technology port Angistronomy It contains a very detailed report published regarding the specifications of AMD RDNA 3 GPUs such as the Navi 31, Navi 32, and Navi 33 chips that will continue to power the next generation Radeon RX 7000 series graphics cards.
AMD Navi 31, Navi 32, Navi 33 “RDNA 3” GPUs for Radeon RX 7000 Graphics Cards Detailed
Information previously leaked inside its Linux patches open That there will be at least four RDNA 3 IP addresses released to the consumer sector. This includes Navi 31, Navi 32, Navi 33, and APU design.
|GC IP VER||GFX ID||AMDGPU_FAMILY (via Coelacanth-Dream)||Writes|
|11.0.0||gfx1100 (Navi31)||AMDGPU_FAMILY_GC_11_0_0 (FAMILY_GFX1100)||dGPU|
|11.0.2||gfx1102 (Navi32)||AMDGPU_FAMILY_GC_11_0_0 (FAMILY_GFX1100)||dGPU|
|11.0.3?||gfx1101 (Navi33)?||AMDGPU_FAMILY_GC_11_0_0 (FAMILY_GFX1100)?||dGPU?|
Angstronomics has now reported information such as die sizes, GPU configurations, and features that may end up in final retail products. An interesting statement was released by the tech outlet which reveals that the configurations for the Navi 3X GPUs were first formulated in 2019 and then finalized sometime in 2020. These specs have not changed since then, however, she indicated. Different rumors have two different specs for the GPUs which means that either AMD has already made changes since its final designs or that those rumors simply aren’t in line with the truth. With that, let’s see what kind of configurations we can expect with Navi’s 3x “RDNA 3” GPU lineup.
AMD confirmed That its RDNA 3 GPUs are coming later this year with a massive performance boost. The company’s senior vice president of engineering, Radeon Technologies Group, David Wang, said the next generation of Radeon RX 7000 series GPUs will deliver more than 50% performance per watt increment versus current RDNA 2 GPUs. Some of the key features of RDNA 3 GPUs highlighted by AMD will include:
- 5nm process node
- Advanced chiplet packaging
- Revisited unit of account
- Enhanced graphics pipeline
- The next generation of AMD Infinity cache
- >50% Perf/Watt vs RDNA 2
Gamers should expect fluid 4K gaming to be incredibly smooth on these graphics cards and with the FidelityFX suite offering support for next-generation FSR, SAS, and SAM, we might see up to 60fps playable in 8K.
AMD Navi 31 “Plum Bonito” GPU Configuration
The AMD Navi 31 GPU, the flagship RDNA 3 chip, will power next-generation enthusiast cards like the Radeon RX 7900 XT graphics card. We’ve heard that AMD will drop CUs (computing units) in favor of WGP (working set processors) in the next generation of RDNA 3 GPUs. Each WGP will contain double CUs (computational units) but with twice the SIMD32 combinations versus only 2 in each CU within RDNA 2. Rumors are that AMD has the option to choose between Samsung and TSMC for the 6nm die.
- AMD Navi 31: 12288 cores, 384-bit bus, 192MB infinity cache, 308mm2 @5nm GPU
- AMD Navi 21: 5120 cores, 384-bit bus, 128MB infinity cache, 520mm2 @7nm GPU
according to Latest informationThe AMD Navi 31 GPU with RDNA 3 architecture is expected to offer a single GCD with 48 WGPs, 12 SAs and 6 SEs. This will give a total of 12288 SP or stream handlers. This is a 2.4x increase in cores compared to the 5120 SPs found in the Navi 21 GPU. It is said that the GPU or Navi 31 GCD measures 308 mm2 It will come packaged in TSMC’s own 5nm process node.
The Navi 31 GPU will also carry 6 MCDs which will have a 16MB Infinity cache per die, and likely also 64-bit (32-bit x 2) memory controllers that will supply the chip with a 384-bit bus interface. While this equates to 96MB of Infinity cache which is down from the 128MB found in current Navi 21 GPUs, there is also a 3D solution in the works that Recently mentioned This would double the 32MB Infinity cache (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 graphics processing unit (GPU) with microchip and 3D stacked designs. These chips or MCDs will be manufactured on a 6nm 37.5mm TSMC process node2 All.
The port reported that there was a 288MB above MCD package (16MB 0-Hi + 16MB 1-Hi + 16MB 2-Hi) planned but this was canceled for cost and revenue reasons. There is also talk of a shortened variant of the Navi 31 GPU which will have 42 WGP or 10752 cores and 5 MCDs for 80MB Infinity Cache over the 320-bit bus interface.
Now, this will lead to a higher power draw and it looks like AMD has confirmed That’s to the point that their next-generation graphics lineup will feature higher power consumption but still be a more efficient option than what NVIDIA offers. The AMD Radeon RX 6950 XT already has a TBP of 335W, so for a 2x performance boost. The cards are expected to retain a dual 8-pin plug input for power and feature a triple-fan cooling design which is slightly longer than the card currently in use.
AMD Navi 32 “Wheat Nas” GPU Configuration
The AMD Navi 32 GPU codenamed “Wheat Nas” is also one of two MCM GPUs featured in the RDNA 3 family. The GPU will contain one GCD (Graphics Compute Templates) and four MCDs (Multi-Cache Template) temporary). The GPU or GCD is said to measure 200mm2 While each MCD will maintain the same size at 37.5mm2.
- AMD Navi 32: 7680 cores, 256-bit bus, 64MB infinity cache, 200mm2 @5nm GPU
- AMD Navi 22: 2560 cores, 192-bit bus, 96MB infinity cache, 335mm2 GPU Die @7nm
The GCD will be equipped with 3 Shader Engines and each Shader Engine will contain 6 Shader Arrays (18 in total). The GPU will contain a total of 30 WGP or 60 compute units for a total of 7,680 cores. This rounds out up to 50% more cores than the flagship Navi 31.
Each MCD must also have two memory connection links (32-bit). This is a total of 8 32-bit memory controllers for a 256-bit bus interface. Each MCD will maintain 16MB 0-Hi but there doesn’t appear to be any 1-Hi variant which means the chip will only have 64MB max configurations. It is said that the Navi 32 GPU will be launched in 2023 and will be the top-of-the-line chip for mobile and a mid-tier chip for desktop computers. The GPU will be launched first in the mobile segment and may be announced at CES 2023 which makes perfect sense as we will have Phoenix Range and Dragon Range laptops with Zen 4 cores ready by then.
AMD Navi 32 “Hotpink Bonefish” GPU Configuration
The AMD Navi 33 GPU codenamed “Hotpink Bonefish” will start the monolithic part of the RDNA 3 family. The GPU will contain a single die. The die is very similar to the flagship Navi 21 GPU and is expected to use a 6nm process node for manufacturing with a die size of 203mm.2.
- AMD Navi 33: 4096 cores, 128-bit bus, 32MB infinity cache, 335mm 2 GPU Die @6nm
- AMD Navi 23: 2048 cores, 128-bit bus, 32MB infinity cache, 237mm 2 GPU Die @7nm
The Navi 33 GCD is expected to have two Shader Engines and each Shader Engine has two Shader Arrays (2 per SE/4 in total). This rounds out up to 16 WGP or 32 compute units for a total of 4,096 cores which is higher than the Radeon RX 6800 “Navi 21 XL” GPU. The GPU will come packed with 32MB of Infinity Cache, the same size as the Navi 23 GPU, and across a 128-bit bus. Just like the Navi 32 GPUs, the Navi 33 chipset will focus first on the mobile family with early graphics performance putting it ahead of the Intel Arc Alchemist lineup at less than half the cost and at the same time draws less power.
The port also talks about two new features coming to AMD RDNA 3 Navi 3x GPUs for Radeon RX 7000 series graphics cards such as:
One of the features in the RDNA 3 graphics pipeline is OREO: an opaque random export command, which is just one of many space-saving technologies. With gfx10, pixel shaders work out of order, as the output goes to the rearrange buffer before moving to the rest of the pipeline in order. With OREO, the next step (mashup) can now receive and execute operations in any order and export to the next stage in that order. Thus, the ROB can be replaced with a much smaller sliding buffer, which saves space.
The size of memory-attached cache blocks (MALL) is halved, doubling the number of banks for the same amount of cache. There are also changes and additions that increase the fee on MALL’s bandwidth and reduce the exit penalty to VRAM.
The AMD Radeon RX 7000 series “RDNA 3” GPU based on Nav 3x GPUs is expected to be launched later this year with the Navi 31 flagship reportedly launching first, followed by the Navi 32 and Navi 33 GPUs.
AMD RDNA 3 Navi 3X GPU Configurations (Initial)
|GPU name||Navi 21||Navi 33||Navi 32||Navi 31||Navi 3X|
|Code name||Sienna Cichlid||Hotpink Bonfish||wheat people||Bloom Bonito||to be announced later on|
|GPU operation||7 nm||6 nm||5 nm / 6 nm||5 nm / 6 nm||5 nm / 6 nm|
|GPU Package||congener||congener||MCM (1 GCD + 4 MCD)||MCM (1 GCD + 6 MCD)||Million cubic meters (TBC)|
|GPU Template Size||520 mm 2||203 mm2 (GCD only)||200 mm2 (GCD only)
425 mm2 (with MCDs)
|308 mm2 (GCD only)
533 mm2 (with MCDs)
|to be announced later on|
|SP per WGP||128||256||256||256||256|
|Calculation units (per template)||80||32||60||96||128 (per GPU)
|cores (per die)||5120||4096||7680||12288||8192|
|Memory bus||256 bit||128 bit||256 bit||384 bit||384 bit x2?|
|Memory capacity||up to 16 GB||up to 8 GB||up to 16 GB||up to 24 GB||up to 32 GB|
|Memory speed||16-18 Gbps||to be announced later on||to be announced later on||20 Gbps||to be announced later on|
|memory bandwidth||512-576 GB/sec||to be announced later on||to be announced later on||960 GB/sec||to be announced later on|
|infinity cache||128 MB||32 MB||64 MB||96/192 MB||to be announced later on|
|Flagship SKU||Radeon RX 6900 XTX||Radeon RX 7600 XT?||Radeon RX 7800 XT?
Radeon RX 7700 XT?
|Radeon RX 7900 XT?||Radeon Pro|
|TBP||330 W||~ 150 watts||~ 250 watts||~ 350 W||to be announced later on|
|release||Fourth Quarter 2020||Fourth quarter 2022?||Fourth quarter 2022?||Fourth quarter 2022?||2023?|
Products mentioned in this post
“Communicator. Music aficionado. Certified bacon trailblazer. Travel advocate. Subtly charming social media fanatic.”