قالب وردپرس درنا توس
Home / Technology / The NVIDIA GeForce RTX 2080 Super Review: Memories of the Future

The NVIDIA GeForce RTX 2080 Super Review: Memories of the Future

In the past two months, there has been a lot of activity in the PC graphics card industry. Both AMD and NVIDIA went back and forth and made announcements and product launches as part of a wider transformation of mid and high end graphics card markets in response to the launch of AMD's Radeon RX 5700 graphics card series. Performance is on the rise, while last-minute maneuvers have lowered prices slightly, while the two main players in the GPU sector are trying to consolidate their product offerings and secure their competitive edge over the competition.

For all practical purposes, NVIDIAs The GeForce RTX 20 Super Series is the company's answer to AMD's new graphics cards. The Super Cards are a combination of a mid-range kicker and a rebalancing of the product stack associated with AMD. They have given NVIDIA's product range a modest performance increase in almost every way. With NVIDIA launching the first two cards in this sub-series with the GeForce RTX 2070 Super and the GeForce RTX 2060 Super earlier this month, NVIDIA is back again to release the last (?) Of Super Cards: the previously announced GeForce RTX 2080 Super.

Since gamers had a few weeks to chew on this one ̵

1; the specifications and all – there are no last-minute surprises here. The GeForce RTX 2080 Super has a small performance penalty for the GeForce RTX 2080. It offers a few more SMs and higher clock speeds to increase performance a little while the price is $ 699. To be honest, it's not a map that sets the world on fire or sets new records. But it's a card that will deliver better performance than the RTX 2080 at this time of the day, and it will restore some latitude between the NVIDIA 2070 and 2080 cards, as the new RTX 2070 Super suits the vanilla RTX 2080 as well as has made superfluous.

NVIDIA GeForce RTX 20 Series
Cards Price
GeForce RTX 2080 Ti RRP: 999 USD
[Price: 1149 USD]
GeForce RTX 2080 Super [1965900UVP:699USD
GeForce RTX 2080 EOL (699 USD)
GeForce RTX 2070 Super MSRP: 499 USD
GeForce RTX 2070 EOL (499 USD) [1965900819659008] RRP: 399 Dollar
GeForce RTX 2060 MSRP: $ 349

Not surprisingly, the vanilla RTX 2080 is on its way out. As part of the introduction of the super card, the models RTX 2080 and RTX 2070 are set. So for all practical purposes, the RTX 2080 Super is the new RTX 2080. That is, the fastest of NVIDIA's half-affordable high-end graphics cards, the surname of the 80-card family name, and a portion aimed at a mix of 1440p and 4K gaming.

And, as we'll see, it's the least and most interesting introduction to the super card.

NVIDIA GeForce specification comparison
RTX 2080 Super RTX 2080 Ti RTX 2080 RTX 2070 Super
CUDA cores 3072 4352 2944 2560
ROPs 64 88 19659031] 64
Core Clock 1650 MHz 1350 MHz 1515 MHz 1605 MHz
Boost Clock 1815 MHz 1545 MHz [196590311710 15.5 Gbps GDDR6 14 Gbps GDDR6 14 Gbps GDDR6
Memory Bus Width 256-bit 352-bit 659031] 256-bit
VRAM 8 GB 11 GB 8 GB 8 GB
Single Precision Perf. 19659008] 11.2 TFLOPS 13.4 TFLOPs 10.1 TFL 9.1 TFLOPS
TDP 250 W 250 W 215 W 215 W
GPU TU104 TU102 TU104 TU109004 13.6B 18.6B [19659027] 13.6B 13.6B
Architecture Turing Turing Turing
Manufacturing Process TSMC 12nm "FF TSMC 12nm" FFN " TSMC 12nm" FFN " TSMC 12nm" FFN "
Start date 23.07.2013 27.09.2008 09/20/2018 07/09/2019
Introductory price 699 $ $ 999 $ 699 $ 499

Let's look at the technical data from above. When NVIDIA launched the GeForce RTX 20 Series cards last year, the 2080 and 2080 Ti models were somewhat conservatively configured. Both cards were shipped (among others) with some disabled SMs, meaning they used less than fully enabled GPUs. This of course helps with binning, but also gives NVIDIA a little air to increase performance, if ever needed. That's exactly what they do with the RTX 2080 Super.

For the first time in a GeForce RTX 20 card, we finally get a fully functional TU104 GPU. NVIDIA enables the last 2 SMs and adds another 128 CUDA cores to the pool for a total of 3072 CUDA cores. With the last two SMs, the corresponding PolyMorph engines and RT cores are shipped, so the RTX 2080 Super can come up with everything the GPU has to offer in the areas of shading, texturing, geometry and raytracing. The notable exception here is the render backend. RTX 2080 ships with all enabled 64 ROPs and the associated L2 cache. Therefore, there are no hardware improvements here.

Instead, higher clock rates and a very generous TDP of 250 W are needed to push the card even further. As RTX 2080 overclockers already know, TU104 can easily handle higher clock rates than NVIDIA's RTX 2080 reference clocks. However, the cost is a higher power consumption and all the associated disadvantages. The RTX 2080 Super has increased base and boost clock rates by 9% and 6%, respectively, and the card should now be able to withstand 1815 MHz (or better) in most games – and this may be somewhat conservative. In contrast to enabling more SMs, this benefits every facet of the graphics card, making front-end and back-end performance faster and faster.

The penalty for this is that the 215W TDP RTX 2080 is replaced by the 250W TDP RTX 2080 Super. This happens to be the same TDP as the even faster GeForce RTX 2080 Ti. So it's easy to see where this is going: NVIDIA boosts overall performance at the expense of energy efficiency. As we will see in our benchmarks, the situation is a bit more complex but not immense. Perhaps the wildest factor is that the RTX 2080, unlike the RTX 2080 Ti, has problems filling its 250 W TDP. NVIDIA has increased the TDP to make sure the card has enough headroom for higher clock rates. This has definitely succeeded. As a result, the card runs much faster than the original 2080 or the 2080 Ti or the 2070 Super.

Faster GDDR6 memory for graphics cards

But perhaps the most interesting The RTX 2080 Super does not even change the GPU itself, but what NVIDIA does to feed the beast on the memory side. Since the introduction of the first GDDR6-equipped graphics card a year ago, each card has been shipped with 14 Gbps (or less) of memory. 14 Gbps was the fastest speed stage offered by major memory vendors at launch, but not the fastest on the roadmap. The 16 Gbps memory is also in the planning stage, and with the introduction of the new NVIDIA card we get our first card with this memory.

For the introduction of their new cards, NVIDIA has accessed Samsung's new 16 Gbps card. 8Gbit chips for the new cards. With a 256-bit memory bus for the card itself, the RTX 2080 Super retains the same 8GB VRAM as the original RTX 2080, but gives the card more bandwidth to play. However, in a rare situation for NVIDIA, they are somewhat conservative in terms of memory clocks – the RTX 2080 Super achieves only 15.5 Gbps instead of the 16 Gbps the memory can deliver.

According to NVIDIA, the ultimate reason for this is the decision to make the Super cards as simple as possible and to specify the cards so that board partners (including themselves) can use them the same Boards and coolers from their Vanilla RTX 2080 cards. For this reason, the RTX 2080 Super is based on a board originally designed and qualified for the 14 Gb / s GDDR6. For true 16 Gb / s memory clocks, a new board is apparently required, for which board manufacturers do not want to spend the money with a clock speed increase of 3% (and a lower performance).

Of course I clocked our card with ease and without any problems at 16 Gbps; But then our stability standards are not as high as those of NVIDIA (no matter what mass production is concerned). For curious overclockers, there is a bit more space here, with the usual warnings of overclocking. The net result for all others is that the RTX 2080 Super has 11% more memory bandwidth than the RTX 2080 it replaces (and the RTX 2070 Super).

As a result, and the higher TDP, the RTX 2080 Super is usually more tied to the GPU than most NVIDIA cards. The higher memory speeds more than make up for the fully enabled GPU and its higher clock speeds. So while no GPU has memory bandwidth available, RTX 2080 Super is pretty busy. This also means that overclocking will be a bit more one-dimensional, with GPU clock rates being a major part of the major factor.

Product Positioning, Availability, and Competition

As with the other RTX 20 Series supercards, NVIDIA and its partners will be launching from reference cards to fully custom cards. These are the same TU104 GPUs whose memory has been validated to run on the same circuit boards. As a result, board partners can quickly leverage their existing 2080 designs to bring new cards to market. Similarly, the card volume should not be a big problem – although the demand, as we have seen in the now sold-out RTX 2070 Super, from the outset is high and therefore there is still the potential for sell-off.

Also with other supercards NVIDIA again publishes its own reference card. Unlike the factory-overclocked Founders Edition RTX 2080 of the previous year, NVIDIA will not overclock its RTX 2080 Super or charge a premium for it. Your reference card will still be distributed under the Founders Edition brand name, but should be the same as all other reference cards. In the meantime, as usual, the cards are sold directly by NVIDIA.

NVIDIA also completes the latest video game bundle for the new card. The RTX 2080 Super receives a two-game bundle consisting of Wolfenstein: Youngblood and Remedy's upcoming Control . NVIDIA, for its part, has avoided most of the game bundles for its high-end cards in recent years. Therefore, it is remarkable that they are included here. As part of the reorientation of the entire product stack, NVIDIA is trying to increase the value of their cards without actually lowering prices. So this is another way to do that.

The competition for the RTX 2080 Super is now a mixed group. With the vanilla RTX 2080 on the way out (along with the RTX 2070), the competition on the NVIDIA side is the RTX 2080 Ti top and the RTX 2070 Super bottom. The RTX 2080 Ti is by far the fastest NVIDIA GeForce card. The ticket prices have been unpredictable for months and it has not stopped yet. Even the cheapest (new) card will bring you $ 1150, a $ 450 premium over the RTX 2080, so close that you probably will not be able to save the money by asking for the card you want to select. Meanwhile, the RTX 2070 Super is definitely the spoiler for the NVIDIA stack. The card offers a slightly lower performance than the original RTX 2080 for $ 200 (perhaps it's not surprising that the card is currently sold out).

As far as the AMD charts are concerned, the Radeon VII is the technically closest competition for AMD launched in January of this year as a surprise. The card generally offers an RTX 2080-like performance for $ 699. But even before the launch of RTX 2080 Super, it was probably born dead in the consumer & PC gaming world – it could not do anything better than the RTX 2080 – and the RTX 2080 Super is the last nail in this coffin. Far more interesting at a fraction of the price is the new Navi-based Radeon RX 5700 XT. It's not an RTX 2080 super competitor, but may be an even bigger spoiler than the RTX 2070 Super, which offers a lot of performance for $ 300 less. Overall, the RTX 2080 Super may be half-affordable compared to the RTX 2080 Ti, but it's still decidedly a high-end NVIDIA card, so the value per dollar is not a strong selling point here.

Source link