Today, AMD will host its “Together we advanced the gamelivestream where the company will unveil its Radeon RX 7000 “RDNA 3” GPUs.
Chiplets are coming to gaming GPUs! The AMD Radeon RX 7000 “RDNA 3” graphics card unveiling is only 24 hours away, tune in to the livestream here
The livestream has a unique focus and purpose, it will mark the unveiling of AMD’s next generation RDNA 3 GPUs which will power the Radeon RX 7000 series graphics cards. software ecosystem that will revolve around new gaming chips.
Join AMD for “together we advance_gaming”, announcing the next generation of GPUs based on the revolutionary AMD RDNA™ 3 architecture bringing new levels of performance, efficiency and functionality to gamers and creators.
Our top-of-the-line RDNA 3 GPUs will deliver big increases in performance and performance per watt over our current products and will include new features to support high resolution, high frame rate gaming. We look forward to sharing more details later this week.
AMD confirmed that its RDNA 3 GPUs will arrive later this year with a huge performance boost. The company’s Senior Vice President of Engineering, Radeon Technologies Group, David Wang, said next-generation GPUs for the Radeon RX 7000 series will deliver over 50% performance per watt over existing RDNA 2 GPUs . Some of the key features of RDNA 3 GPUs highlighted by AMD will include:
- 5nm process node
- Advanced chip packaging
- Redesigned computing unit
- Optimized graphics pipeline
- Next Generation AMD Infinity Cache
- Improved ray tracing capabilities
- Refined adaptive power management
- >50% performance/watts compared to RDNA 2
Join us on November 3 for the unveiling of #RDNA3 and the world premiere of exciting new footage from @CallistoTheGame, @Forespokena sneak peek at something amazing from our friends at @UbisoftAnd much more…
Live at 4 p.m. on: https://t.co/GE89dZgxqF pic.twitter.com/Yuy4BpRz94
— Radeon RX (@Radeon) November 2, 2022
The initial lineup is expected to include two cards based on the flagship Navi 31 GPU, the 24GB Radeon RX 7900 XTX and the 20GB Radeon RX 7900 XT.
AMD Radeon RX 7900 XTX 24GB Graphics Card Specifications
The AMD Navi 31 GPU, the flagship RDNA 3 chip, is said to power next-gen enthusiast cards such as the Radeon RX 7900 XTX/XT graphics card. We’ve heard that AMD will be ditching CUs (computing units) in favor of WGPs (workgroup processors) on its next-gen RDNA 3 GPUs. Each WGP will house two CUs (computing units), but with twice as many SIMD32 clusters instead of just 2 on each CU in RDNA 2. Rumors say that AMD has the option to choose between Samsung and TSMC for chip 6 n.
- AMD Navi 31 XTX: 12288 cores, 384-bit bus, 192MB Infinity cache, 308mm2 5nm GPU chip
- AMD Navi 21 XTX: 5120 cores, 384-bit bus, 128 MB Infinity cache, 520 mm2 GPU chip at 7 nm
According to latest information, the AMD Navi 31 GPU with RDNA 3 architecture is expected to offer a single GCD with 48 WGP, 12 SA and 6 SE. This will give a total of 12,288 SPs or stream processors. This is a 2.4x increase in core count over the 5120 SP of the Navi 21 GPU. The GPU or Navi 31 GCD would measure 308mm2 and will be packaged on TSMC’s 5nm Process Node.
The Navi 31 GPU will also carry 6 MCDs which will include 16MB of Infinity Cache per chip and are also likely to carry the 64-bit (32-bit x 2) memory controllers which will provide the chip with a 384-bit bus interface.
While this equates to 96MB of Infinity Cache, which is less than the 128MB featured on current Navi 21 GPUs, there is also a 3D-Stacked solution in the works that was highlighted recently and that would double the Infinity Cache with capacities of 32MB (16MB 0-hi + 16MB 1-hi) for a total of 192MB of cache. This is a 50% increase over the current Navi 21 design and it also makes the Navi 31 the first GPU with both stacked, chiplet and 3D designs. These chiplets or MCDs will be fabricated on TSMC’s 6nm process node and will measure 37.5mm2 each.
AMD Radeon RX 7900 XT 20GB Graphics Card Specifications
There will also be a scaled down variant which will feature the Navi 31 XT GPU core. This chip will contain 42 WGP or 10,752 cores. That’s 12.5% fewer hearts than the full-fat variant. The GPU will also run at slightly lower clock speeds, but the power consumption will be lower.
- AMD Navi 31XT: 12288 cores, 320-bit bus, 160MB Infinity cache, 308mm2 5nm GPU chip
- AMD Navi 21XT: 4608 cores, 256-bit bus, 128 MB Infinity cache, 520 mm2 GPU chip at 7 nm
The graphics card will also feature 20GB of GDDR6 memory on a 320-bit wide bus interface. Since there are only 5 MCDs enabled, the card will end up with 80MB of Infinity Cache, which is 16MB less than the top variant and a decrease of 16.6%.
AMD Radeon RX 7900 XT “preliminary” specs:
| Graphic card | AMD Radeon RX 7900 XTX | AMD Radeon RX 7900XT | AMD Radeon RX 6950XT | AMD Radeon RX 6900XT |
|---|---|---|---|---|
| GPUs | Navi 31 XTX | Navi 31 XT | Navi 21 KXTX | Navi 21 XTX |
| Process node | 5nm + 6nm | 5nm + 6nm | 7nm | 7nm |
| Die size | 308mm2 (only GCD) 533 mm2 (with MCD) |
308mm2 (only GCD) 533 mm2 (with MCD) |
520mm2 | 520mm2 |
| Transistors | To be determined | To be determined | 26.8 billion | 26.8 billion |
| WGP-GPU | 48 | 42 | 40 | 40 |
| Stream processors | 12288 | 10752 | 5120 | 5120 |
| TMU/ROP | To be determined | To be determined | 320 / 128 | 320 / 128 |
| game clock | To be determined | To be determined | 2100MHz | 2015MHz |
| Boost clock | >3GHz | >3GHz | 2310MHz | 2250MHz |
| TFLOP FP32 | >75 TFLOPs | >65 TFLOPs | 23.65 TFLOPs | 23.04 TFLOPs |
| Memory size | 24 GB GDDR6 | 20 GB GDDR6 | 16 GB GDDR6 | 16 GB GDDR6 |
| Infinite cache | 96 MB | 80 MB | 128 MB | 128 MB |
| Memory bus | 384 bit | 320 bit | 256 bit | 256 bit |
| Memory clock | ~20 Gbps | ~20 Gbps | 18 Gbps | 16 Gbps |
| Bandwidth | ~960 GB/s | ~800 GB/s | 576 GB/s | 512 GB/s |
| Effective Bandwidth | To be determined | To be determined | 1728.2 GB/s | 1664.2 GB/s |
| PAD | ~400W | ~350W | 335W | 300W |
| PCIe interface | PCIe 5.0 x16 | PCIe 5.0 x16 | PCIe 4.0 x16 | PCIe 4.0 x16 |
| Price | To be determined | To be determined | US$1099 | US$999 |
Next Generation Adaptive Power Management and Infinity Cache Confirmed for RDNA 3
AMD SVP & Technology Architect Sam Naffziger already highlighted that next-generation RDNA 3 GPUs featured on Radeon RX 7000 GPUs and next-generation iGPUs, will offer a range of new technologies, including refined adaptive power management technology to set performance-specific operating points. workload, ensuring that the GPU uses only the power needed for the workload. The GPUs will also include next-generation AMD Infinity Cache which will deliver higher density, lower power consumption caches and reduced power requirements for graphics memory.
And after?
Looking ahead, we continue our efforts for more efficient gaming with the AMD RDNA 3 architecture. As the first AMD graphics architecture to take advantage of the 5nm process and our chip packaging technology, AMD RDNA 3 is On track to deliver >50% better estimated performance per watt than AMD RDNA 2 architecture, truly bringing high-end gaming. performance to gamers in cool, quiet, and energy-efficient designs.
Contributing to this power-efficient design, AMD RDNA 3 refines AMD RDNA 2 adaptive power management technology to set workload-specific operating points, ensuring that each GPU component uses only the power it has. needed for optimal performance. The new architecture also introduces a new generation of AMD Infinity Cache, which is expected to deliver even denser, lower-power caches to reduce graphics memory power requirements, helping consolidate AMD RDNA 3 and Radeon graphics. as true leaders in efficiency.
We’re excited about the improvements we’re making with AMD RDNA 3 and its predecessors, and believe there’s even more to be gained from our advanced architectures and process technologies, delivering unmatched performance per watt across the entire stack. as we continue our push for a better game.
AMD will unveil its RDNA 3 GPU architecture and Radeon RX 7000 graphics cards on November 3 while the actual commercial launch will take place a few weeks later.