipei Times suggested that this new architecture would result in an increase of up to 50 percent, in GPU performance, and at the same time reduce energy consumption to half in comparison with Turing.
The YouTube channel, Moore”s Law is Dead, reported that the RTX 3080 should put Nvidia at the head of the performance once more, with a capacity of 21 teraflops, or nine teraflops more of what is possible in the Xbox Series X. When it comes to gaming, the Nvidia graphics cards of the series RTX-3000 will show that computer games still have a performance advantage over the consoles.
Expected different consumer versions of Ampere’s, although there are some conflicts in your numbering scheme due to leaks early. It is widely believed that the GA102 is the version used for the RTX-3090, which debut under the line RTX Titan, the RTX-3080 Ti or the Super version and the RTX 3080 standard. The version Titan could get up to 24 GB of memory, while the card improved line It or Super will come with 12 GB of memory, and the 3080 standard will be sent to you with 10 GB of RAM.
However, a leak is separated from the user of Twitter KittyCorgi shows the series 3080 with the model number GA103. Regardless of how they are labeled, the performance of these cards is not discussed.
Hardwareleaks published a leak about a test of 18,257 points in 3DMark Time Spy of what is believed to be obtained with the RTX 3080 You. If true, this represents a performance improvement of 30 per cent on the RTX 2080 You.
By way of comparison, the reference points are filtered above for the GPU Big Navi from AMD, which uses the architecture RDNA 2 of the company, suggested that the card works only 10 to 15 percent better than the RTX-2080 Ti Nvidia.
Although, it is expected an improvement in generation performance to a new family of GPU, it is possible that the RTX 3080 It has significantly better performance in ray Tracing, the distinctive feature of the line RTX and an area of struggle for the series RTX 2000 current.
The improvements of Ampere with artificial intelligence and DLSS should help to improve the performance of Ray tracing, especially when the games are rendered at higher resolutions. In addition, it is expected that the card comes with 108 cores for acceleration of ray Tracing, more than double that of the 48 are supported by the RTX-2080 Super.
Source: Tweak Town
Along with the novel design of the card is double-sided for a pc, it is rumored that Nvidia uses a new coprocessor cross to help improve the performance of ray Tracing. This coprocessor would be the chip in charge of Ray tracing to the GPU, according to Tweak Town, and may serve to attach nuclei RT in a same chip.
It is not clear how this will develop this architecture, and if Nvidia will download all or some of the claims of ray Tracing to this coprocessor cross, but gamers should be able to expect a much better performance in games with Ray tracing enabled, especially in settings gamng enhanced in high resolution.
CyberPunkCat on Twitter
A spec sheet leaked to the RTX-3080 Ti published by the Twitter user CyberCatPunk revealed that the card will come with 5,376 cores CUDA, and support a memory GDDR6 from 12 GB to 18 bps. The TDP (design Power thermal, for its acronym in English), for the card is set to 320 W.
It is reported that the clock speeds of around 2.2 GHz, and we have heard rumors in the past that indicate that the card will come with 18 GB of video memory. A leak recently suggested an increase of the clock speed of only 1,935 MHz, and it is not clear at this time if this will change or if Nvidia has made sufficient improvements in efficiency that do not require a higher speed.
It says that the card is PCIe 4.0, and will be provided with three ports, DisplayPort 2.0, a HDMI 2.1 and a USB port-C.
However, for gamers, the feature killer of the GPU based on Ampere is that it will be compatible with the specifications DLSS 3.0 next generation for any game that is rendered using techniques TAA or temporal anti-aliasing.
The implementation of the compatibility with DLSS 3.0 still requires drivers Game Ready, but will still be easier for developers to incorporate DLSS to your game with some code work involved.
For gamers, this will result in a greater performance in games, and it is rumored that Nvidia enables DLSS 3.0 out of default, according to the YouTube channel of Moores Law is Dead.
In general terms, the RTX 3080 could debut with a new and controversial. The leaked images showed a very different system of cooling for the use of a double fan on the card.
Instead of having both fans on the same side, are placed on opposite sides. Although, the fins that wrap around the card will probably serve to dissipate heat, the on-line readers noted that this gives the series RTX 3000 an appearance of a heat sink.
The leaked images more recently reveal that Nvidia may be moving towards a new system of power interface PCIe 12 pin to the graphics card high-end that is capable of delivering up to 600 W of power.
The first rumors suggest that the card flagship from Nvidia will require at least 400 W. The new interface of 12-pin would be used in Nvidia GPUS with plate numbers PG142, that it is speculated would be for the 3000 series of high-end, including the RTX-3080 Ti/RTX 3090, RTX 3080, RTX 3070 Ti/RTX 3070, according to Wccftech.
A scheme on the site technological chinese FCPOWERUP shows the power interface, which seems to be a single connector with two connectors stacked six pin states. This leads some in the tech community to speculate that the gamers who do not want to pay more for the connector pin 12 can still join two connectors are six-pin.
However, this solution would require an additional cable, which will not only give the inside of your cabinet look messy, but that could present new challenges thermal in regard to the flow of air, especially given the requirements of most power out of the card.
Apparently, the optimal solution would be to also upgrade their power supplies to accommodate the new connector and meet the requirements of higher power of the cards flagship of Nvidia.
This cost is added to the to update a computer’s existing. The publication also detailed a interface separate four-pin adjacent to the new power connector 12-pin, although the purpose of this connector is unknown at this time.
Fortunately, if the rumors of a launch of fall are finalized, we will not have much time to learn more about the upcoming great event of Nvidia.
*Updated by José Luis Plascencia on July 24, 2020.
Recommendations of the editor