Until recently, AMD had a big gap between the cheaper RX 7800 XT and the more powerful RX 7900 XT. After the launch of the RTX 4070 Super, which hit it, AMD responded with the Radeon RX 7900 GRE. Even though the Gigabyte model is one of the most affordable, its features will satisfy most interested parties.
The first Radeon RX 7000 was introduced by AMD in November 2022. They use graphics chips with the RDNA3 architecture, which has a redesigned ALU with dual stream processors (dual SIMD units), separate units for accelerating AI calculations, separate shader clocks from the front-end, a new generation of units used for ray tracing calculations with new functions, support for compression and decompression of the AV1 format and support for the new DisplayPort 2.1 standard. You can find a more detailed description in the article introducing the RDNA3 architecture.
More powerful models built on the largest Navi 31 chip are Radeon RX 7900 XTX and Radeon RX 7900 XT. In China, they were joined last year by the weaker RX 7900 GRE model. The GRE got its unusual postscript precisely because of the Chinese market. Last year belonged to rabbits in China, hence the abbreviation GRE – Golden Rabbit Edition. Although the RX 7900 GRE will officially arrive on the European market only in the year of the dragon, it was already devoted to rabbits once, and that’s how it stayed.
All three Radeon RX 7900s are equipped with the largest chip with the RDNA3 architecture, codenamed Navi 31. It is the first gaming graphics chip divided into a large computing core (GCD) and smaller chiplets (MCD), into which the infinity cache and memory controllers have been moved .
The RX 7900 XTX uses Navi 31 with a fully active core, six active MCDs, 96MB of infinity cache and 24GB of GDDR6 memory at 20Gb/s connected via a 384b bus. The recommended price from AMD at launch was $999, which is roughly 29,000 crowns including tax at the current exchange rate, today the prices of the available RX 7900 XTX start at just over 25,000.
In the RX 7900 XT, the chip has slightly worse parameters, it has only five functional MCD chiplets forming 80 MB of infinity cache and a 320b bus, on which 20 GB of GDDR6 memory with a speed of 20 Gb/s is connected. The recommended introductory price of $899 is 26,100 crowns with VAT at the current exchange rate, but since then the prices of the RX 7900 XT have dropped significantly, and in our country they start at 20,000 crowns with the RX 7900 XTX from XFX.
AMD originally targeted the competing 16GB GeForce RTX 4080 with both new models. But Nvidia has since introduced a number of other models and “facelifted” models from the RTX 40 Super series. One of the new models is the RTX 4070 Super, which ranks between the RTX 4070 and RTX 4070 Ti in terms of performance. Against the RTX 4070 Super, AMD did not have a direct competitor in the Western markets – the RX 7800 XT is built against the original RTX 4070, and the RX 7900 XT is significantly more expensive.
AMD solved it simply – they decided to introduce the Radeon RX 7900 GRE to markets where it was not officially sold yet.
You can guess where the Radeon RX 7900 GRE with the weakest variant of the Navi 31 chip was saved in the first place – it lost another of the MCD chiplets and part of the video memory. The large CCD is supplemented by only four MCD chips, and the 4GB memory block has also disappeared from the other memory controllers. So the cards are equipped with four controllers and 16 GB of memory just like the RX 7800 XT. But AMD reduced the computing units and, unfortunately, also the memory clocks.
In addition to the lower number of stream processors and texturing units, the number of rasterization units is even lower compared to the RX 7900 XTX and RX 7900 XT. While both older models had 192 active units, the RX 7900 GRE has only 160. The significant decrease was due to the narrower bus and slower chips even in terms of memory bandwidth, it even fell below the level of the RX 7800 XT. But this is where you can really help by overclocking, with the latest version of AMD drivers you can go from the reference 18 Gb/s to over 20 Gb/s. The positive thing is that along with the chip parameters, the stated consumption and efficiency of the chip also dropped significantly – the card has a higher performance than the RX 7800 XT with similar consumption.
A detailed comparison of the specifications of individual models is in the table below.
| Model | Radeon RX 7900 XTX | Radeon RX 7900 XT | Radeon RX 7900 GRE | Radeon RX 7800 XT | Radeon RX 6950 XT | Radeon RX 6800 XT |
| Computing Units (CU) | 96 | 84 | 80 | 60 | 80 | 72 |
| Rasterization units (ROPs) | 192 | 192 | 160 | 96 | 128 | 128 |
| Stream processors | 6144 | 5376 | 5120 | 3840 | 5120 | 4608 |
| Texture units | 384 | 336 | 320 | 240 | 320 | 288 |
| Ray tracing accelerators | 96 | 84 | 80 | 60 | 80 | 72 |
| Infinity cache | 96 MB | 80MB | 64MB | 64MB | 128MB | 128MB |
| Basic beat | 1929 MHz | 1387 MHz | 1287 MHz | 1890 MHz | ||
| Game tempo | 2300 MHz | 2000 MHz | 1880 MHz | 2124 MHz | 2100 MHz | 2015 MHz |
| Tact with boost | Up to 2500 MHz | Up to 2400 MHz | Up to 2245 MHz | Up to 2430 MHz | Up to 2310 MHz | Up to 2250 MHz |
| Memory bus width | 384 b | 324 b | 256 b | 256 b | 256 b | 256 b |
| Memory capacity | 24 GB | 20 GB | 16 GB | 16 GB | 16 GB | 16 GB |
| Memory type | GDDR6 | GDDR6 | GDDR6 | GDDR6 | GDDR6 | GDDR6 |
| Memory speed | up to 20 Gbps | up to 20 Gbps | up to 18 Gbps | up to 19.5 Gbps | up to 18 Gbps | up to 16 Gbps |
| Memory throughput | up to 960 GB/s | up to 800 GB/s | up to 576 GB/s | up to 624 GB/s | up to 576 GB/s | up to 512 GB/s |
| Effective throughput (with infinity cache) | up to 3500 GB/s | up to 2900 GB/s | up to 2250 GB/s | up to 2708 GB/s | up to 1794 GB/s | |
| Total power input (TBP) | 355 W | 315 W | 260 W | 263 W | 335 W | 300 W |
Gigabyte AMD Radeon RX 7900 GRE 16GB
The Gigabyte RX 7900 GRE Gaming OC is the only Radeon RX 7900 GRE model in the Gigabyte range. The game clock has increased from the reference 1880 to 2052 MHz, the boost clock from the reference up to 2245 to 2391 MHz. The same values as the previously tested RX 7900 XT Nitro in overclocking mode.
The card has a switch with which you can choose between Silent and OC mode. The default is OC mode with factory overclocking, but in Silent mode the clock settings are not different and only the fan regulation settings are different. The performance of the card is only slightly lower due to the fact that the temperature of the chip rises slightly and with it its consumption, which is why it then runs at slightly lower clocks. The difference in performance is still more like tenths of a second and almost immeasurable, while the difference in noise is significant and worth switching to a quieter mode. Therefore, we will use the Silent mode as the basic one for all tests, including performance tests. Operating characteristics that differ more significantly are then measured in both modes.
On the left, BIOS in Silent mode, on the right in OC mode
A lot of data can be obtained from the monitoring, including hotspot temperatures (the hottest spot on the chip), total power consumption and power consumption on individual power branches. In common utilities such as GPU-Z, you can monitor a few more important values, but the card can reveal much more about itself in HWiNFO – in addition to the usual chip and hotspot temperature, VRM temperature, memory, consumption of individual parts of the card and their workload, power limits and more .
GPU PPT (Sustained) power limit is in OC mode increased from 212 to 244 W and peak power limit – GPU PPT limit (Short) is increased from 254.4 to 292.8 W. These are exactly the same values as Sapphire’s Nitro+ in OC mode. But it also had a quiet mode with reference clocks, a GPU PPT limit of 212 W and a peak limit of 254.4 W.
I really don’t know what to think about this anymore, it seems to me that AMD lists some reference specs on their site with lower clocks and lower consumption (260W) which looks better on paper, but in the end the manufacturers will get a premade one directly from AMD some “OC BIOS” where clocks and power limits are higher.
Cards with reference parameters practically do not exist, in the database on TechPowerUP you will find only overclocked models, with the exception of a single model from XFX, which looks like a card in a reference design. On paper, the RX 7900 GRE should have 25 W less power than the 285 W RTX 4070 Ti, but in reality they are practically the same.
Data that can be monitored in GPU-Z and HWiNFO monitoring, on the left is always BIOS in Silent mode, on the right in OC mode
on the left always BIOS in Silent mode, on the right in OC mode
And still supported formats for hardware video decoding:
On the back of the packaging, in addition to the list of AMD technologies, which is common to all Radeons, Gigabyte draws attention to the company’s technologies that the RX 7900 GRE uses – a WindForce cooler with three counter-rotating fans, a passive heatpipe and a large copper base that cools the memory and technology in addition to the graphics chip screen cooling, which is the name Gigabyte uses for coolers with free-through fins that allow heated air to flow over the card – but in the case of the RX 7900 GRE, the openings are minimalistic. There are still some ARGB LEDs left under the logo on the side of the card, the colors of which can be controlled via the Gigabyte Control Center, and a metal backplate on the back of the card that helps strengthen the card and protects the circuit board. You will only find a brief installation guide in the accessories.
For the record, I will add that it is a GV-R79GREGAMING OC-16GD model in revision 1.0, it may happen that a second revision will appear over time, which may have a different design, for example in the cooler, and thus slightly different operating characteristics.
Tags: Gigabyte AMD Radeon GRE Gaming test good features price
-
