Is AMD lastly focusing on Nvidia? Radeon RX 7800 XT & RX 7700 XT launch particulars, FSR 3, and extra

What simply occurred? AMD formally introduced the Radeon RX 7800 XT and Radeon RX 7700 XT graphics playing cards. We’ve got all the main points, together with specs, launch date, pricing, and AMD’s efficiency claims – plus extra info on FSR 3, Anti-Lag+, Hypr-RX, and extra. So let’s get began.

Pricing, Specs and Availability

Let’s start with the 2 new graphics playing cards, which will likely be obtainable on September sixth. The RX 7800 XT will likely be priced at $500 US, whereas the RX 7700 XT is barely much less at $450 US. Each nonetheless lie in the midst of the present market in direct competitors with Nvidia’s GeForce RTX 4070 and RTX 4060 Ti.

These new playing cards are primarily based on AMD’s RDNA3 Navi 32 GPU die, which is chiplet primarily based. It has one GCD within the center, a smaller GCD than was used with Navi 31, however continues to be constructed on TSMC’s N5 node. Flanking this are 4 MCDs utilizing a TSMC N6 course of. So, from a GPU design perspective, it is extra just like the RX 7900 XTX and 7900 XT than the RX 7600, which used a monolithic Navi 33 die.

The 7800 XT makes use of a completely unlocked Navi 32 die with 60 compute items and all 4 MCDs, providing a 256-bit reminiscence bus and 64 MB of infinity cache paired with 16GB of GDDR6 reminiscence at 19.5 Gbps. In the meantime, the RX 7700 XT is reduce right down to 54 compute items, and it has three of the 4 MCDs, so a 192-bit reminiscence bus and 48 MB of infinity cache. This permits for a 12 GB GDDR6 reminiscence configuration, though it is also clocked barely decrease than the 7800 XT, at simply 18 Gbps. Each playing cards have an identical recreation clock, 2171 MHz for the 7700 and 2124 MHz for the 7800, though the rated increase clock for the 7700 is considerably increased. Usually, the sport clock is a extra sensible reflection of in-game clock speeds.

As well as, each fashions help new RDNA3 options like AV1 video encoding and DisplayPort 2.1. Board energy rankings are mid-tier, as anticipated, with the 7800 XT listed at 263W whereas the 7700 XT runs at 245W. Provided that AMD is pitting these fashions in opposition to the 200W RTX 4070 and 160W RTX 4060 Ti, this means inferior effectivity for Navi 32 in comparison with Nvidia’s AD104 and AD106 silicon.

RX 7700 XT Efficiency Claims

Based mostly on specs and pricing alone, there are some issues in regards to the value of the RX 7700 XT. The 7700 XT is 10 % cheaper than the 7800 XT, for a 10-percent discount in compute items. That sounds cheap. Nevertheless, the reminiscence system is considerably lowered. The 7700 XT, with its 192-bit bus and slower reminiscence, finally ends up with 31 % much less reminiscence bandwidth than the 7800 XT.

Realistically, the efficiency hole between these two fashions will likely be greater than 10 %. That is definitely what AMD is anticipating primarily based on this slide, the place the 7800 XT is, on common, 22 % quicker than the 7700 XT, albeit throughout simply 4 video games.

If that is an correct reflection of efficiency, the 7800 XT will likely be round 20 % quicker for only a 10-percent value premium, giving the 7800 XT a greater price per body. This seems to be a really related scenario to the 7900 XT versus the 7900 XTX, the place the XTX was solely round 10 % costlier for 20 % extra efficiency at launch. We’ll have to attend and see the place the ultimate benchmarks lie in our testing, however our preliminary evaluation factors towards the 7700 XT being lifeless on arrival, at the least in AMD’s line-up.

AMD’s central efficiency claims for the 7700 XT examine it to Nvidia’s RTX 4060 Ti 16GB. On common, throughout this 20-game configuration pattern (beneath), AMD has the 7700 XT 12 % forward of the 4060 Ti. Nevertheless, this needs to be taken with a wholesome dose of skepticism, given these are first-party benchmarks, and AMD hasn’t at all times been tremendous correct with the efficiency of this technology. A few of these outcomes are ray-traced video games, and a few use rasterization, so if we cut up these off individually, AMD is claiming a 17-percent lead in raster however a four-percent deficit in ray tracing.

At first look, this seems cheap. AMD priced this new card 10 % beneath the 4060 Ti 16GB however 13 % forward of the 4060 Ti 8GB, which we consider are terrible graphics playing cards with horrible costs. VRAM capability additionally slots within the center, providing 4GB greater than the anemic 8GB 4060 Ti, however you are additionally paying a premium for this privilege. Equally, whereas it’s cheaper than the 16GB mannequin, Nvidia has a VRAM benefit.

AMD may battle with the 7700 XT in comparison with a few of their RDNA2 fashions, that are nonetheless obtainable to buy. The RX 6800, particularly, is at present priced round $430 to $450, and in our testing, the 6800 is about 16 % quicker than the 4060 Ti in rasterization. This is able to point out related efficiency to the 7700 XT at a comparable or cheaper price whereas additionally packing extra VRAM, 16GB versus 12GB. So, that will likely be a really fascinating comparability for our closing evaluation to see exactly the place price-to-performance ratios lie, particularly if the 7700 XT cannot fairly attain these efficiency figures.

Our preliminary feeling is that the 7700 XT is simply too costly and needs to be a $400 GPU to suit with the remainder of the RDNA3 line-up and older RDNA2 fashions whereas additionally providing sturdy competitors to GeForce opponents. With that pricing, AMD may have claimed a quicker GPU with extra VRAM on the identical value because the RTX 4060 Ti 8GB mannequin. As a substitute, they’ve chosen to not take any huge benefit within the mid-range.

RX 7800 XT Efficiency Claims

The 7800 XT has been pitted in opposition to the RTX 4070 in AMD’s efficiency slide (beneath), exhibiting a four-percent achieve on common throughout these 19 video games. This breaks down right into a nine-percent benefit in rasterization and a six-percent deficit in ray tracing, which is a little more favorable than the 7700 XT, given that is what AMD is providing at a $100 cheaper price level, to not overlook the additional VRAM coming in at 16GB versus 12GB for Nvidia’s competitor.

It also needs to result in a extra vital efficiency hole in comparison with the RTX 4060 Ti 16GB on the identical value, on condition that our testing put the 4070 at 28 % quicker than the 4060 Ti, and AMD is claiming a slight lead over the 4070. Clearly, a thorn in AMD’s aspect on this value tier will likely be ray tracing efficiency, which AMD suggests is decrease than the 4070, and so they did not even check a few of the most intensive RT titles, reminiscent of Cyberpunk 2077.

The place this mannequin might be shaky is when evaluating it to the nearest-priced RDNA2 mannequin – the RX 6800 XT – which is at present obtainable for round $530. Our testing has the RTX 4070 and RX 6800 XT providing related rasterization efficiency, whereas AMD claims a nine-percent raster lead over the 4070. This does level to a greater price per body for the 7800 XT, nevertheless it hinges fully on the hole between it and the 6800 XT. If it is close to 10 %, as steered, that does give it an inexpensive benefit in worth to the newer 7800 XT, given it is also barely cheaper. Nevertheless, if the playing cards find yourself extra equal, it’s only a slight generational uplift. We’ll should see the way it all performs out in our complete evaluation.

A minimum of on the graphics card launch aspect of issues, what AMD is exhibiting seems cheap, not tremendous superb must-buys, not awfully insufficient fashions, however simply okay. Once more, the 7800 XT seems higher from what AMD has proven than the 7700 XT, however with each GPUs, it is actually going to come back right down to closing efficiency, options, and worth. There are various first rate choices on this value vary, together with the RX 6800 XT, RX 6800, RX 6700 XT, and possibly Nvidia’s finest RTX 40 collection GPU exterior the 4090 within the RTX 4070. One factor AMD has going for it with this launch is it might probably take the crown for the most cost effective present technology GPU to supply greater than 8GB of VRAM within the 7700 XT, succeeding the RTX 4060 Ti as a result of its $50 cheaper price tag and 12GB capability.

AMD Gives Extra FSR 3 Particulars

Throughout its Gamescom presentation, AMD additionally spent a while speaking about new options for Radeon GPUs and video games as an entire. One of many greatest ones is FSR 3, which is hotly anticipated after the corporate introduced it means again in November of 2022. We’re practically 10 months on from that announcement, and regardless of revisiting FSR 3 at Gamescom, AMD continues to be not able to launch this function and hasn’t put a agency launch date on it. The very best we may get from AMD is it can roll out in “early fall,” which may imply as early as September.

We additionally realized that opposite to latest rumors, AMD just isn’t aiming to have Starfield as their first FSR 3 launch title. Actually, this presentation didn’t point out that FSR 3 could be coming to Bethesda’s recreation in any respect. Starfield just isn’t listed on AMD’s future recreation help slide, and Bethesda just isn’t listed as a associate for FSR 3. AMD particularly advised us that Forspoken and Immortals of Aveum could be the primary video games to obtain FSR 3. That is a little bit of a blow, on condition that Starfield is at present AMD’s flagship sponsored title. Nevertheless, FSR 3 is coming to Cyberpunk 2077, which is nice to see given the approaching launch of that recreation’s growth, Phantom Liberty.

We did not study far more about FSR 3 that AMD hadn’t beforehand claimed. Their demonstrations present body technology expertise working as anticipated to roughly double the body price, as we anticipated primarily based on what the corporate stated final 12 months. As with DLSS 3 body technology, you aren’t getting exactly double the FPS, as seen beneath, however the expertise is exhibiting a big leap in smoothness with it enabled. AMD is branding its body technology tech as “Fluid Movement Frames.” When carried out in FSR 3, it makes use of body interpolation and movement vectors to generate these extra frames (fairly just like DLSS 3, by the sounds of issues).

The place FSR 3 will differ from DLSS 3 is in {hardware} help. FSR 3 will likely be obtainable for many graphics playing cards. Once we requested, AMD stated help goes again to RDNA1-based GPUs and “a broad vary of competitor options, together with NVIDIA GeForce RTX 20, 30, and 40 Sequence graphics playing cards.” Nevertheless, it additionally really useful utilizing FSR 3 on RDNA2 or RDNA3 GPUs, suggesting the tech is optimized for newer architectures and should endure on these older playing cards. It is unclear whether or not FSR 3 will likely be blocked fully from working on GPUs that are not on this supported record, for instance, the trusty outdated RX 580; it might be a scenario the place the structure of older fashions does not embody key performance that FSR 3 requires.

We requested AMD whether or not FSR 3 makes use of FSR 2 for temporal upscaling or whether or not the upscaling element has been upgraded.

“FSR 3 contains the most recent model of our temporal upscaling expertise, which has been optimized to be totally built-in with the brand new body technology expertise. Nevertheless, our primary focus with AMD FSR 3 analysis and improvement has been on creating high-performance, high-quality body technology expertise that works throughout a broad vary of merchandise.”

So it does not sound just like the temporal upscaler has been upgraded, however it can use the most recent model. AMD additionally stated it has optimized FSR 3 to reconstruct recreation UIs, saying that “UI processing is an integral a part of our resolution to reduce any impression body technology can have on it.”

Alongside FSR 3, AMD will likely be integrating body technology into their driver as a standalone Fluid Movement Frames function, just like what it did with Radeon Tremendous Decision, bringing FSR 1 into the motive force. It will provide broad recreation compatibility and assist you to use a really rudimentary model of body technology in any title in your Radeon GPU.

All of it sounds good in idea, and we’re certain many will remark about how this can put AMD forward within the “generated frames” race, and loads of fanboy wars will ensue. Nevertheless, primarily based on what we have seen with DLSS 3 body technology, recreation integration and using movement vectors are essential to producing visually first rate frames. The body interpolation element is commonly the weak hyperlink to border technology, inflicting blurry, garbled messes each second body. With DLSS 3, normally solely components of the display screen should depend on interpolation, with the remainder being processed primarily because of movement vector and recreation engine information (e.g., shift this whole body barely to the left for the subsequent generated body).

With AMD saying that Fluid Movement Frames within the driver is simply the movement interpolation element – after all it’s; the motive force can not entry information reminiscent of movement vectors – we have now fairly substantial query marks over the visible high quality of this function. With out that essential movement information, is your complete body simply going to be a blurry, smeared mess mixing two full frames collectively? If it is something like what we have seen from interpolation in DLSS 3 however now utilized to your complete generated body each single time, it may look terrible, and we’re undecided it is a function that will likely be value utilizing. So, that one would require some fairly heavy scrutiny earlier than we get enthusiastic about it.

New Anti-Lag Plus Characteristic for RDNA3

Latency is the place it will get a bit difficult. You may need noticed on these FSR 3 slides one thing known as “Anti-Lag+,” a brand new function being introduced. It’s separate from FSR 3’s built-in latency-reducing expertise. So, to be clear, while you allow FSR 3, it can robotically apply latency-reducing tech, however then Radeon house owners can use Anti-Lag+ on prime of this on the driver stage. AMD’s slides make it seem to be Anti-Lag+ is a part of FSR 3, like how Nvidia Reflex is a part of DLSS 3 and is enabled when body technology is toggled on. Nevertheless, AMD clarified this is not the case.

Anti-Lag+ is a brand new driver-level latency-reducing function obtainable solely for RDNA3 GPUs and presumably the rest shifting ahead. It is going to be obtainable alongside the outdated Anti-Lag however guarantees some enhancements. AMD defined it to us as follows:

“AMD Radeon Anti-Lag helps synchronize CPU and GPU processing throughout gameplay to cut back enter latency. Anti-Lag’s synchronization level, nonetheless, is positioned at a hard and fast location within the rendering pipeline. AMD Radeon Anti-Lag+ additional reduces latency via clever synchronization placement that considers the holistic rendering pipeline. Consequently, Anti-Lag+ can additional cut back latency the place it issues most.”

AMD did not go into specifics, however this appears extra akin to Nvidia Reflex than their earlier Anti-Lag expertise. Reflex intelligently adjusts to reduce latency within the pipeline, successfully offering the advantages of a dynamic body cap to barely alleviate any GPU bottleneck, a confirmed method for bettering latency. Nvidia’s expertise is built-in into the video games and is claimed to incorporate optimizations on a per-title foundation. We do not know if Anti-Lag+ works like this but, and implementation does differ in that it is driver-side as an alternative of game-side. Nevertheless, the best way AMD has defined it makes it sound prefer it’s attempting to attain one thing related. Hopefully, it is like Reflex, an amazing expertise that works properly, and up thus far, AMD did not have a competitor for it.

It is also unclear whether or not Anti-Lag+ will work throughout all video games or only a number of titles. AMD confirmed that it’s not built-in into video games however is driver profile-based. The corporate appeared to recommend just some titles will help Anti-Lag+. We’ll should see how that each one works when it is built-in into the motive force.

How Anti-Lag, Anti-Lag+, and FSR3 intertwine is a bit complicated, particularly regarding how DLSS3 and Reflex function on the Nvidia aspect. So right here is one of the best clarification of the way it all works:

AMD’s FSR3 body technology expertise has built-in latency-reducing expertise that works throughout all GPUs. Avid gamers utilizing an RX 7000 collection GPU can go into the motive force and allow Anti-Lag+ on prime of FSR3’s latency-reduction tech in supported titles. Anti-Lag+ just isn’t enabled by default in FSR3 titles; it is an elective function that customers should activate manually. House owners of older Radeon GPUs can apply Anti-Lag (the non-plus model) on prime of FSR3’s latency-reducing tech as an alternative, once more an elective function.

“The built-in latency discount expertise in AMD FSR 3 will present worthwhile latency discount when utilizing FSR 3 body technology, and our in-driver latency discount options are a bonus that stacks on prime to offer the very best expertise when utilizing FSR 3,” AMD defined.

As for why Anti-Lag+ is simply supported on RDNA3, AMD stated, “Anti-Lag+ [relies] on Mental Property that was launched with AMD RDNA 3 and isn’t obtainable on earlier generations of {hardware}. At AMD, we at all times intention to help earlier generations of {hardware}. […] Whereas we’re investigating the power to do […] Anti-Lag+ in earlier {hardware}, we can not commit that that is potential.”

We wish to know precisely what mental property is restricted to RDNA3 that’s required for this function, however we’ll have to attend to search out out.

For normal FSR 2 customers, AMD is releasing a brand new native anti-aliasing mode, which lets you use FSR 2 expertise with none upscaling utilized. That is successfully what Nvidia is doing with DLAA. So, it offers AMD parity with Nvidia in video games that combine this function. This native AA mode will present players with the next high quality anti-aliasing method than many video games’ built-in TAA resolution, at some price to efficiency, whereas additionally that includes FSR’s built-in sharpening. Many individuals have been requesting one thing like this, so AMD is now delivering it.

HyperRX

Lastly, we have now Hypr-RX, AMD’s one-click toggle to use all its performance-enhancing options. First introduced means again with RDNA3 final 12 months, this function is lastly prepared and will likely be obtainable natively in some video games and as a driver toggle. Basically, it goals to make it simple for novice customers to allow options like Radeon Enhance, Anti-Lag, Tremendous Decision, and even Fluid Movement Frames by hitting one button. In-game implementations will use FSR when Hypr-RX is on, whereas the motive force toggle will use Radeon Tremendous Decision. Avid gamers can then go in and additional customise what options are enabled, so they may activate Hypr-RX after which disable Enhance or RSR if they do not like these options.

Hypr-RX is the primary time that each one of those options will be enabled collectively. Beforehand, you could not use Enhance and Radeon Tremendous Decision concurrently. AMD has labored round that limitation, the caveat being that, like Anti-Lag+, Hypr-RX is simply supported on RDNA3 GPUs and newer. AMD provided the identical technical purpose for this as with Anti-Lag+, which appears fairly questionable.

We do not assume Hypr-RX will likely be a necessary function for many fanatic players, as they will most likely wish to tweak every of AMD’s options individually.

To recap: The RX 7800 XT and RX 7700 XT are approaching September sixth, priced at $500 and $450, respectively. AMD FSR 3 is sort of prepared and will likely be debuting in early fall, beginning with Forspoken and Immortals of Aveum. Anti-Lag+, Fluid Movement Frames, and Hypr-RX are additionally coming to AMD’s Radeon driver shortly.