ASUS ROG Astral GeForce RTX 5080 OC graphics card review: the first space speed?

They cause both delight and disappointment. They are argued about, tried to ignore, dream of buying and calculate how many game consoles can be bought for their price. However, they do not leave indifferent any active PC user. We are talking, of course, about top-end video cards from NVIDIA. This year, the developers introduced the GeForce RTX 50 series based on GPUs with Blackwell architecture. As usual, the flagship models were the first to enter the market, so we start our practical acquaintance with the new graphics adapters with the GeForce RTX 5080. And here the original ASUS ROG Astral GeForce RTX 5080 OC will help you get into the spirit of new top solutions.

GeForce RTX 5080 lineup

The GeForce RTX 5080 is the second-highest-end graphics card in NVIDIA's new graphics lineup. The model is equipped with a GB203 graphics processor based on the Blackwell architecture in a modification with 10,752 active compute units. The new GPUs are manufactured using TSMC's 5-nanometer process with some additional optimizations according to NVIDIA's requirements.

In fact, these are the same standards that were used for the previous generation of chips based on the Ada Lovelace architecture. Therefore, it is not surprising that the physical characteristics of the new crystal are almost identical to those of the AD103 chip. The area of the GB203 crystal is 378 mm², the total number of transistors is 45.6 billion. The density of the elements is 120.6 million/mm². So, the quantitative indicators here have almost not changed.

In the case of the GeForce RTX 5080, the GB203 processor received a frequency formula of 2295/2617 MHz. The video card retained a 256-bit bus and is equipped with 16 GB of memory. As for the type of memory, here, instead of GDDR6X, new GDDR7 standard chips are used. Moreover, the RTX 5080 has a maximum speed mode with an effective frequency of 30,000 MHz (the RTX 5090 has 28,000 MHz). The combination with a 256-bit bus allows you to get a bandwidth of 960 GB/s. This is 30% more than the same GeForce RTX 4080 SUPER.

The claimed power consumption level of the GeForce RTX 5080 is 360 W. This is more than its predecessor, the RTX 4080 SUPER, which has a TGP of 320 W.

The GeForce RTX 5080 is offered by the developer in the reference Founders Edition version, which is almost identical in appearance to the flagship GeForce RTX 5090. A compact 2-slot design with a miniature board and a pair of fans providing dual through-flow. Of course, NVIDIA partners have prepared many of their own original developments.

The recommended price of the GeForce RTX 5080 is $999, but even for the American market, this MSRP is rather arbitrary. Most models from manufacturers initially have significantly higher price tags. In addition, against the background of a certain shortage of GPUs for the new line, the existing shortage provokes an additional increase in retail prices.

ASUS ROG Astral GeForce RTX 5080 OC

With the announcement of the top RTX 50 graphics cards, ASUS has prepared a line of ROG Astral graphics adapters within the well-known gaming series.

The most powerful modification of the GeForce RTX 5080 series was the ASUS ROG Astral GeForce RTX 5080 OC model, which received increased operating frequencies of the GPU and increased power limits. This is the version of the video card that we reviewed.

This time, the developers obviously did not restrain themselves by the limits of compact forms, so even the overall dimensions of the video card are impressive here - 358 x 150 x 76 mm. If we add that the adapter has a 3.8-slot design, and its weight is 2.9 kg, then this even more conveys the feeling of monstrosity of the new model. Taking the ASUS ROG Astral GeForce RTX 5080 OC in your hands, the effect of a “precious thing” only intensifies. A heavy bar with a cast metal frame, which has a high-quality assembly and fit.

ROG Astral models feature a new cooling system with four fans at once. The cooler is based on a solid radiator block consisting of aluminum plates with increased density, which are pierced by eight heat pipes.

The design also uses an oversized evaporation chamber with milled heat pipe slots to accelerate heat transfer from the GPU and memory chips. A phase transition thermal interface is used to improve contact density between the GPU die and the heat sink surface.

The overall design with through-flow ventilation for graphics cards has already become a steady trend when it comes to powerful models. The four-fan version actually offers a further development of this concept.

There are three axial fans installed on the front panel, and another one is mounted on the back and helps to blow air through the radiator block.

The developers claim that this way they can increase the air pressure in the vertical channel by 20%. In turn, this allows to increase the density of the radiator plates, and accordingly the total dissipation area, without worsening the blowing. All technical tricks are aimed at improving the temperature regime of the video card elements.

Since the printed circuit boards for the reference GeForce RTX 5090/5080 models have a rather specific layout and design, all adapters of the corresponding series from partners use original PCBs of their own development. In the case of the ASUS ROG Astral GeForce RTX 5080 OC, we are dealing with a rather powerful power stabilization subsystem. The GPU circuit uses a 20-phase circuit with Vishay SIC654A (DrMOS) power assemblies, designed for an operating current of up to 50 A. The VRM parameters are controlled by the MP29816 controller from Monolithic Power Systems.

To power the memory chips, a 3-phase circuit is used with the same DrMOS assemblies, which are controlled by the uPI uP9529Q. As for the DRAM chips themselves, they are located in a U-shaped manner next to the GPU. In this case, we are talking about Samsung K4VAF325ZC-SC32 chips, which according to the specification are capable of operating at an effective 32,000 MHz.

Compared to the reference PCB from the Founders Edition adapter, the ASUS model has a significantly larger board, while effectively using its entire area. A solid base without additional connections is rather a certain advantage, especially when it comes to using the PCI Express 5.0 interface. At least we can note that the video card had no problems with operation in this mode.

To connect additional power to the video card, there is one 16-pin 12HVPWR connector on the top side. In case the power supply does not have a suitable cable, the kit includes an adapter of three 8-pin PCI-E 12V connectors. As for the power supply of the video card in general, the developers recommend equipping the platform with a power supply with a capacity of more than 850 W. The total Total Graphics Power (TGP) for the GeForce RTX 5080 is 360 W, however, forced models of graphics adapters may have higher power limits.

As expected, a video card of this level is equipped with two BIOS chips, which allows you to use both standard settings (P Mode) and the “Q Mode” mode with a reduced noise level, which provides a less aggressive cooler operation algorithm.

ASUS ROG Astral GeForce RTX 5080 OC also has two 4-pin PWM-controlled connectors for connecting an appropriate number of additional case fans. If necessary, you can adjust the algorithm for their rotation speed depending on the GPU temperature.

The video card has received additional lighting. The decoration is neat, but quite accented. The ARGB strip is located on the side of the adapter, so it is clearly visible in the classic layout of the system unit. AURA Sync support allows you to synchronize the illumination with that of other system components, and support for addressable lighting allows you to increase the variety of lighting effects.

On the back side, the video card is covered by a metal plate, which has general protective functions, and also comes into contact with certain elements through heat-conducting stickers. Therefore, the plate also acts as an additional cooler. In the GPU area, an additional bracket is installed, which balances the pressure of the cooling system against the graphics processor crystal.

The fan on the back of the graphics card still looks unusual. Although we have seen a similar approach on NVIDIA reference adapters, this case is special, and the design differences are also obvious.

The interface panel of the ASUS graphics card is somewhat different from the standard set of reference GeForce RTX 5080. In addition to the basic three full-size DisplayPort 2.1b and one HDMI 2.1b for the series, the adapter offers another additional HDMI 2.1b port. This makes switching easier when, for example, you need to connect a TV and certain models of VR headsets to the system.

Test system configuration

• Motherboard: ASUS ROG STRIX Z890-F GAMING WIFI
• Processor: Intel Core Ultra 7 265KF (8P+12E; 3.9/5.5 GHz + 3.3/4.6 GHz)
• Cooling: ASUS ROG STRIX LC III 360
• Memory: G-SKILL Trident Z5 RGB DDR5-6400 32GB (F5-6400J3239G16GX2-TZ5RK)
• Graphics card: ASUS ROG Strix GeForce RTX 4080 SUPER 16GB GDDR6X OC Edition (ROG-STRIX-RTX4080S-O16G-GAMING)
• Drive: KINGSTON NV3 2 TB (SNV3S/2000G)
• BZ: ASUS TUF Gaming 1000W Gold
• Case: ASUS TUF Gaming GT502 PLUS

ASUS ROG Astral GeForce RTX 5080 OC came to us for testing as part of an assembled system from Artline. The PC used an Intel LGA1851 platform with a 20-core Core Ultra 7 265KF processor and an ASUS ROG STRIX Z890-F GAMING WIFI motherboard on the top-end Intel Z890 chipset.

To cool the chip, an ASUS ROG STRIX LC III 360 liquid cooling system with a three-fan layout, a radiator block of appropriate dimensions, and a fairly powerful pump was used.

The platform was complemented by a dual-channel G-SKILL Trident Z5 RGB DDR5-6400 RAM kit with a total capacity of 32 GB, and a 2 TB KINGSTON NV3 drive was used for data storage.

Considering the recommendations regarding the power of the PSU for a platform with a GeForce RTX 5080 (850+ W), the available “kilowatt” of the ASUS TUF Gaming 1000W Gold is quite enough even for overclocking experiments.

The PC was assembled in an ASUS TUF Gaming GT502 PLUS case. The modular two-chamber model with good ventilation of the internal space provides the necessary temperature regime for a fairly powerful platform.

In operation

Initially, we were interested in the general performance of the video card – frequencies, heating, and fan speed of the cooling system.

In idle mode, the video card switches to passive cooling mode, stopping the fans. The GPU temperature in such conditions was kept within 30C, the memory chips - 40C. The total power consumption of the video card is 10 W.

Under gaming load, the GPU warmed up to 66C, the GDDR7 chips – up to 68C. To maintain such values, the fan speed at the peak rose to 1800 rpm. For video cards of the GeForce RTX 50 line, the developers for some reason disabled the ability to monitor the temperature of the hottest point of the GPU (Hot Spot). As for the average value, there are no comments here.

Under load, the GPU accelerated to almost 3 GHz, the memory frequency was 1875 MHz, which for GDDR7 is equivalent to an effective 30,000 MHz. At rest, the GPU operated at 255 MHz, and the memory chips at 50 MHz (800 MHz).

We also tried the operation of the video card in “Q Mode” mode, switching the BIOS selection toggle switch to the appropriate position. What changes did we get? Under prolonged load, the temperature of the graphics processor rose to 72C, the memory chips – to 78C. At the same time, the fan speed was only 900 rpm. In such conditions, the video card worked much quieter than in standard mode.

Despite the recent release of the video card, a firmware version has already appeared on the manufacturer's website, which offers an even more “economical” fan operation algorithm for “Q Mode”. However, it seems that a rational balance has been achieved even in the basic version. Silence lovers can rest assured about this point, the available “Q Mode” allows you to get very good results.

Test results

To assess the capabilities of the new video card, it would be indicative to compare its performance with that of its predecessor, the GeForce RTX 4080. However, the latter was actually replaced by the upgraded version of the GeForce RTX 4080 SUPER. The super modification had quite modest differences in performance from the basic version, but had a recommended price tag of $999 instead of $1199. Therefore, comparing the performance of the GeForce RTX 5080 with that of the GeForce RTX 4080 SUPER also makes practical sense and will be quite indicative.

Although after the announcement and start of sales of the top models of GeForce RTX 50, the availability of older graphics cards of the previous generation has significantly deteriorated. NVIDIA has been carefully preparing for the release of new graphics cards, allowing partners to free up their stockpiles of adapters on GPUs with the Ada Lovalace architecture. So even for comparison, getting a GeForce RTX 4080 SUPER turned out to be not so easy - it is out of stock, everything is sold out.

In the end, we still managed to get the ASUS ROG Strix GeForce RTX 4080 SUPER 16GB GDDR6X OC Edition (ROG-STRIX-RTX4080S-O16G-GAMING) for these purposes. The overclocked version has a frequency formula of 2295/2640 MHz instead of the recommended 2295/2550 MHz.

So, first, we suggest you familiarize yourself with the screenshots of the ASUS ROG Astral GeForce RTX 5080 OC testing results in graphic and computational tests. Let's analyze the indicators and compare them with the values obtained on the previous generation video card. For convenience, in the presented diagrams, the ASUS ROG Strix GeForce RTX 4080 SUPER OC model, which was used in its standard mode, is marked as “ GeForce RTX 4080 SUPER OC ”.

Key stages of the 3DMark test give an undeniable advantage to the new generation graphics card. The GeForce RTX 5080 will receive an average of 21% more points than the RTX 4080 SUPER. A fairly significant difference, which further demonstrates the potential of the new architecture.

Compute tests generally also favor the RTX 5080, but the difference in performance depends on specific stages. In GeekBench, the new product turned out to be 10% more productive in OpenCL calculations. The test scene in Blender using the GPU (Optix) received a 7% acceleration. At the same time, V-Ray 6 using the RTX engine added as much as 23%. So the difference in performance can vary quite significantly and depends on the conditions.

Gaming performance

How will the power of both graphics cards compare in real games? According to the results of tests in several diverse projects using a resolution of 1920×1080 and maximum quality settings, the actual fps difference was 2–12%.

The average figure for this mode is 6.5%. It is unlikely that anyone will buy a GeForce RTX 5080 for Full HD games, but in certain cases (Black Myth: Wucong and STALKER 2: Heart of Chernobyl) at maximum settings, even a video card of such power will not be superfluous. In addition, ray tracing technology was not used during these tests. Perhaps the overall difference in Full HD would be somewhat more noticeable when using the most powerful processors with 3D V-Cache from AMD, which are revealed precisely under such conditions.

With the transition to a resolution of 2560×1440 , the advantage of the ASUS ROG Astral GeForce RTX 5080 OC over its forced predecessor has increased to 11% on average. The minimum difference (6%) was observed in Assassin's Creed Mirage, but in Cyberpunk 2077 the new Blackwell improved the performance of the RTX 4080 SUPER by as much as 23%.

When using 4K, the advantage of the GeForce RTX 5080 was already 17% on average, ranging from 10% to 33% (all the same Cyberpunk 2077). In addition to the architectural features of the GPU, we assume that the increase in the difference is due to the significantly higher (+30%) memory bandwidth of the new model.

DLSS 4 and Multi Frame Generation (MFG)

One of the key features of the GeForce RTX 50 series graphics cards is support for the next-generation DLSS 4 smart scaling technology based on transformer models, as well as the Multi Frame Generation (MFG) function. While some of the DLSS 4 options are available to previous-generation RTX graphics cards, MFG is an exclusive feature of adapters based on the Blackwell architecture.

The ability to generate up to four additional frames has already been implemented in some projects. We experimented with Cyberpunk 2077 and Alan Wake 2 in 4K resolution.

Cyberpunk 2077 initially used the RT Overdrive mode with maximum image quality, which is capable of “bringing to its knees” even video cards of this level. The starting 20 fps is definitely not the level of performance that you can put up with. When using DLSS (Balanced), the average performance increased to 44–50 fps, and after activating the basic Frame Generation, the performance increased to an acceptable 78–87 fps. The GeForce RTX 5080 video card turned out to be almost 12% faster in such conditions. However, the new product on Blackwell has MFG support, which allowed it to actually double the number of frames. The average 166 fps in 4K with RT Overdrive looks good.

In general, a similar situation can be observed in the game Alan Wake. In 4K mode with the highest quality ray tracing (RT Ultra), we initially have only 15–18 frames/s. But if we add scaling and multiple frame generation technologies (for RTX 5080), we get 113 fps at the output.

Note that at the previous stage, in the mode with high-quality DLSS and basic Frame Generation, the GeForce RTX 4080 SUPER video card was no longer able to achieve an average of 60 fps, while the additional +15% performance of the RTX 5080 allowed us to achieve the minimum desired indicator. Of course, this is an isolated case, which, however, has a certain symbolism.

The number of projects with DLSS 4 and MFG support will grow, so the GeForce RTX 50xx series models will gradually gain more advantages.

Overclocking

In addition to performance indicators, it was interesting to evaluate the frequency potential of the new series of video cards. It is worth noting right away that the ASUS ROG Astral GeForce RTX 5080 OC already in the basic version has an increased Power Limit of 400 W, so this parameter does not change further.

Experimentally, the base frequency of the graphics processor was increased by as much as 400 MHz - from 2295 MHz to 2695 MHz. Under such conditions, the average acceleration rate increased to 3160 MHz. GDDR7 memory chips also have a good frequency margin. We managed to increase the actual frequency from 1875 MHz (30,000 MHz) to 2125 MHz (34,000 MHz) and this is clearly not the limit, because this was the limitation of the configuration utility.

After additional overclocking, the GPU peaked at 3375 MHz, although typical average values under gaming load were slightly lower. Surprisingly, despite the increase in operating frequencies, during measurements, the temperature values for the GPU/memory remained at the same level as before overclocking - 66/68C. Are there any signs of through-flow, a giant radiator with an evaporation chamber, or a thermal interface with a phase transition? It probably works in combination.

After additional overclocking, the graphics performance in the 3DMark Time Spy Extreme test increased by 9% - from 16,910 to 18,439 points. A good indicator, especially for cases where the cooling system copes with increased load without any special difficulties. The obtained values \u200b\u200bare at least worth considering when planning a purchase. An additional plus of ~10% of performance cannot always be obtained even when migrating to another generation of discrete graphics.

Power consumption

The GeForce RTX 5080 series graphics cards have a claimed power consumption of 360 W. This is 12.5% more than the GeForce RTX 4080 SUPER. Despite using GPUs with similar technological parameters, higher processor frequencies with the same manufacturing process require additional energy consumption.

Returning to performance monitoring, we note that the ASUS ROG Astral GeForce RTX 5080 OC under high gaming load (Cyberpunk 2077) consumed up to 391 W. The GeForce RTX 4080 SUPER model required about 320 W.

The calculation of the overall energy efficiency will depend on the conditions. For certain games with a small increase in performance, it will be lower than in its predecessor, in Cyberpunk 2077 and 3DMark load - comparable values, and if we take into account the increase in fps after activating MFG, then here the novelty is out of competition. However, in this case, it will be necessary to introduce a new term - the number of watts of energy consumed per one generated fps.