NVIDIA RTX 5080 is a very broken graphics card since its launch. It is a smaller, more powerful GPU than its predecessor Ada and does not really offer much more silicone than RTX 4080 great. Yes, it is seemingly the same price, but it provides only about 15% higher frames of the frame per second at 4K, and less at lower resolutions.
Oh, boy, I didn’t expect it. For years I have not seen this level of acting in a graphic processor issued.
This did not stop them from all sales when they appeared for sale. After all, the impressive performance of generating many frames and its price stickers for 999 USD make it much cheaper than $ 2000 USD+ RTX 5090 and is still able to offer a vigorous generation in DLSS 4 support games.
I felt quite uninitiated by the GPU RTX Blackwell of the second level if the truth is said. It will be a great basis for high -class gaming computers, but in raw silicon the needle barely moved from Ada to Blackwell.
Well, I was quite uninhap until I started showing this thing. And oh, boy, I didn’t expect it. I have not seen this level of head sealing in the issued graphic processor in years. There is a lot of raw performance on the floor of the engineering room, and this means that what was quite average performance above the RTX 4080 is more persuasive.
I see over 500 MHz additional performance for fun, which gives me virtually another 10% higher numbers of frames per second. And all without actually taxing the power or thermal performance of the founding plaque.
Since we got into the areas of the top -up clock, and knowledgeable GPUs generally change clock speeds in flight depending on the available thermal headrest and power supply, there was not much for overlap.
Sure, you can augment the frequency shift, but rarely by a significant amount, especially not to a level that would register as something more than a type of variance of the number of frames per second, which can be seen only from standard tests. Even the factory excessively rounded cards barely offered something particularly thrilling. Look at the ASUS RTX 4080 great great and you are lucky if you choose another 90 MHz from it.
It differs here, in the case of RTX 5080, it is that even these factory transmitted cards offer something different. Asus Rog Astral RTX 5080 offers overclock +173 MHz and Gigabyte Aorus RTX 5080 A +188 MHz Overclock. We have not seen this level of trust in production when it comes to their OC cards for a long time.
Now, if people can suggest, I have won a silicon lottery – and as much as Nvidia wants to point out that users should not expect that each board will be able to hit the same clock buckle by 0.5 GHz, which I see – this – it It is obvious to me that there is a lot of head in GP 203 GPU.
I’m not a great overclocker. Although I have been playing with the GPU professionally for the best 20 years, I am a very basic strainer. MSI Afterburner is my jam and I’m not going too deep. I played with a scanner of OC, which is trying to adapt the tension curve to the capabilities of your graphic processor, which can ensure a higher stable performance, but it did not really move the needle using RTX 5080.
However, I just go Buckwild with MHz sliders, certainly yes. After starting Black Myth Wukong Benchmark I could quickly go through gears and achieve a solid shift +500 MHz. However, this did not give a 3.1 GHz clock, which I wanted, so I squeezed GPU even more and hit +525 MHz.
However, it would not go very much at +550 MHz. In this way there was a catastrophic failure, black screens and a graphic processor, which was rather uncertain when it became so high again when I started again. Letting it lie when I returned to +525 MHz and has remained stable since then.
I was rather a bachelor with memory clocks and went up to +1,000 MHz, where I just pulled the line out of fear.
As you can see, there is an augment in performance. Almost 10% unevenness at 4K and near 1440p (which is significant, as if you were running DLSS in quality at 4K Really action).
Combine this with the unevenness RTX 5080 RTX 5080 in RTX 4080 Super, and now you are talking about a 25% augment instead of 15%. This is a much more convincing character and makes the RTX Blackwell card a more tempting option, especially because it is the same price, as well as equipped with a multi -class gene as an additional panacea.
Of course, it is still far below the augment in the number of frames per second by 50%+, which you saw from RTX 3080 to RTX 4080, but this is a better look than the RTX Blackwell RTX card, and much closer to 30% up to 30% augment in RTX 5090 gets RTX 4090. And in fact it also provides several opportunities in which RTX 5080 is able to match the raw power of rendering this best ADA graphics processor.
It barely moved thermals or GPU performance requirements.
Another thing that impressed me and made me feel like it was more than just a numbers, but I strain that I could legally live with a full -time job, is that she barely changed thermals or requirements for GPU performance. Half, I expected an exponential augment in obtaining an additional 10%, which would explain why Nvidia was so conservative in the case of top -up clocks. But no, it has barely changed.
Quite Why I am not sure if there is such a high performance, I am not sure. I now have only one edition of the RTX 5080 founders, so it will be intriguing to see what cards of other manufacturers they provide on this front, but I expect that I will still see some performances.
It is intriguing that the upcoming RTX 5070 Ti also uses the same GPU GB 203. Sure, it will be quite a cut version, but also has the default top -up clock well below RTX 5080. If this card is the same head on this card, it can be a winding king upper average range.
Despite this, I intend to heat up in the glory of winning in a silicone lottery. And continue to run the 2024 football manager. I mean that he will have to do it for some time, now the 2025 version has been canceled, eh 🫗.
