NVIDIA GeForce RTX 4080 Founders Edition Review in Spanish
NVIDIA GeForce RTX 4080 Founders Edition Review in Spanish
La NVIDIA GeForce RTX 4080 finally it has been released at a price from 1199 US dollars (MSRP USA US$) and this launch has been bumpy. From the cancellation From the 4080GB RTX 12 version (which seems to be the RTX 4070 Ti) to adapter/connection issues with the new 12VHPWR input, the RTX 40 series hasn't sailed the seas in the best of conditions.
In this stage land la RTX 4080 and let's analyze what it offers us.
The NVIDIA GeForce RTX 4080 has been sampled by NVIDIA for today's review and future testing.
Table of Contents
NVIDIA GeForce RTX 4080 – Technical Section/Specifications
The RTX 4080 will be promoted by NVIDIA for content creation, streaming, and gaming alike, due to the improvements to the Ada Lovelace architecture that we detailed above, as well as new technologies that we will detail in this section.
Regarding specifications, the RTX 4080 offers 9728 CUDA cores and 49 Shader TFLOPS, 304 fourth-generation Tensor Cores with 780 Tensor TFLOPS for AI graphics processing and DLSS; and 76 third-generation RT cores that will offer 113 RT-TFLOPS for ray-traced graphics.
Other technical features include 16GB of GDDR6X memory, the evolution of technologies such as NVIDIA Broadcast, Studio and DLSS 3.0; and other innovations seen in the Ada Lovelace architecture for the RTX 4090. NVIDIA promises that the RTX 4080 will be up to three times more powerful than the RTX 3080 Ti in titles that support RTX raytracing.
The emphasis on raytracing is evident in NVIDIA's promises for DLSS 3, which evolves the supersampling technique to display graphics at higher resolutions with lower dropped frames per second. One of the main changes is that DLSS 3 can render entire frames instead of just pixels. According to the company, it works with an optical flow accelerator, game motion vectors, an AI frame-generating convolutional autoencoder, and a super-low latency pipeline using Reflex.
In our review of the 4090 we explained how the optical flow accelerator works and a good part of DLSS 3. But it is noted that the accelerator indicates to a neural network the speed and direction of the pixels so that they are rendered, and then this network generates the intermediate frames. In this way, full frames are generated and a performance of up to four times that achieved in previous architectures is offered, without having to resort to "brute force" to render each frame. And the company provided benchmarks with different games that show its position.
Finally, DLSS 3 will be compatible with "old" games that use DLSS 2, and it will be easier for developers to integrate the new version into their games through an SDK known as NVIDIA Streamline.
Evolution in specifications
Using TSMC's 4nm process for its manufacturing node, the RTX 4080 follows the mold of the 4090 along with other architectural highlights. Ada Lovelace, counting on 45.9 billion transistors.
In our review of the RTX 4090 we explain how the fourth generation tensor core works. The RTX 4080 will offer 480 TFLOPS of performance for AI inference tasks, allowing it to “predict” frames as we described above. These will be rendered at 1440p resolution and upscaled to 8K resolutions using tensor cores.
Meanwhile, third-generation RT cores will reduce the workload of streaming multiprocessors (SMs), in conjunction with a micromap opacity engine and another microgrid engine, promising twice the performance of Ampere.
And to keep the GU at proper temperatures, the RTX 4080 has an improved cooling design, which while it looks similar to the RTX 3090 to the naked eye, has an improved internal layout. The company ensures that the airflow has increased by 15% in front of the Ampere flagship with improvements in the design of the heatsink and fans. And the RTX 4080 would be 10% more efficient than the 3080 Ti in terms of power efficiency. In this section, NVIDIA recommends a minimum 750W power supply, and the GPU will offer 3-pin PCIe 8-cable connectivity, or a 12-pin Gen XNUMX PCIe connector cable.
additional technologies
In addition to hardware and brute force improvements, NVIDIA promises performance improvements in gaming and other fields by using technologies that have evolved with the new architecture.
For streaming, the RTX 4080 will support dual AV1 encoders, which will deliver higher streaming quality than H.265 / H.264 with the same bit rate. Depending on the platform, connection and configuration, improvements of up to 50% in video encoding efficiency will be seen compared to Ampere. There are also two dedicated engines for encoding compared to the single engine in Ampere, which will be able to work in parallel.
Other important technologies to improve performance include NVIDIA FrameView, to measure responsiveness and smoothness of motion, NVIDIA ICAT to analyze performance and image quality, and enhancements to NVIDIA Broadcast to increase the role of artificial intelligence in streaming.
These technologies, as well as DLSS 3 and architectural improvements, also promise to positively impact content creation, video rendering, raytracing for video and animation creation, specialized drivers for content creators under NVIDIA Studio, and tools for developers to take advantage of. RTX as Reflex and RTX Remix.
Specifications table
| GeForce RTX 3080 | GeForce RTX 4080 | |
| Graphics processing clusters | 7 | 7 |
| Texture Processing Clusters | 34 | 38 |
| Streaming Multiprocessors (SM) | 68 | 76 |
| CUDA Cores | 8704 | 9728 |
| tensor nuclei | 272 (3rd Gen) | 304 (4th Gen) |
| RT cores | 68 (2nd Gen) | 76 (3rd Gen) |
| texture units | 320 | 304 |
| ROPs | 112 | 112 |
| boost clock speed | 1710 MHz | 2505 MHz |
| memory clock speed | 9500 MHz | 11200 MHz |
| Memory data rate | 19 Gbps | 22.4 Gbps |
| L2 cache size | 5MB | 65536 KB |
| total video memory | 10 GB GDDR6X | 16 GB GDDR6X |
| Memory interface | 320-bit | 256-bit |
| Total Memory Bandwidth | 760.3 GB / sec | 716.8 GB / sec |
| texture rate | 465.1 Gigatexels/sec | 761.5 Gigatexels/sec |
| Fabrication process | Samsung 8 nm 8N NVIDIA Custom process | TSMC 4N NVIDIA Custom Process |
| Number of transistors | 28.3 billion | 45.9 billion |
| Connectors | 3 x Display Port, 1 x HDMI | 3 x Display Port, 1 x HDMI |
| Maximum resolution | 4K at 240Hz or 8K at 60Hz with DSC | 4K at 240Hz or 8K at 60Hz with DSC |
| Forma | double groove | triple groove |
| power connectors | 1×16 pin (2×8 pin adapter) | 1×16 pin (3×8 pin adapter) |
| minimum power source | 750 Watts | 750 Watts |
| Total Graphics Power (TGP) | 320 Watts | 320 Watts |
| GPU max temperature | 93°C | 90°C |
| PCI Express interface | Fourth generation | Fourth generation |
NVIDIA GeForce RTX 4080 – Photos and unboxing
According to the photos, it appears that the supplier of the adapter of the RTX 4080 Founders Edition es ASTRON, due to the shape of the anchoring system. Anyway, I will try to confirm this in the following days.
Synthetic and gaming benchmarks (1080p, 1440p, 2160p)
With the release and announcement of new generation video cards, we have to update our benchmark (once again). The best processor that we have in our hands to carry out the necessary tests is the processor Intel Core i9-12900KF.
A statistic reminder...
Before detailing our setup system, let's refresh a bit what AVG FPS and 1% LOW are.
AVG FPS (Average FPS): As the name says, it is the average number of frames per second within a specific sequence. It is the most used measure, but it does not detail the whole story, since there are FPS drops.
1%LOW: Within an entire frames per second dataset, the 1% LOW is the value equal to the lowest 1% within the frames dataset (of the specific sequence ordered ascending). In simpler terms, it is the frame where you see the FPS drop that exists within a specific sequence.
Also, we suggest you read any additional feedback about the gaming experience (jerks) that may or may not have existed during testing.
Benchmarks (GPU Benchmarks – Raster – 2022)
Our revamped benchmark features the best processor we have in stock, the Intel Core i9-12900KF. We use said processor, since it is the one that will generate the least bottleneck for the GPUs tested in scenarios where the limiter may be the CPU.
The focus is aimed at achieving 100% in the performance of the video card, NVIDIA GeForce RTX 4080 Founders Edition. However, there are scenarios where the GPU will not scale more at lower resolutions. (1080p and even 1440p).
Also, we've upgraded to Windows 11, but we've VBS disabled (Virtualization-Based Security), because in the middle of our tests, we noticed that it took considerable performance away in certain scenarios or generated stuttering.
CPU: Intel Core i9-12900KF (Power Limiters Disabled) (https://amzn.to/3fXVGCb)
Board: Z690 Maximus Hero (BIOS 2004) (https://amzn.to/3GuXC0j)
RAM: Kingston Fury DDR5 5200C40 – 2x16GB (https://amzn.to/3tqmWwv)
T.video (what we are testing): NVIDIA GeForce RTX 4080 Founders Edition
Operating system: Windows 11 Home Edition 22H2 – VBS OFF
Liquid refrigeration: Custom Water Loop (EKWB)
SSD: HP FX900 1TB + Samsung 970 EVO PLUS 2TB
Driver: NVIDIA Press Driver
Power supply: EVGA P+ 1300W
Gaming – Rasterization
All tests are done at the highest quality available, unless otherwise specified.
Let's see the first title, Assassin's Creed: Origins.
Assassin's Creed: Origins (1080p, 1440p, 2160p)
Game Engine: AnvilNext 2.0
Borderlands 3 (1080p, 1440p, 2160p)
Game Engine: Frostbite 3
Monitor (1080p, 1440p, 2160p)
Game Engine: Northlight Engine
Death Stranding (1080p, 1440p, 2160p)
Game Engine: Decima
F1 2022 (1080p, 1440p, 2160p)
Game Engine: EGO “Boost” Engine 4.0
God of War (1080p, 1440p, 2160p)
Game Engine: Proprietary
Shadow of the Tomb Raider DX 12 (1080, 1440p, 2160p)
Game Engine: Foundation
Shadow of War (1080, 1440p, 2160p)
Game engine: LithTech Jupiter EX
Strange Brigade DX12 + Async (1080p, 1440p, 2160p)
Game Engine: Asura
The Witcher 3 (1080p, 1440p, 2160p)
Game Engine: REDengine 3
Final Analysis – The successor to the RTX 3080 comes in the middle of a storm
The second graphics card in the NVIDIA GeForce RTX 40 (Ada Lovelace) family, the NVIDIA GeForce RTX 4080 hits the market, one month after the RTX 4090 launch. new architecture compared to its predecessor, the GeForce RTX 3080.
If we compare the RTX 4080 versus the RTX 3080 in price, there is an increase of 71.53%. This is noticeable for the end user's pocket, although it should also be noted that the RTX 3080 came with 10/12GB of VRAM. By contrast, the RTX 4080 comes with 16GB, which can be useful for certain scenarios (productivity).
In terms of price, the NVIDIA GeForce RTX 4080 is closer to what it once was BEST graphics card gamer from NVIDIA, the GeForce RTX 3080 Ti. This has a suggested price to the end user (USA) of 1119 US$. If we compare between these two, the increase in the price of the RTX 4080 over the 3080 Ti is 7.14% more.
Now let's see how the RTX 4080 performs against the RTX 3080 and RTX 3080 Ti and the RTX 4090..
NVIDIA GeForce RTX 4080 Founders Edition ($1199) Versus | |||
| 1080p | 1440p | 2160p |
NVIDIA GeForce RTX 4090 ($1599) | -5.38% | -15.16% | -25.90% |
NVIDIA GeForce RTX 3080 Ti ($1119) | 22.17% | 31.03% | 29.40% |
NVIDIA GeForce RTX 3080 ($699) | 29.08% | 41.28% | 41.35% |
NVIDIA GeForce RTX 3060 Ti ($399) | 80% | 110.50% | 118.24% |
Against its predecessor, the GeForce RTX 3080, there is a 41.30% improvement in 4K UHD (2160p). That improvement is similar at 1440p as well. At 1080, the improvement is less because we are in several games where there is a bottleneck generated by the game or by the processor.
The improvement is minor compared to the GeForce RTX 3080 Ti, with 29.40% higher performance at 2160p, 31.03% at 1440p, and 22.17% at 1080p.
The NVIDIA GeForce RTX 4090 remains the biggest beast NVIDIA brings to the table with performance in another class and is the most powerful gaming graphics card on the planet today.
NVIDIA GeForce RTX 4080 in difficult times
Unlike other launches that we have had coverage with NVIDIA products, this is the first that has several short-term adversities. The issue of the much debated 12VHPWR connector will undoubtedly be a variable that will make many think more than once, if they want to be early adopters of the new format.
The NVIDIA GeForce RTX 4080 has the same connector as the RTX 4090, although in terms of consumption and TGP, it is considerably smaller. The problem, which does not have a final solution at the time of writing, lies both with the PCIe adapters, and also with what appears to be a revision to the 12VHPWR connector by PCI SIG.
Not having any official announcement, which I would have liked from NVIDIA (it is still under investigation) it is difficult to recommend the GeForce RTX 4080 at this time due to a few factors:
-NVIDIA is missing a final resolution on the connector problem.
-In a little over three weeks AMD launches its new graphics cards that will be direct competition to the RTX 4080 and the "supposed" RTX 4070 Ti.
It is time for him to play the role of the Devil's lawyer. As of date, there are 26 reported cases (via Reddit) of affected units (RTX 4090). This is probably not even 1% of all RTX 4090 video cards sold across the globe. Under another scenario, this will probably go unnoticed until hundreds of users complain.
Being a high-cost product (RTX 4090) the few cases to date turn this low rate into a snowball. At the moment, there are only reports by media outlets that have tried to replicate the problem, as well as have obtained privileged information that a problem does exist (although there is still no final resolution).
I imagine that in the next few weeks NVIDIA will pronounce itself, although I speculate that this will happen after Black Friday. I think the reason for the silence is because reproducing the problem and putting a solution on the table takes time and a lot of human resources.
Having said that, I think there is another factor that perhaps plays a part in the momentary silence and that is to clear out the remaining stock of GeForce RTX 30, from RTX 3070 and up. The traditional 8-pin PCIe connector is perceived to be “safe” and momentary noise will favor moving the remaining stock.
These dates are a good occasion to liquidate stock.
Removing the 12VHPWR connector factor…
We think the RTX 4080 could have a place for those who can't afford a GeForce RTX 4090 and game at 2160p with better performance than what one was getting with the RTX 3080, RTX 3080 Ti, 3090, or 3090 Ti. The raster performance is present, although the RTX 4090 is a beast on another level (and that there is the possibility of a 4090 Ti).
The price increase of the 4080 compared to the 3080 takes away a lot of its appeal and perhaps the card that could shine the brightest this generation, could be the alleged RTX 4070 Ti.
DLSS3 is a factor to consider and in the coming months we will see more games with this technology. They have just announced a remastered title that we hope will attract the attention of many players and could be NVIDIA's workhorse in the short and medium term... The Witcher III.
Ray tracing is slowly being adopted, but a gamer who hasn't tried Geralt of Rivia in his most popular iteration is a missed opportunity. CD Projekt RED announced yesterday the remastering that will bring the game once again, now with Ray Tracing.
This could be the first ray tracing title that every gamer, lover and even the one who says “Ray tracing doesn't work”. Let's hope that the level of ray tracing implemented in the remaster lives up to expectations and that the experiences of Cyberpunk 2077 have served to improve the game GOTY of 2015.
In a nutshell, The Witcher III Remastered may be the accelerator the industry "needs" for ray-traced games to become the norm, but it will depend a lot on CD Projekt RED's implementation of one of the most acclaimed games in recent years. last ten years.
Going back to the GeForce RTX 4080…
The cost per frame is not very convincing. The next table shows it:
NVIDIA GeForce RTX 40 series | |
| Cost per FPS (2160p) |
NVIDIA GeForce RTX 4090 ($1599) | $ 9.00 |
NVIDIA GeForce RTX 4080 ($1199) | $ 9.11 |
NVIDIA GeForce RTX 3080 Ti ($1119) | $ 11.00 |
NVIDIA GeForce RTX 3080 ($699) | $ 7.51 |
NVIDIA GeForce RTX 3060 Ti ($399) | $ 6.62 |
The RTX 3080 still makes a good case at 2160p with decent performance, although it doesn't get the FPS achieved by the GeForce RTX 4090 or RTX 4080, nor does it have DLSS3, among others.
The NVIDIA GeForce RTX 4080 is a mixed story with enough bumps at launch that hopefully the coming weeks will clear things up.
NVIDIA needs to release a statement before RDNA3 cards are available on the market and have a final resolution to the 12VHPWR issue.
The power requirement of the RTX 4080 is much lower than the RTX 4090, but due to association error, the end user will link that, because the RTX 4090 has problems and they use the same connector, the same happens with the 4080 (they could still exist). cases, as there are also burned PCIe 8-pin connectors).
What we can assure in the short and medium term is that the GeForce RTX 4090 is in a higher league than the 4080 and it will remain there without any dispute, apparently, as the king in high-performance gaming.
Whether or not you decide to buy, the only suggestion we can give you is that the only thing that protects you as end users until there is a statement is your warranty. So we don't suggest using third party cables, but the ones that came with your video card box. So far, there are no known affected RTX 4090 units locally (and hopefully for the sake of end users, there is no such scenario).
