Fanatic-Class RDNA3 For 1440p Gaming

With the Gamescom conference happening in Germany this week, AMD is utilizing Europe’s largest online game commerce present because the venue for his or her newest Radeon video card bulletins. This morning the corporate is asserting their long-awaited center members of the Radeon RX 7000 collection, the Radeon RX 7800 XT and Radeon RX 7700 XT. Aimed on the 1440p gaming market and primarily based on AMD’s new Navi 32 GPU, the brand new playing cards are designed to fit in to the center of AMD’s product stack, providing a set of potent RDNA 3 structure video playing cards for avid gamers who don’t want the bleeding-edge efficiency (and wallet-bleeding prices) of the Radeon RX 7900 collection playing cards.

With immediately’s announcement being simply that – an announcement – the retail launch of those playing cards, as well as AMD’s new Hypr-RX enabled drivers, will observe in a few weeks, on Wednesday, September 6^th. That date additionally not-so-coincidentally occurs to be the discharge date for Bethesda’s ARPG Starfield, for which AMD is the sport’s unique PC {hardware} associate, and which AMD shall be providing with the brand new Radeon playing cards as a part of the corporate’s newest recreation bundle. So for AMD, the celebs are aligning to make September 6^th a giant day for the corporate’s GPU division.

However as for us, we’re right here to speak about {hardware}, so let’s talk about the Radeon RX 7800 XT and 7700 XT, in addition to the Navi 32 GPU that underpins them.

For this technology of video playing cards, AMD opted to eschew the same old top-to-bottom GPU and video card technique in favor of releasing the highest video playing cards (Radeon RX 7900 XTX & 7900 XT), and following that up with the underside of the stack (Radeon RX 7600). Consequently, that is a kind of uncommon launches the place the center components, broadly encompassing the “fanatic” section, are the ultimate section to launch. The exact order is usually immaterial, nevertheless it means we’re coming into the RX 7800 XT launch already realizing how the AMD playing cards above it and beneath it carry out – and thus the place it must be to greatest slot in to the product stack.

And with rival NVIDIA having already launched their full stack of video playing cards for this technology, AMD additionally will get the benefit of realizing the place to greatest place their playing cards towards the competitors. On this case, AMD is positioning the Radeon RX 7800 XT roughly towards the GeForce RTX 4070, whereas its smaller sibling, the Radeon RX 7700 XT, will go up towards the GeForce RTX 4060 Ti. NVIDIA has already proven their hand for 1440p gaming, so with this launch it’s as much as AMD to resolve the best way to greatest beat it.

Radeon RX 7800 XT & Radeon RX 7700 XT: By The Numbers

Main issues off is the Radeon RX 7800 XT, which AMD shall be positioning as their premiere 1440p gaming card. The RX 7800 XT has a near-full Navi 32 configuration (extra later) with 60 CUs built-in, affording 7680 ALUs (3840 stream processors). AMD is clocking this half rather less aggressively than what we’ve seen with the flagship RX 7900 XTX, giving us a peak enhance clock of 2430MHz and a median recreation clock of 2124MHz. On paper this offers the cardboard a median FP32 compute throughput of 32.6 TFLOPS, although as has been the case with the remainder of the RDNA 3 household, the architectural adjustments make throughput comparisons towards prior AMD playing cards difficult.

By the numbers, the RX 7800 XT presents about 76% of the compute throughput of the subsequent card up, the RX 7900 XT. Notably, nonetheless, the variety of ROPs has been halved versus each RX 7900 collection playing cards, so the precise rendered pixel throughput is nearer to 50%. Few video games are completely ROP-bound, so the precise efficiency hole between this card and the 7900 XT shouldn’t be almost as large, nevertheless it underscores the truth that we’re not taking a look at one thing as easy as 75% of a 7900 XT.

With 4 lively MCDs, the RX 7800 XT is connected to a 256-bit GDDR6 reminiscence bus. Whereas we’ve seen the 7900 collection playing cards clocked at 20Gbps, AMD has taken the odd option to dial down the clocks on the 7800 XT only a bit, to the noticeably fractional 19.5Gbps. This offers the ensuing card a complete of 624GB/second of reminiscence bandwidth. That, in flip, is additional augmented with 64MB of Infinity Cache unfold among the many MCDs, two-thirds of the cache discovered on AMD’s flagship playing cards.

The usage of a 256-bit reminiscence bus implies that AMD is well in a position to equip the cardboard with 16GB of VRAM, which goes to be a significant advertising and marketing level from the corporate in competing with the GeForce RTX 4070 and its 12GB of VRAM sitting on a 192-bit reminiscence bus. Consequently, AMD has the sting in each reminiscence capability and reminiscence bandwidth, although as we’ve seen up to now, that gained’t essentially translate right into a efficiency benefit for AMD.

In the meantime, with the general design of Navi 32 favoring value over energy effectivity, the ability consumption of the 7800 XT goes to run at a comparatively excessive 263 Watts, 37W lower than the 7900 XT. AMD shouldn’t be going to have the ability to win the ability effectivity race with NVIDIA right here (and arguably they by no means even supposed to attempt) so as a substitute the RX 7800 XT is configured to maximise its efficiency. Word that these are AMD’s baseline specs, and manufacturing unit overclocked AIB associate playing cards can go increased.

Making generational comparisons here’s a bit difficult, each as a result of architectural adjustments from RDNA 2 to RDNA 3, in addition to some quirks in AMD’s personal card positioning for this technology. Whereas the RX 7800 XT is AMD’s 3^rd/4^th tier card total (behind the three 7900 collection playing cards), its direct predecessor, the RX 6800 XT, was AMD’s second-tier card, behind solely the flagship 6900 XT. Lengthy story quick, the 6800 XT/7800 XT comparability isn’t particularly flattering for AMD – the variety of CUs has gone down by a bit – although so has energy consumption and worth. The extra apt comparability is arguably the RX 6800 vanilla, however I digress.

In any case, the large generational improve push from AMD goes to be pitting the RX 7800 XT towards playing cards that are actually developing on 4 years previous, such because the Radeon RX 5700 XT and GeForce RTX 2070. The RX 7800 XT ought to roughly double the efficiency of these playing cards, making them a extra palatable improve for system house owners whose present 1440p playing cards are now not maintaining.

Sapphire Pulse RX 7700 XT

Lastly, standing in line behind the RX 7800 XT, we’ve got the RX 7700 XT. This can be a cut-down Navi 32-based card that, regardless of the decrease collection quantity, can also be aimed toward 1440p gaming.

{Hardware}-wise, we’re taking a look at 54 CUs, with a most enhance clock of 2544MHz and a median recreation clock of 2171MHz. Consequently, the compute efficiency of the RX 7700 XT is, on paper, fairly near its greater brother, with the smaller card slated to ship 92% of the RX 7800 XT’s compute efficiency. And, notably, there isn’t any reduce to the variety of ROPs right here, so the general pixel pushing energy is technically ever so barely within the RX 7700 XT’s favor.

As an alternative, what shall be driving the majority of the efficiency variations between the 2 playing cards shall be reminiscence bandwidth. The RX 7700 XT solely options 3 lively MCDs (just like the 7900 XT, the inactive MCD is stored as a dummy spacer), which suggests it’s dropping 1 / 4 of its reminiscence bus, in addition to 1 / 4 of its Infinity Cache. Coupled with that, AMD has reduce the reminiscence clocks a bit extra, bringing them all the way down to 18Gbps/pin. The web result’s that the RX 7700 XT presents 69% of the RX 7800 XT’s reminiscence bandwidth, and 75% (48MB) of the Infinity Cache capability. All of which implies that it will likely be simply how reminiscence and cache bandwidth-bound a given recreation is that determines how shut the RX 7700 XT can get to the RX 7800 XT.

The smaller reminiscence bus additionally implies that the cardboard comes with much less VRAM total. On this case, we’re taking a look at 12GB of GDDR6. With AMD positioning this card towards NVIDIA’s 128-bit GeForce RTX 4060 Ti, meaning they’ve a reminiscence capability and bandwidth towards the baseline 8GB mannequin, although NVIDIA’s companions are additionally providing a 16GB mannequin that leaves the RX 7700 XT smack dab within the center.

In any other case, with clockspeeds working even increased than the RX 7800 XT, the RX 7700 XT picks up little or no in energy financial savings. AMD has the cardboard’s official TBP at 245W, simply 18W lower than the RX 7800 XT. Assuming what real-world energy consumption matches this determine (which isn’t all the time the case), then it implies that the RX 7700 XT goes to be much less energy environment friendly of the 2 playing cards total. And, for that matter, far much less energy environment friendly than NVIDIA’s 160W RTX 4060 Ti.

The RX 7700 XT shall be changing the RX 6700 XT inside AMD’s product stack, which on the time of its introduction, was the highest Navi 22 card. Whereas we don’t have any comparative efficiency figures, the RX 7700 XT total enjoys a 35% bump in CU depend, albeit at a ~10% regression in common clockspeeds. All of the whereas energy consumption has trickled up a bit from 230W. So it ought to make for a modest generational soar over its direct predecessor.

Navi 32: A Second, Smaller Chiplet GPU

Driving each playing cards is AMD’s new RDNA 3 architecture-based Navi 32 GPU. That is the second client chiplet GPU from the corporate, and is actually a smaller tackle Navi 31 that powers the RX 7900 collection. On this case “smaller take” implies that AMD is utilizing a brand new Graphics Compute Die (GCD) with fewer useful blocks and transistors, buying and selling off efficiency for decrease energy consumption and manufacturing prices. In the meantime the GCD is being paired with 4 of AMD’s Reminiscence Cache Dies (MCD), that are being reused from Navi 31.

As with Navi 31, the intricate GCD is fabbed on TSMC’s 5nm course of, constructing the scalable and efficiency important components of the GPU on a modern TSMC course of. In the meantime the MCDs proceed to be produced on TMSC’s 6nm course of, housing components of the GPU that don’t scale properly with smaller nodes, such because the Infinity Cache and GDDR6 reminiscence controllers. So Navi 32 is a mixture of new and previous(er) course of nodes close to its fabrication, whereas it’s on the slicing fringe of packaging.

Broadly talking, the whole Navi 3x GPU lineup is designed to be cost-efficient from a producing standpoint, and Navi 32 is not any totally different. AMD is barely producing the important bits on a forefront 5nm node – and even then, not utilizing TSMC’s bleeding-edge 4nm variant – whereas producing the remainder on a less expensive 6nm node. With out going too far down the rabbit gap for this announcement, this has efficiency implications that change into a think about how AMD configures and positions their playing cards. However at a excessive stage, the takeaway is that the novelty of a chiplet-based GPU apart, Navi 32 follows a notably totally different design technique than previous enthusiast-class AMD GPUs, or for that matter rival NVIDIA’s monolithic designs. As we’ve seen with the RX 7900 collection, that is going to be probably the most noticeable in energy consumption, and is probably going the largest think about why each Navi 32 playing cards have TBP specs properly above their NVIDIA competitors.

For now, we’re nonetheless ready on AMD to verify a number of the specs of Navi 32, specifically the die dimension of the GCD and its transistor depend. Based mostly on AMD’s off-angle die picture, we’re at the moment estimating a tough 211mm² primarily based on the identified dimension of the MCDs (37.5mm²), however that’s a tough guess at greatest. That might put it at round two-thirds the scale of the Navi 31 GCD.

At this level we’re additionally awaiting affirmation on the overall variety of CUs accessible inside the GPU. Whereas the RX 7800 XT is the highest tier half, AMD has not confirmed whether or not Navi 32 itself additionally solely has 60 CUs, or if it has extra. On condition that Navi 31 was 96 CUs and Navi 33 was 32 CUs, it’s probably that Navi 32 is a 64 CU design, however stranger issues have occurred.

In any other case, Navi 32 is RDNA 3 by means of and thru. This implies it incorporates the identical architectural adjustments we’ve seen in the remainder of the RX 7000 collection lineup, akin to AV1 encode/decode help, improved ray tracing efficiency, AMD’s new DisplayPort UHBR13.5-capable Radiance show controller, and the second financial institution of ALUs for dual-issuing appropriate directions and bettering AI efficiency.

Reference Playing cards, Companion Playing cards, & Product Positioning

As famous earlier, each playing cards shall be launching in a few weeks, on September 6^th. Almost about reference playing cards, AMD is actually splitting the distinction – they are going to be promoting a reference Radeon RX 7800 XT by means of their web site, however not a reference Radeon RX 7700 XT. As an alternative, AMD’s board companions shall be swooping in to help the launch of each playing cards, with AIB designs slated to be accessible for each playing cards on the launch day.

Taking a quick take a look at AMD’s reference card, it’s a reasonably sizable twin-fan open air design, copying the design aesthetics from AMD’s reference RX 7900 collection playing cards. At two followers it’s smaller total, however not immensely so. Notably, this can be a 2.5 slot card, which not the entire photos present very properly, so it would want some further room to breathe.

Powering the cardboard is a pair of 8-pin PCIe energy sockets. Strictly talking, that is technically overkill, since a 6+8 configuration would nonetheless be good for 300W. However an increasing number of we’re seeing 6-pin sockets go away completely, so it’s not completely stunning to see AMD go for 8+8. As famous earlier, the TDP of the reference card is 265W, nonetheless AIB companions can (and can) clock a few of their playing cards increased with increased TDPs to match.

In the meantime, AMD’s port configuration has taken a little bit of a departure from the RX 7900 reference designs. Whereas nonetheless providing 4 ports, AMD has thrown out the USB-C port in favor of one other full-size DisplayPort. The result’s a trio of DisplayPorts that help DP 2.1 at knowledge charges of as much as UHBR13.5, in addition to a sole HDMI port supporting HDMI 2.1.

Sapphire’s Nitro RX 7800 XT  & RX 7700 XT

As for the associate playing cards, the entire regular suspects have RX 7800/7700 designs within the works, together with ASUS, GIGABYTE, Sapphire, PowerColor, and extra. AMD despatched over a few dozen photographs displaying off the totally different designs, loads of that are triple-fan playing cards that seem like their current RX 7900 designs.

Sadly, most of those photographs are head-on pictures, so we are able to’t inform whether or not many (if any) are true dual-slot designs (versus the two.5 slot design of the AMD reference card). Because of the comparatively excessive TBPs, the RX 7800/7700 collection places AMD and its companions in a little bit of a troublesome place in relation to producing SFF-sized playing cards. We’ll have extra data as soon as the board companions publish their official product pages, however at first blush, it doesn’t seem like we’ll see a lot in the best way of small playing cards – a minimum of initially.

When it comes to competitors, AMD is establishing the RX 7800 XT and RX 7700 XT roughly towards the GeForce RTX 4070 and GeForce RTX 4060 Ti respectively – with AMD believing they’re within the lead, in fact. It’s a little bit of an odd match-up because the RTX 4060 Ti shouldn’t be primarily aimed toward 1440p gaming the best way the RX 7700 XT is, however that is hardly the primary time AMD and NVIDIA’s playing cards don’t completely line up.

Simply how properly they examine, nonetheless, stays to be seen. Not only for impartial efficiency benchmarks, however as a result of AMD shouldn’t be disclosing the value of both card to the press forward of time. So whether or not AMD goes for iso-pricing or in the event that they intend to undercut NVIDIA in some style stays to be seen.

Replace: Now that the official announcement has gone out, AMD has disclosed the value of the playing cards. The Radeon RX 7800 XT will begin at $499, whereas the Radeon RX 7700 XT will begin at $449. The remainder of the unique article follows as beneath.

However, in an indication that AMD probably isn’t going to be too aggressive on pricing, the corporate is extending their present Starfield promotional recreation bundle to cowl the brand new Radeon playing cards; so each the RX 7800 XT and RX 7700 XT shall be coming with a recreation bundle proper off the bat. Starfield is AMD’s large recreation partnership for the 12 months, so the corporate is eager to advertise what they see as one of the crucial vital video games of the 12 months. On the identical time, recreation bundles are historically used to prop up video card costs in lieu of worth cuts by including worth to the general bundle.

In the end, simply how profitable the Radeon RX 7800 XT and RX 7700 XT are shall be completely in AMD’s palms. RDNA 3 is a stable structure, and due to the corporate’s cost-optimized chiplet-based design technique, they shouldn’t have any hassle hitting worth factors that make each playing cards aggressive. The one actual query is simply what number of playing cards AMD needs to promote – and thus what worth factors they aim.

Wrapping issues up, the Radeon RX 7800 XT and RX 7700 XT shall be launching on September 6^th. AMD shall be carrying their reference RX 7800 XT instantly on their web site, in the meantime Newegg, Amazon, and different retailers and etailers shall be carrying the varied AIB playing cards that the companions have put collectively.