Nvidia Geforce RTX 3080 (Ti) from Q3 2020?
This is rumored to be the current status for the Ampere flagship for gamers – see the update from April 30th, 2020, for more information see the latest news.
Status May 15, 2020: Yesterday, Thursday afternoon, Nvidia officially introduced its new GPU generation. Shortly after AMD presented the Radeon VII Pro (see next page, May 14th), the A100, the Ampere-Flagship (or rather the Ampere-Mother Ship) for professional applications, will be born.
After 12nm on the predecessor Volta, Nvidia, like AMD, also relies on 7-nanometer production. Nvidia bundles 54 millilarden transistors on a chip area of massive 826 mm².
108 streaming multiprocessors with 64 FP32, 32 FP64 and 4 Tensor cores each get support for up to 40 GB HBM2 memory with a bus of 1.6 TB per second.
The predecessor Tesla V100 based on Volta combines “only” 21.1 billion transistors on almost the same area of 815 mm². For the relevant performance data (FP32, FP64, Tensor cores), Nvidia gives TFLOPS data that beats its predecessor by around 50 percent.
Nvidia sells amps in a deep learning server called DGX A100, which combines eight amp GPUs.
Two 64-bit Rome processors from AMD, 1TB RAM and 15TB SSD memory justify a price of $199,000.
The National Argonne Laboratory of the US Department of Energy has purchased the first DGX A100 – and is using it for research on Covid-19.
For the common player the card is of course nothing. But it is the basis for the upcoming gaming graphics cards. That makes you want to play the next months. We will keep you up to date.
Status 12. May 2020: A recent YouTube video (following the current info update) promises a big leak that 3dcenter.org exposes as a summary of known rumors from several sources.
The promised specs (see March 13th) correspond to the circulated technical data of the base chip (amps).
Based on earlier publications, details such as the number of shader clusters or memory bandwidth are estimated. For more information, see the linked article and the following video.
“Leak” to the Geforce 3080 Ti
As of April 30, 2020: wccftech.com came across an article in the China Times, which was taken offline shortly afterwards.
According to this article, Ampere for gamers will appear in the third quarter of 2020 (September/October).
Board partners are in preparation to clear their stocks with Turing cards. Soon there will be sales and bargains for the current GPU generation to make room for the successor.
ASUS, Gigabyte and MSI are mentioned.
The beginning should be Geforce RTX 3080 Ti, 3080 and 3070 – possibly with a staggered, quarterly release starting in Q3.
The non-RTX cards of the Geforce GTX 16xx series should remain for the time being. Therefore, an RTX 3060 could be correspondingly longer to wait for.
For comparison: Almost one and a half months passed between the start dates of Geforce 1080/70 and the midrange Geforce GTX 1060; almost three months between Geforce RTX 2080/70 and the RTX 2060.
Status April 27th, 2020: On May 14th, Nvidia will probably hold an online event where everything revolves around amps. That’s the name for the next generation after Pascal (Geforce 1xxx) and Turing (Geforce 2xxx) – so the base for the upcoming “Geforce RTX 3xxx” cards.
This can be concluded from a current Nvidia press release, which promises new technologies around AI, deep learning, automotive driving, robotics and professional graphics applications and introduces it with the words “Get Amped”. “Amped” in this context means: “Get ready!”
The event will be available on the YouTube channel of Nvidia. It will probably not be about gaming graphics cards, but upcoming features that gamers can look forward to with Nvidia’s new generation.
As of April 8, 2020: Nvidia and board partners respectively are using GDDR6 memory for Geforce 1650 (Ti) and 1660 (Ti).
GDDR5 chips were used before. The larger RTX and RTX super series already have the faster memory. In terms of price, the newer variants are minimally more expensive due to the newer technology after a first look.
We observe this at least for gigabyte cards at Alternate.
The new memory primarily brings a lower energy consumption and of course more data throughput.
14 to 16 instead of 8 to 12 GB can be transferred per second and processed by the GPU. This can have a positive effect on games with many additional details.
Fewer bottlenecks then mean for example fewer pop-ups in the game world or display errors.
Status April 2, 2020: The rumor has come true.
Nvidia has announced the RTX Super series for notebooks. In fact an RTX 2060 Super is missing.
The models come as standard and with a Max-Q design, which means a lower clock rate and less power consumption compared to the Max-P counterparts.
With “Dynamic Boost” there is an intelligent balancing between CPU and GPU. With Advanced Optimus, notebooks get an improved switching mechanism for a dedicated graphics card and the often additionally integrated graphics chips from Intel or AMD.
In addition, GDDR6 graphics memory with a low voltage is also included.
From April 15th, devices for the gaming sector will come with the following chips (see table).
Besides Ryzen 4000 laptops, Comet Lake H from Intel will also play a role. These are the new chips from AMD and Intel for 2020.
While AMD is bringing a new platform, Intel is bringing a high-end extension to the already available U series. Devices can be found on Thursday afternoon (15:10) for example from 2,099 Dollar. In a current news we show more notebooks for 2020 with new Intel and Nvidia chips.
|RTX 2080 SUPER||1560 Mhz||8GB GDDR6||150W+|
|RTX 2080 SUPER (Max-Q)||1080 Mhz||8GB GDDR6||80W|
|RTX 2070 SUPER||1380 Mhz||8GB GDDR6||115W|
|RTX 2070 SUPER (Max-Q)||1150 Mhz||8GB GDDR6||80W|
As of March 26, 2020: Parallel to Intel’s Comet Lake H (Intel’s 10th generation for mobile high-end devices), Nvidia’s super cards for notebooks will be launched.
According to the Chinese website ITHome (machine translation), April 2 will mark the launch.
The following cards should be available, there are several uncertainties in several places in the question of whether an RTX 2060 Super for notebooks will be included. The following
Status 17 March 2020: The corona virus causes Nvidia to set other priorities.
The GTC news can wait for now. Now is the right time to take care of family, friends and the community. Researchers and developers will later have the opportunity to share their achievements in streams.
Nvidia does not want to expose employees, partners, media and analysts to unnecessary risks because of the virus.
As of March 13, 2020: There are current rumors about the Geforce RTX 3000 series.
These come from Twitter, among others, and make you sit up and take notice. The Geforce RTX 3080 Ti is not mentioned directly, but is indicated by the code name GA102.
In addition, there are other codenames with specs that correspond to cards up to RTX 3050. In addition there are also performance data in comparison to the direct predecessor – the RTX 2080 Ti.
Samsung’s 10 nanometer process would thus be used with RTX 3000 instead of 7nm as with AMD.
The flagship is supposed to offer up to 40 percent more performance than the RTX 2080 Ti, which fits to rumors from early March (see below). SLI will only be available for that high-end card, RTX for the entire amp portfolio.
Geforce RTX 3080 Ti: Alleged Specs
Surprisingly, the next gen chip will be based on Samsung 10nm node 2 Not very solid about what GA102 really got, may struggle to be 40% above 2080ti, but bound to be under that number if compared against full bore TU102 3 SLi will be available to GA102 ONLY 4 RTX for everyone pic.twitter.com/b5YAKmOmgf
- KittyCorgi (@CorgiKitty) March 11, 2020
The Geforce RTX 3080 Ti’s VRAM should increase to 12 GB (11 GB in the predecessor). An RTX 3080 (GA103) would increase to 10 GB (8 GB with the RTX 2080), for the RTX 3070 (GA104) 8 GB would be available as with the predecessor. It is remarkable that the RTX 3070 is supposed to come on 95 percent of the achievement of a RTX 2080 Ti. With 6 or 4 GB would come 3060 (GA106) and probably 3050 (GA107).
In addition, Twitter user KittyCorgi states that the next generation of Geforce maps will not be released until the end of 2020 – planned in response to Big Navi. The information at this point is only rumors.
We will continue to report. You can find out more at wccftech.com.
Status 10. March 2020: After the cancellation for the GTC with the promise of a live stream, this one has to fall victim to Covid-19 as well. Nvidia will announce the upcoming news classically via press release.
This will follow on 24 March. Nvidia CEO Jensen Huang will then only be available for investors – in a telephone conference. Other planned replacement online services of GTC are to remain in place: Webinars, recorded speeches and panels, as well as research, training and other sessions.
From GTC we expect to receive first information about Ampere, Nvidia’s base for the next workstation and gaming graphics cards.
Of course, the upcoming GPUs of the Geforce RTX 3000 series are primarily of interest.
As of March 3, 2020: Nvidia has declared the GTC (GPU Technology Conference) a pure online event.
The reason is the risk of the new corona virus (Covid-19), which has already caused the Mobile World Congress, the Facebook conference F8, the Games Developer Conference and more events in 2020.
The GTC was supposed to take place in San Jose from 22nd to 26th March.
In a live stream at another, as yet unknown time, Nvidia boss Jensen Huang will present new products. In addition to workstation applications and cloud technologies, the focus will probably be on Ampere, Nvidia’s new GPU platform.
This will form the basis for the RTX-3000 series and new professional graphics cards.
Benchmark results have possibly been found for the latter in the past few days.
VRAM values of 24 or even 48 GB, among others, suggest this. According to the scores, the cards have beaten the current high-end chips RTX 2080 Ti and Titan RTX by about 40 percent. That lets further hope for the summer.
Status 18. February 2020: Since the beginning of the year it remains quiet around Nvidia.
For the coming action role-playing game Cyberpunk 2077, the manufacturer has at least presented the right graphics card: the Nvidia Geforce RTX 2080 Ti Cyberpunk 2077 Edition.
It is a special version of Nvidia’s currently fastest gaming graphics card. It is not intended for sale, it is only available in a limited edition of 200 pieces. 77 of them are raffled off by Nvidia in a raffle.
Details can be found in the linked announcement. The following video shows the design of the new card with an aluminum cooler in neon yellow cyberpunk design and matching lettering.
Nvidia RTX 3000 Amp Release & Specs – All info, leaks and rumors
What about Nvidia’s upcoming high-end graphics cards? We are gathering all relevant information on possible specifications, performance and release.
Due to the Corona crisis, it’s been quiet lately about Nvidia’s upcoming graphics card generation Geforce RTX 3000 for gamers. Important trade fairs are cancelled, others like Computex have already been postponed.
However, especially in the run-up to the GPU Technology Conference (GTC), which was actually scheduled for March, a lot has been reported about Nvidia’s next-gen graphics cards:
The Ampere GPUs will be faster than their predecessors, that much is self-evident, and the Californians also continue to rely on real-time ray tracing. How much faster the latest generation will be, however, is another matter.
In the following, we’ll summarize what there is to know so far about Nvidia’s Geforce RTX 3000 amps, what rumors and speculations are floating around and when they might appear.
Last update: 13.5.2020
Newly added: Current leaks to the VRAM of the RTX 3080 Ti and to the raytracing performance in the sections “The possible specifications” and “How much performance can be expected from RTX 3000”.
The possible specifications
Production in a 7-nanometer process is considered to be assured. This is to be carried out both by our long-standing partner, contract manufacturer TSMC, and by Samsung.
It is assumed that 128 SMs (streaming multiprocessors), each with 64 CUDA computing cores, will be used for the full expansion of the ampere chip (GA100) for the professional segment.
This results in an impressive 8,192 computing cores, which are said to be linked to up to 48 GB of HBM2e video memory.
Variants of the chip for player graphics cards:
- GA102 (RTX 3080 Ti) – 84 SMs / 5,376 CUDA cores / 12/24 GByte GDDR6 384 bits (possibly 18 Gbps with a bandwidth of 864 GB/s)
- GA103 (RTX 3080) – 60 SMs / 3,840 CUDA cores / 10/20 GByte GDDR6 320 Bit
- GA104 (RTX 3070) – 48 SMs / 3,072 CUDA cores / 8/16 GByte GDDR6 256 Bit
- In addition, the clock rate should be increased by 200 to 300 MHz compared to Turing (RTX 2000, GTX 1600) – we are talking about a 2.2 GHz turbo clock.
In this context, Igor Wallossek from igor’sLAB brought an interesting detail to light. He claims to have found out that Nvidia is fine-tuning the boards in close cooperation with its board partners.
Put simply, Nvidia is trying to improve the signal path for data transmission, which ultimately enables higher clock and throughput rates.
How much performance can be expected from RTX 3000?
The rumor mill provides some hints on this, which, as usual, should be taken with a certain amount of caution. The direction is however clear, RTX 3000 is to become correctly fast and put down a clear performance jump measured at RTX 2000-Turing.
Thus performance increases of 40 per cent are assumed in relation to the Titan RTX and the RTX 2080 Ti. Also from an increase of 70 per cent already the speech was.
Both statements refer to OpenCL scores from Geekbench 5.
The measurement results at a glance:
|Nvidia Graphics 3*||222.377 Punkte||7.936 Shader||32 GB|
|Nvidia Graphics 2*||184.096 Punkte||7.552 Shader||24 GB|
|Nvidia Graphics 1*||141.654 Punkte||6.912 Shader||47 GB|
|Tesla V100||154.606 Punkte||5.120 Shader||32 GB|
|Titan RTX||131.974 Punkte||4.608 Shader||24 GB|
|RTX 2080 Ti||129.019 Punkte||4.352 Shader||11 GB|
Ray tracing should also experience a significant increase in performance. A significant increase in the number of RT cores (RT stands for raytracing) is assumed here.
Exact details are not available, but this is generally considered to be certain.
A current leak (as of 13.5.2020) assumes, however, that the number of RT cores does not necessarily have to increase at all, since the performance of the individual cores is significantly increased – the so-called intersections are said to be processed up to four times faster.
When can a release be expected?
Currently, a release in autumn of this year is expected. The first teasers are expected in June or July, and many believe that a product launch is likely at Computex, which has been postponed to September 28 to 30.
A current rumour from China supports a possible release in autumn. As a result, the large Taiwanese hardware manufacturers Asus, Gigabyte and MSI are lowering the prices for RTX-2000 graphics cards in order to clear the stocks for the upcoming new high-end models from the RTX-3000 series.
However, there are also voices that assume that Nvidia first wants to wait for AMD’s next steps in the GPU segment in the form of Big Navi before releasing their own new graphics cards.
By the way, we can expect the first more tangible, new details on the amp architecture in a few weeks.
Nvidia will catch up on the keynote for the GTC via YouTube on May 14th at 3 pm Central European Summer Time.
It’s not to be assumed that there will be concrete hints about player GPUs, but the information from the professional graphics card sector should allow conclusions about the RTX 3000 models.
Nvidia Geforce RTX 3080 Ti Ampere coming in 09/2020?
RTX 3080 Ti GA102 chip
New information to the RTX 3080 Ti appears in the Internet! The 3080 Ti could have 84 shader clusters, which results in 5376 shader units!
More likely it should be that it could have 80 shader clusters with 5120 shader units!
What is very interesting is that the GA102 chip will be manufactured in 7 nm
- 2.2 – 2.5 GHz
- 21 teraflops
- GDDR6 VRAM (video memory)
- 18 Gbps
- 863 Gigabyte per second
- Up to 35 percent more power possible compared to the RTX 2080 Ti
- RTX 3080 Ti Release date possibly 09/2020
As they say on YouTube and the Internet… all rumors and that’s where it’s going to stay until it’s officially confirmed by NVIDIA! But it remains exciting!
Corona and Nvidia Ampere RTX 3080 still 2020?
Corona… everywhere people are getting sick and working from home, which of course also influences the further development of new techniques and graphics cards!
If there is really a chance that there will be new graphics cards from AMD and Nvidia this year, we will all see!
I will definitely keep you up to date with news about new graphics cards.
New details about the Ampere RTX 3080 Ti have been released!
New details of a still unknown graphics card have appeared on the net. What is now becoming more and more important is not an increase in performance of 75% but rather of 40-50% more performance compared to the current RTX 2080 Ti Turing graphics card.
As you can see in the table, the Nvidia Graphics 1 could become the new RTX 3080 Ti and would have a total of 24GB memory.
Because of the 3200 Cuda cores, which it has more compared to the RTX 2080 Ti, it could finally also bring up to 40-50% more performance!
It is also possible that it will only have 20GB RAM, which is sufficient for gaming at the moment! It will probably not be released in March 2020 but rather in summer or autumn 2020.
Here the newest details to the RTX 3080 in the table:
|GPU||GEEKBENCH 5||CUDA||VIDEO MEMORY|
|Nvidia Graphics 1||184.096||7,552||24 GB|
|Nvidia Graphics 2||141,654||6,912||47 GB|
|Titan RTX||131,974||4,608||24 GB GDDR6|
|RTX 2080 Ti||129,019||4,352||11 GB GDDR6|
Nvidia Ampere already in March?
And again new details come to light, so it is said now already that Nvidia with the new RTX 3080 / RTX 3080 Ti / Amperer could possibly come already in March 2020 into the trade!
Also the rumors about the increased performance of the graphics card, is supposed to move towards 75%, at least in raytracing, where more performance is really needed.
The power consumption should be kept within limits with about 300 watts of power consumption!
In March 2020 the GTC (GPU Technology Conference) from 22.03 – 26.03.2020 in San José, where a series of new product introductions is planned and where there could be talk of a new Nvidia Ampere graphics card aka RTX 3080.
When the release of the new Nvidia Ampere 3080 Ti will be, is not certain! The current RTX 2080 graphics cards are very popular and sell well and AMD is currently no competition for Nvidia in terms of graphics cards!
Therefore, they have no stress to bring a new Ampere graphics card in the market as soon as possible! Let’s wait and see what will happen at the GTC and what will really be presented!
The new Ampere graphics cards called RTX 3080 (if that’s what they’ll be called!) are said to have 75 percent more power than the previous Turing RTX 2080 graphics cards, according to an Internet rumor mill!
THIS IS VERY UNLIKELY! The GTX 1080 Ti has about 25% less performance than the current RTX 2080 Ti. So in my opinion 25-30% more performance is realistic!
Also the production of the new Ampere GPU, should be manufactured in 12-nm and not, as AMD has already done in its new graphics cards, with 7-nm.
Current status (February 2020) and possible prospects, but these are not confirmed:
New name: RTX 3060, RTX 3070, RTX 3080, RTX 3080 Ti
Instead of RTX 3080 Ti I could also imagine that it is called NVIDIA Ampere RTX! (as the NVIDIA TITAN RTX was also called)
7 nanometer EUV (Extreme Ultraviolet) manufacturing at TSMC and Samsung
completely new architecture (Ampere)
Focus on raytracing, where the increased performance can also be noticed later!
First alleged specifications for the new Nvidia RTX 3080 and RTX 3070 appeared on the Internet! PCGH has new data about the new Nvidia RTX 3080 graphics cards.
According to this, the RTX 3080 should work with 64 CUDA cores, 3,840 shader cores and a memory interface connected with 320 bits.
(Interface memory capacity of 10 or 20 Gigabyte GDDR6). 10-20 GB would thus be 2-12 GB more performance compared to the RTX 2080!
And the RTX 3070 accordingly works with 48 CUDA cores, 3,072 shader cores and a 256 bit wide memory interface. 8 or 16 Gigabyte video memory would be possible. With 16 GB video memory, the video memory would increase by 100% compared to the current RTX 2070!
Here the question arises nevertheless, if the RTX 3070 / 3080 already have 16-20GB video memory, what should a possible RTX 3080 Ti bring with itself then? RTX 3080 Ti with 24 GB or even 32 GB video memory? We will see!
|RTX 3080||GA103||3.840||10/20 GB GDDR6||320 Bit|
|RTX 3070||GA104||3.072||8/16 GB GDDR6||256 Bit|
|RTX 2080 Ti||TU102||4.352||11 GB GDDR6||352 bit|
Like every year, there is always a new graphics card series available at Nvidia! At the latest in June 2020 is planned to bring the RTX 3080 Ti, as well as the RTX 3080 into the trade!
Who thus at present, over it thinks itself a new RTX diagram map to buy, which should still wait, if it can!
Clearly, if the new Nvidia RTX 3080 graphics cards are to be bought, then the prices of the older RTX 2080 and RTX 2080 Ti models will also fall! Just who is also ready to buy a used graphics card, will find many of them in the classified ads market, because owners of the RTX 2080 will probably buy the successor models 3080 now!
This is the perfect time to buy the high performance graphics of the RTX 2080 Ti series! Usually the prices for new RTX 2080 graphics cards in the trade fall around 30%, so you can buy a now current 1300 Dollar RTX 2080 Ti, then for approximately 899 Dollar new. Used the prices fall then also gladly around the 50%, which then 650 Dollar would be! This could be seen also well with the GTX 1080 Ti models, when the new 2080 models were for sale!
RTX 3080 Ti and RTX 3080 7-nanometer method
Nvidia is currently still keeping a low profile with information about the new 3080 series! The new graphics card generation will be equipped with ampere chips, which are manufactured in a 7-nanometer process.
Compared to the current 12 nm architecture of the RTX 2080 series, this could lead to a great performance leap.
As far as is known on the Internet, the new graphics cards will not have more GPU memory than the current 2080 Ti and 2080 graphics cards.
That means that the RTX 3080 will also have 8GB GDDR6 and the 3080 Ti 11GB GDDR6 memory. The GDDR6 memory works with 8000 MHz.
Beautiful it would be however, if Nvidia would equip surprisingly way, the RTX 3080 with 11GB and the RTX 3080 Ti with 16GB memory!
Which will probably not be the case, I assume that here only the GPU clock of 1350 MHz, around and at 1550-1750 MHz and the boost clock in OC mode will be around 2000 MHz.
A performance increase of the new RTX 3080 Ti in relation to the current RTX 2080 Ti will be according to my assumption between 15% and 25%, depending upon manufacturer differently!
ASUS will probably use the fastest clock rates again and the model will be called ASUS GeForce RTX 3080 Ti ROG MATRIX GAMING.
I guess the cheapest RTX 3080 Ti graphics cards will cost 1450 Dollar, for example from Zotac (ZOTAC GeForce RTX 3080 Ti) and the top 3080 Ti graphics cards around 2000 Dollar with water cooling and lots of RGB lighting.