With Hardware-Accelerated Ray Tracing Capabilities
Taken from Wccftech …AMD’s Radeon RX graphics cards based on the RDNA 2 GPU architecture are coming and so far, all eyes have been set on the ‘Big Navi’ GPU which will be launching later this year to compete against NVIDIA’s flagship Ampere Gaming graphics cards. But it looks like we should switch gears for this particular topic and take a look at the mainstream RDNA 2 GPUs too.
The latest findings come from Rogame who has confirmed the GPU ID for the Navi 23 chip. As per the findings, the Navi 23 GPU will have the GFX1032 ID for its graphics architecture. In fact, majority of the Navi 2X family will feature the GFX103* IDs which were originally found by Komachi and are as listed below:
- Navi 21 – GFX1030
- Navi 22 – GFX1031
- Navi 23 – GFX1032
- Van Gogh – GFX1033
…. From what I can tell, the Navi 23 GPU might be the mainstream replacement for the Navi 2X family, succeeding the Radeon RX 5500/5600 series. Since reports suggest that Big Navi ‘Navi 21’ GPU will be the first RDNA 2 chip launching in the gaming market, it is likely that a Navi 10 refresh could keep things interesting in mainstream segment until AMD launches its RDNA 2 mainstream lineup with the Navi 22 and Navi 23 GPUs. We can definitely expect much lower price points than existing Radeon RX 5000 series cards for the refreshed parts so the added performance would be a bonus. The expected die sizes of these three GPUs are also listed below:
- AMD Navi 21 (505mm2)
- AMD Navi 22 (340mm2)
- AMD Navi 23 (240mm2)
You can tell by the die size alone that Navi 23 is expected to be almost as big as the Navi 10 GPU which measures at 251mm2. Again, we haven’t seen how hw-accelerated ray tracing blocks affect the overall die size of AMD GPUs. NVIDIA’s RTX GPUs saw a slight increase in logic area of around 20-25% on average with the added RT cores. This might explain why even the smaller GPUs are so large compared to current-generation offerings. Or the other possibility is AMD just going for the brute force route and increasing the number of SMs by a huge factor.
Source: Wccftech