AMD Radeon VII Anti-Aliasing & 5K, 8K Media Signs vs. RTX 2080 | GamersNexus


This content piece is active as an academic exercise, which means we will leaving "world-wide" usage issues at that time to explore how the cart is carrying and levels on the basis. This means to & # 39; Look at unacceptable situations, similar to 8K configuration, just to see if AMD has a scaling over NVIDIA; in theory, based on what AMD has told us about performance, it should, so we just need to confirm that. Anti-aliasing produces many samples of each pixel, often 2x, 4x, 8x, or 16x, and the impact of achievement is at the same level as greater reach. It will reach the same part of the pipeline.

If AMD's memory bandwidth is really beneficial to him in games, what would we do would include a & # 39; Happening as a delta between Radeon VII and the RTX 2080 as an analysis and anti-aliasing increase, with a gap expanding or increasing; expand, leaving RTX as a compromise reduction. We have already shown how Radeon VII can close the gap at higher expectations at times when we posted our survey, and discuss how the gap – similar to F1 2018 – a & # 39; expand at 1080p. Now, it's time to see if it is turning or running; closing to high expectations; However, note that the obvious thing is that we are entering the boundaries where the content is unbalanced at the frame, and so it's the same. call this academic exercise.

Heavy rain

Starting with screens to & # 39; Set to "Low" texts set to "High" and adjustment changes at 4K, 5K, and 8K, we saw NVIDIA lead a 27.4% over base, and # 39; move to 27.9% – in error – over a 5K base line, and then 8.1% drive 28.1%. There is no change here from 4K to 8K, regardless of scaling, although it is clear that these ambitions reduce their gameplay as they increase.

We then reset a test with "Extreme" screens, which found AMD to have a & # 39; get land in this situation. AMD worked at 1566 average points, with NVIDIA at 1504 points. We will enter this test 4 hours and we gave it access to, and # 39; confirm that this delta is outside the normal separation.

Having learned that greater differences will appear with Extreme Screenshots, we will Testing from 1080p to 4K. At 1080p, NVIDIA at RTX 2080 played about 25.8% better than the Radeon VII card at a basic level of 100%. At 1440 p, the RTX 2080 benefit had rapidly decreased to 16.1%. The 4K results are too close and they create a messenger record, but the RTX 2080 is virtually enhanced by Radeon VII here, with Radeon VII a & # 39; lead average couple point. Now, in terms of a "true" framework, the criteria worked at 11-12 PHPS AVG. This has not been played, clearly, and no benefit has been achieved since they would not really use those settings, but it is interesting to see how two card cards can be distributed in different exam situations.


There are many anti-aliasing options at GTA V, so we use both MSAA from 2x to 8x and then we will add 8x reflection on MSAA. This will start with just a FPS AVG chart at the different settings, which will best show how to change its & # 39; gap per test.

At 4K and Ultra and Ultra settings for everything, MS2A 2x gives us 58.8FPS AVG for the 2080 and 51.4FPS AVG for the Radeon VII, but the most important indicator is that 2080 a & # 39; 14.4% over the Radeon VII S Post 4x MSAA card, that direction does not change significantly and drop to 14.1%, effectively in error. At 8x MSAA, we'll start to & # 39; see some move. The AVG FPS 27.5 has on the RTX 2080 Radeon VII and 31FPS, with an increase in RTX 2080 to 13%, down from an original 14.4%. At that time also compile the 8x MSAA anti-aliasing reflection, we & # 39; Seeing a different variation of 10.8% benefit for FE 2080. Looking at its & # 39; line graph as a whole, we can see that AMD starts to & # 39; recover some loss to the end of the scale, that is just that we are in unlimited boundaries for other purposes.

For a comment, we ran the level of test trips, the general divide that ran to around 0.3FPS AVG. These results are very correct.

AA Firestrike results

Firestrike is still one of the best tools for this kind of work of messengers. For this, we will watch GT1 and GT2 individually. GT1 will pack the GPU with polyons and split, shadow shade and light, and use computer screens for post-processing and grain physics. GT2 is very heavy when it works on a blow tool and it will greatly increase the process of processing each frame, but it is less than 50% of it; work. So GT2 should be better at Radeon VII of the GT1, mostly speaking.

Starting with GT2, we & # 39; see that NVIDIA is comparable to Firestrike Ultra stocks under AMD performance, at 86.7% of the basic performance of Radeon VII 100%. This is unique to this commendation, so far. At 2x MSAA, that gap will flow to 90.7% of basic performance. 4x MSAA brings us to 94.6% baseline, with the last 8x MSAA and # 39; allows RTX 2080 to exceed AMD and maintain at 104.7% of basic performance.

It makes sense that AMD is usually a GT2 competitive, where there is a focus on both work and memory. In terms of why NVIDIA starts to & # 39; Ongoing once MSAA is taken to 8x, we work with a 3DMark team and others to try to gain a better understanding of the particular behavior. It is estimated that this may need to be done with a memory co-operation of NVIDIA, or which can be done; find anti-aliasing anti-aliasing anti-aliasing software, or other attainment level advantage.

For GT1, where heavier and pollen are heavier, we see the two cards that are; Starting up current under Firestrike Ultra stock settings, with NVIDIA receiving 16% at 2xMSAA, 21% at 4x MSAA, and 27% at MSAA 8x.

At 8K with 8x MSAA, we will reach visually impaired visually impaired levels, but we 're Seeing that Radeon VII is a long way ahead of the RTX 2080. This is how he fully remembered the RTX 2080, and at that time he started to start. He caught hold of him and reminded him.

Far Cry 5 HD Textures + AA Testing

Far Cry 5 is coming. For this, we tested with SMAA, TAA, and HD texts under different settings.

In the first 4K test and its & # 39; Using HD Textures with SMAA, we measured NVIDIA at 62.8FPS AVG and AMD at 60.6FPS AVG, adding NVIDIA around 3.5% onwards. 1% was very different, as 0.1% were sufficient.

In order to test scale with different settings, we also tested with a smaller amount of VRAM but with similar situations. We can now change to that cart. The results, which hinder HD texts, work equally to the HD text output. It is not surprising that we are inside a mistake error, because texts do not affect performance if VRAM is not the limit. On both devices, in contrast, the knowledge has not changed from the previous chart. Qualitatively, the product is the same on each device. Again, NVIDIA is about 3.3% faster here.

Moving the place to TAA, performance will be Falling slightly for each device, but NVIDIA still benefits 3.3% in AVG FPS. HD texts and anti-alias changes in Far Cry 5 are not enough to reduce the delta display or by card-to-card switch.

Decision: Not just what we expected

We did not think it was right. It seems we've put it in a few moments, but Firestrike threw us out of the way. We have come to the 3DMark team and NVIDIA engineers to try to better understand a particular performance setting. The only example we saw was a huge increase for Radeon VII when we put off the VRAM on the RTX 2080, which was done by operating Firestrike Ultra at the 8K and 8x MSAA resolution. This is, of course, totally non-play, and so we will stress again in saying that this is not just an exercise in this research, not one who is especially practical. At 8K / 8xMSAA, we see that Radeon VII operates around 4.7-4.8FPS AVG to 2080 ~ ~ 1-1.7FPS AVG. This is a "big" percentage development, and it is because we have a " shaped on a woodwood umbrella on 2080, but clearly it is not easy to play on the same device. Despite this, it is interesting data, and it can help to understand other features; cart in future exams.

Editor, Test: Steve Burke
Video: Andrew Coleman

Source link