AMD, starting with an extremely busy summer for AMD, which has seen some of the most important product launches in recent history for both processors and processors, AMD is preparing to end the year also on a high note. This morning, the company is announcing the next series in its Radeon RX 5000 camcorder, the Radeon RX 5500 series. Based on the same RDNA architecture as the flagship part of the AMD RX 5700, the new RX 5500 series is built around a new, smaller Navi GPU that will allow AMD bring your newest Radeon technologies to the more affordable, basic graphics cards, and for the first time.
The AMD announcement corresponds to the usual cascade of GPU products that we see in every generation. With the launch of the next-generation GPU router ̵
As is often the case, AMD goes straight down the stack. After launching this summer in the Radeon RX 5700 series – and the fanfare that follows it – AMD is now turning its attention to the market below. Even the Radeon RX 5700 (vanilla) card, which starts at $ 349, has plenty of room under the RX 5700 for AMD to offer cheaper cards to a wider audience, and that's what AMD aims to do. New details of the Radeon RX 5500 series will focus on 1080p games, bringing RDD architecture and AMD features to the popular high-volume gaming market.
However, with everything in mind – and before we get to the specifications and other technical issues – today's announcement needs to be put in context, since today is not a retail launch of a video card. In fact, AMD doesn't even announce when retail cards will be available outside of the clear fourth quarter. Instead, today's message is essentially about the OEM page of AMD. With the holiday season unfolding, OEM partners need to announce their holiday system lineup to AMD, which is why AMD announces almost everything about the Radeon RX 5500 series, not OEM companies to steal AMD's thunder.
To be sure, AMD provides us with many details, including information about the new GPU, clock speed, TDP, and more. But the specifications here are for OEM products, not retail, and AMD designs them accordingly. In general, in this segment of OEM and retail products are not too different – so do not be surprised if you see the Radeon RX 5500 (vanilla) retail card further down the line with the specifications (if not slightly better than) the OEM card – but context matters , especially since OEMs sometimes announce things.
Meanwhile, not to be outdone, today's announcement also covers a new mobile part, the Radeon RX 5500M. We'll get to that part a little bit later, but it's mentioned earlier, as it's part of a bigger shift in product naming for AMD. In short, after spending the last generation on only one product family for both desktop and mobile devices, discrete mobile devices have returned . The Radeon RX 5500M is the first of these products, and although it is based on the same GPU as the desktop RX 5500, it has its own speed, performance and TDP features. None of this really changes how manufacturers build their laptops, but if nothing else, hopefully it will become clearer how their laptops are configured and run.
|AMD Radeon RX Series Specification|
| AMD Radeon|
| RX 5500 Series (OEM)||AMD Radeon RX 5700||AMD Radeon RX 480||AMD Radeon RX 560|
( 2304 SP)
|Textural units||88||144||144||64  ROPs||32?||64||32||16|
|Basic clock||?||1465 MHz||1120 MHz||1175 MHz|
|Game Clock||Game Clock||= 1717MHz||1625MHz||N / A||N / A||N / A||Clock||<= 1845MHz||1725MHz||1266MHz||1275MHz|
|Bandwidth (FPP) <= 5.2 TFLOP||7.95 TFLOPs||5.1 TFLOPs [1 9659015] 2.6 TFLOP|
|Memory Clock||14 Gbps GDDR6||14 Gbps GDDR6||8 Gbps GDDR5||7 Gbps GDDR5|
|Memory Bus Width||bit bit||256 bit||128 bit|
|VRAM||4GB / 8GB||8GB||8GB||4GB|
|Count Transistor||6.4B||10.3 B||5.7B||3B|
|Typical Board Power||150W||180W  4] 150W||80W|
|Manufacturing Technology||TSMC 7nm||TSMC 7nm||TSMC 7nm||14nm||GloFo 14nm|
|Architecture RD (19659012]||RDNA (1)||GCN 4||GCN 4|
|GPU||Navi 14||Navi 10||] Polaris 10||Polaris 11|
|Launch date||Q4 2019||07/07/2019||06/29/2016||05/2017|
|Start price||TBD  $ 349||$ 239||$ 99|
Starting from card OEMs of desktops, we have what AMD calls the Radeon RX 5500 Series. Based on the new Navi 14 GPU, these cards are market-oriented 1080p, and in terms of performance, these are greatly reduced in version of AMD Navi 10 cards, Radeon RX 5700 AMD series.
As for the CU, all of these cards will ship with 22 CUs, 14 CUs less (-39%) than the next card on the AMD stack, the Radeon RX 5700. However, as the CU count goes down, the clock speed will be slightly up , with the RX 5500 series carrying a maximum clock speed of 1717 MHz, 92 MHz (6%) higher than the RX 5700. Similarly, the maximum clock speed is higher up to 1845 MHz. That being said, since we're talking about OEM cards, the usual disclaimer applies: it's "up to" the frequency, and OEMs can pick up the clock speed because of power and other reasons. Officially, there are no minimum clock speeds for OEM configurations, so it remains to be understood by the original manufacturers.
In terms of overall GPU performance, then at full clock speed the RX 5500 Series will have a peak computing / shader bandwidth of 5.2. TFLOP. This is 35% below the maximum speed of the RX 5700, and is fairly typical of the type of jump we see between the tiers of maps.
Meanwhile, as far as memory is concerned, the RX 5500 series also receives a narrower memory bus. . If it comes with the RX 5700 series and its 256-bit bus, the AMD cards will come with a 128-bit memory bus instead. However, this is only a change for memory; AMD will also use GDDR6 for these maps. So, although the effective drop in memory bandwidth relative to the RX 5700 is significant – from 448GB / s to 224GB / s – it will be the first time GDDR6 has appeared on the GPU. As a result, the RX 5500 Series will still have access to considerable memory bandwidth, and in this respect it should be of paramount importance over its previous 128-bit bus cards. As for the total storage capacity, the cards will support up to 8 GB of VRAM, but this is another "up" specification, and OEMs may also decide to supply some systems with a configuration of 4 GB.
Finally, until it was listed on the official AMD table, the company told me that the TGP for these cards would be 150 watts. This surprised me a little, and I'm curious to see if AMD is simply conservative here for OEMs, or whether real world energy consumption will really be that high. To put that in perspective, the Radeon 5700 (vanilla) is a 180W card, so a 150W TGP is only about 17% lower than the next card on the AMD stack, and the performance difference is what I expect. will be bigger. Or, looking at the latest generation cards, it would give the RX 5500 TGP series comparable to the Radeon RX 480 and 570. If these TGPs are accurate, this could be another AMD case that contributes to absolute performance over productivity for their mainstream parts stream , which was also the case with the upscale Radeon RX 500 previous generation cards.
In terms of performance expectations then, AMD offers only a few numbers. The company claims that the new RX 5500 cards still deliver better performance than their RX 5700 cards, and 1.6 times better performance per watt compared to the latest generation AMD cards. Comparing the new card to the RX 480, AMD notes a 30% drop in power consumption along with a 12% increase in absolute productivity. Note, however, that AMD data is literally referred to as a "GPU", which means that we only look at GPU numbers, not whole cards.
Elsewhere, AMD places game figures in the latest 1080p Medium games, getting over the RX 480 in the 15-20% range.
Even a 12% increase in the absolute performance of the RX 480 will mean that the RX 5500 series cards will work very much like the higher-end parts of the Radeon RX 500 AMD series. The black sheep of this family, the RX 590, performed 15% better than the RX 480, so if the AMD figures are accurate, then the RX 5500 will land between the RX 580 and RX 590, and that's not enough. Which, if AMD can deliver that 150W of performance, would be a significant improvement over the powerful RX 590.
Otherwise, the rest of AMD's performance ratings are at odds with the GeForce GTX 1650 NVIDIA, which of course AMD easily wins. However, without the pricing information, I don't read too much of these figures, as the GTX 1650 is a $ 149 card, and I'm surprised AMD will go down quite low when they add 30-35% performance advantage. .
Navi 14: Navi but smaller
Now that we've seen the specifications for the OEM desktop cards, let's take a brief look at the Navi 14. GPU itself is designed to power AMD RX 5500, Navi 14 desktop and mobile parts is a very typical small version of the advanced AMD technology.
As for the size of the dies and the number of transistors, AMD dropped about 3.9V transistors from the Navi 10, which is 38% less than transistors overall. This means that the die size is 158 mm2, almost perfect for 37% less than the Navi 10. Compared to previous generations of AMD, comparisons are a bit more complicated as performance goals and die sizes have increased overall, but this fits for a transistor by about 12 % higher than Polaris 20.10 30, only 70% of the place. Polaris 11 (used in RX 460/560) might be a more appropriate comparison here; in this case, AMD increased the number of transistors by 113%, while increasing the size of the dies by 28%.
These enhancements come from a 7nm TMSC process, of course. AMD's heart is a whole stack of next-generation products in CPUs and GPUs for 2019, and AMD is re-using this process for its latest Navi navigation platform. This remains a major competitive advantage for AMD, as NVIDIA still uses TSMC 12nm, a 16nm process, for its GPUs. This gives AMD the ability to install smaller transistors operating at lower power levels, and the trade-off is that 7nm is a younger and more expensive process. And according to the reports, the TSMC client is joking to ensure that waffles are launched because of very high demand.
|Navi 14||Navi 10||Polaris 10  Polaris 11|
|Memory Bus Width||128-bit||256-bit||256-bit||128-bit|
|6,4B  10.3B||5.7B||3B|
|Manufacturing process||TSMC 7nm||TSMC 7nm||GloFo 14nm||GloFo 14nm|
|Architecture||RD 19659015] RDNA (1)||GCN 4||GCN 4|
Drilling, according to the AMD cards of the RX 5500 series do not use the fully integrated graphics processor N avi 14. The company does not disclose exactly what the GPU is all about, but the RX 5500 and 22 CU are not. Given that we know about die sizes, AMD couldn't merge too many units here, so it's true that the full Navi 14 GPU comes with a 24 or 26 CU. Assuming that the RX 5500 series comes to the retailer as a multi-card product, and that the OEM RX 5500 series becomes the RX 5500 (vanilla), then using a disassembled part means that AMD can use the fully-accessible part for the Radeon RX 5500 XT if needed.
Otherwise, the Navi 14 is a full-featured Navi GPU in terms of architecture and features. AMD has confirmed that it is a complete RDNA architecture (1), which means that it includes PCIe 4.0 support as well as new display processing units (Radeon Display Engine) and video processing (Radeon Media Engine). The display unit, in turn, shrinks slightly from 6 pipelines to 5 – meaning that the Navi 14 can support up to 5 monitor streams – but otherwise the same with DisplayPort 1.4 support, stream compression, and more. In the meantime, the video processing unit remains completely unchanged from the Navi 10 and should offer the same encoding and decoding performance as high-end AMD cards. So, if AMD really values the Radeon RX 5500 series against the NVIDIA GeForce GTX 1650, expect AMD to use it regularly against NVIDIA, since the GTX 1650 (TU117) comes with an older, less advanced Volta video unit.