Engineering:RDNA 3

From HandWiki
Short description: GPU microarchitecture by AMD

Template:Infobox GPU microarchitecture RDNA 3 is a GPU microarchitecture designed by AMD, released with the Radeon RX 7000 series on December 13, 2022. Alongside powering the RX 7000 series, RDNA 3 is also featured in the SoCs designed by AMD for the Asus ROG Ally and Legion Go consoles.

Background

On June 9, 2022, AMD held their Financial Analyst Day where they presented a client GPU roadmap which contained mention of RDNA 3 coming in 2022 and RDNA 4 coming in 2024.[1] AMD announced to investors their intention to achieve a performance-per-watt uplift of over 50% with RDNA 3 and that the upcoming architecture would be built using chiplet packaging on a 5 nm process.[2]

A sneak preview for RDNA 3 was included towards the end of AMD's Ryzen 7000 unveiling event on August 29, 2022. The preview included RDNA 3 running gameplay of Lies of P, AMD CEO Lisa Su confirming that a chiplet design would be used, and a partial look at AMD's reference design for an RDNA 3 GPU.[3]

Full details for the RDNA 3 architecture were unveiled on November 3, 2022 at an event in Las Vegas.[4]

Architectural details

Chiplet packaging

For the first time ever in a consumer GPU, RDNA 3 utilizes modular chiplets. AMD previously had great success with its use of chiplets in its Ryzen desktop and Epyc server processors.[5]

The decision to move to a chiplet-based GPU microarchitecture was led by AMD Senior Vice President Sam Naffziger who had also lead the chiplet initiative with Ryzen and Epyc.[6] The development of RDNA 3's chiplet architecture began towards the end of 2017 with Naffziger leading the AMD graphics team in the effort.[7]

Memory Cache Dies (MCDs)

With a respective 2.05 billion transistors, each Memory Cache Die (MCD) contains large blocks of L3 cache and two physical 32-bit GDDR6 memory interfaces for a combined 64-bit interface per MCD.[8] The Radeon RX 7900 XTX has a 384-bit memory bus through the use of six MCDs while the RX 7900 XT has a 320-bit bus due to its five MCDs.

Chiplet interconnects

The chiplet interconnects have a bandwidth of 5.3 TB/s.[9]

Process node

According to Naffziger, cache and SRAM do not scale as linearly as logic does on advanced nodes like N5 in terms of density and power consumption so they can instead be fabricated on the cheaper, more mature N6 node. The use of smaller chiplet dies rather than one large monolithic die is beneficial for maximizing wafer yields as more dies can be fitted onto a single wafer.[7]

Compute Units

RDNA 3 includes improved dual-issue shader ALUs with the ability to execute two instructions per cycle. It can contain up to 96 graphics Compute Units that can provide up to 61 TFLOPS of compute.[10]

RDNA 3 has dedicated AI acceleration with Wave MMA (matrix multiply-accumulate) instructions,[11] which can improve AI-based performance by 2.7x and also benefits ray tracing instructions, similar to Nvidia's Tensor cores.[10]

Ray tracing

Each RDNA 3 Compute Unit contains one ray tracing accelerator. The overall number of ray tracing accelerators is increased due to the higher number of Compute Units, though the number of ray tracing accelerators per Compute Unit has not increased over RDNA 2.

Clock speeds

RDNA 3 was designed to support high clock speeds. On RDNA 3, clock speeds have been decoupled with the front end operating at a 2.5 GHz frequency while the shaders operate at 2.3 GHz. The shaders operating at a lower clock speed gives up to 25% power savings according to AMD and RDNA 3's shader clock speed is still 15% faster than RDNA 2.[12]

Cache and memory subsystem

RDNA 3 GPUs use GDDR6 memory rather than faster GDDR6X due to the latter's increased power consumption.

16 MB Infinity Cache is included on each MCD. Theoretically, additional L3 cache could be added to the MCDs via AMD's 3D V-Cache die stacking technology as the MCDs contain unused TSV connection points.[13][14]

Power efficiency

AMD claims that RDNA 3 achieves a 54% increase in performance-per-watt which is in line with their previous claims of 50% performance-per-watt increases for both RDNA and RDNA 2.

Media engine

RDNA 3 is the first RDNA architecture to have a dedicated media engine. It is built into the GCD and is based on VCN 4.0 encoding and decoding core.[15] AMD's AMF AV1 encoder is comparable in quality to Nvidia's NVENC AV1 encoder but can handle a higher number of simultaneous encoding streams compared to the limit of 3 on the GeForce RTX 40 series.[16]

Supported encoding frame rates (FPS) per resolution and video coding format[17]
Resolution H.264 H.265 AV1
1080p
1440p
4K 180 180 240
8K 48 48 60

Display engine

RDNA 3 GPUs feature a new display engine called the "Radiance Display Engine". AMD touted its support for DisplayPort 2.1 UHBR 13.5, delivering up to 54 Gbit/s bandwidth for high refresh rates at 4K and 8K resolutions.[18] The Radeon Pro W7900 and W7800 support UHBR 20. DisplayPort 2.1 can support 4K at 480 Hz and 8K at 165 Hz with Display Stream Compression (DSC). The previous DisplayPort 1.4 standard with DSC was limited to 4K at 240 Hz and 8K at 60 Hz.

Navi 3x dies

Graphics Compute Die
(GCD) 		Memory Cache Die
(MCD)
Navi 31 Navi 32 Navi 33
Ref. [8]
Launch 2022 2023 2023 2022
Codename Plum Bonito Wheat Nas Hotpink Bonefish N/A
Compute units
(Stream processors)
[FP32 cores] 		96
(6144)
[12288] 		60
(3840)
[7680] 		32
(2048)
[4096]
Process TSMC N5 TSMC N6
Transistors 45.4B 19.9B 13.3B 2.05B
Transistor density 150.2 MTr/mm2 101.5 MTr/mm2 65.2 MTr/mm2 54.64 MTr/mm2
Die size 304.35 mm2 196 mm2 204 mm2 37.52 mm2
Max TDP 405 W 263 W 200 W N/A
Products Consumer Desktop
  • RX 7900 GRE
  • RX 7900 XT
  • RX 7900 XTX
  • RX 7700 XT
  • RX 7800 XT
  • RX 7600
  • RX 7700 XT (3×)
  • RX 7800 XT (4×)
  • RX 7900 GRE (4×)
  • RX 7900 XT (5×)
  • RX 7900 XTX (6×)
Mobile
  • RX 7900M
 N/A
  • RX 7600S
  • RX 7600M
  • RX 7600M XT
  • RX 7700S
  • RX 7900M (4×)
Workstation Desktop
  • W7800
  • W7900
  • W7700
  • W7500
  • W7600
  • W7700 (4×)
  • W7800 (4×)
  • W7900 (6×)
Mobile N/A N/A N/A N/A

Products

Desktop

Mobile

Workstation

Desktop Workstation

Integrated graphics processing units (iGPUs)

Model Launch Codename Architecture
& fab 		Die
size 		Core Fillrate[lower-alpha 1][lower-alpha 2][lower-alpha 3] Processing power[lower-alpha 1][lower-alpha 4]
(GFLOPS) 		Cache TDP Bus
interface
Config[lower-alpha 5][lower-alpha 6] Clock[lower-alpha 1]
(MHz) 		Texture
(GT/s) 		Pixel
(GP/s) 		Half Single Double L0 L1 L2
Radeon 740M 2023 Phoenix RDNA 3
TSMC N4 		178 mm2 4 CU
256:16:8:4 		2,500 40.0 20.0 5,120 2,560 80.0 64 KB 512 KB 2 MB 15–30 W PCIe 4.0
×8
Ryzen Z1 2,735 43.7 21.8 5,600 2,800 87.5 9–30 W
Radeon 760M 8 CU
512:32:16:8 		1,000
2,600 		32.0
83.2 		21.3
55.5 		4,096
10,649 		2,048
5,324 		64.0
166.4 		128 KB 1 MB 35–54 W
Radeon 780M 12 CU
768:48:24:12 		2,700
2,800 		129.6
134.4 		64.8
67.2 		16,588
17,203 		8,294
8,600 		259.2
268.8 		192 KB 1.5 MB 35–54 W
Ryzen Z1 Extreme 2,800 134.4 67.2 17,203 8,600 268.8 9–30 W
  1. ↑ 1.0 1.1 1.2 Boost values (if available) are stated below the base value in italic.
  2. ↑ Texture fillrate is calculated as the number of Texture Mapping Units multiplied by the base (or boost) core clock speed.
  3. ↑ Pixel fillrate is calculated as the number of Render Output Units multiplied by the base (or boost) core clock speed.
  4. ↑ Precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  5. ↑ Unified shaders : Texture mapping units : Render output units : Ray accelerators and Compute units (CU)
  6. ↑ GPUs based on RDNA 3 have dual-issue stream processors so that up to two shader instructions can be executed per clock cycle under certain parallelism conditions.

References

  1. ↑ Smith, Ryan (June 9, 2022). "AMD's 2022-2024 Client GPU Roadmap: RDNA 3 This Year, RDNA 4 Lands in 2024" (in en-US). https://www.anandtech.com/show/17443/amds-client-gpu-roadmap-rdna-3-this-year-rdna-4-lands-in-2024. 
  2. ↑ Walton, Jarred (June 9, 2022). "AMD GPU Roadmap: RDNA 3 With 5nm GPU Chiplets Coming This Year" (in en-US). https://www.tomshardware.com/news/amd-rdna3-roadmap-chiplets-5nm. 
  3. ↑ Wickens, Katie (August 31, 2022). "AMD's Lisa Su confirms chiplet-based RDNA 3 GPU architecture" (in en-US). https://www.pcgamer.com/amds-lisa-su-confirms-chiplet-based-rdna-3-gpu-architecture/. 
  4. ↑ "AMD Unveils World's Most Advanced Gaming Graphics Cards, Built on Groundbreaking AMD RDNA 3 Architecture with Chiplet Design". AMD (Press release). Las Vegas, NV. November 3, 2022. Retrieved April 8, 2023.
  5. ↑ James, Dave (June 24, 2022). "AMD suggests a Ryzen-like design for RDNA 3 chiplets would be 'a reasonable inference'" (in en-US). https://www.pcgamer.com/amd-suggests-a-ryzen-like-chiplet-design-for-rdna-3-gpus-would-be-a-reasonable-inference/. 
  6. ↑ Alcorn, Paul; Walton, Jarred (June 23, 2022). "Into the GPU Chiplet Era: An Interview With AMD's Sam Naffziger" (in en-US). https://www.tomshardware.com/features/gpu-chiplet-era-interview-amd-sam-naffziger. 
  7. ↑ 7.0 7.1 Brosdahl, Peter (November 22, 2022). "AMD Lead Engineer Sam Naffziger Explains Advantages of RDNA3 Chiplet Design" (in en-US). https://www.thefpsreview.com/2022/11/22/amd-lead-engineer-sam-naffziger-explains-advantages-of-rdna3-chiplet-design/. 
  8. ↑ 8.0 8.1 Walton, Jarred (November 14, 2022). "AMD RDNA 3 GPU Architecture Deep Dive: The Ryzen Moment for GPUs" (in en-US). https://www.tomshardware.com/news/amd-rdna-3-gpu-architecture-deep-dive-the-ryzen-moment-for-gpus. 
  9. ↑ "AMD Explains the Economics Behind Chiplets for GPUs" (in en-US). November 14, 2022. https://www.techpowerup.com/301071/amd-explains-the-economics-behind-chiplets-for-gpus. 
  10. ↑ 10.0 10.1 Gula, Damien (November 3, 2022). "AMD's RDNA 3 GPUs are Way Cheaper Than the RTX 4090" (in en-US). https://gizmodo.com/amd-rdna-3-gpu-rtx-4090-4080-rx-7900-xtx-xt-price-date-1849741000. 
  11. ↑ Vasishta, Aaryaman (January 10, 2023). "How to accelerate AI applications on RDNA 3 using WMMA". GPUOpen.com. https://gpuopen.com/learn/wmma_on_rdna3/. 
  12. ↑ Olšan, Jan (November 7, 2022). "AMD RDNA 3 details: architecture changes, AI acceleration, DP 2.1" (in en-GB). https://www.hwcooling.net/en/amd-rdna-3-details-architecture-changes-ai-acceleration-dp-2-1-en/. 
  13. ↑ Klotz, Aaron (January 29, 2023). "AMD GPU Appears to Leave Room for Future 3D V-Cache" (in en-US). https://www.tomshardware.com/news/3d-vcache-rdna3-amd-gpu. 
  14. ↑ Ridley, Jacob (January 30, 2023). "Tiny spots on AMD's RDNA 3 GPU hint at massive cache potential" (in en-US). https://www.pcgamer.com/tiny-spots-on-amds-rdna-3-gpu-hint-at-massive-cache-potential/. 
  15. ↑ Shilov, Anton (May 4, 2022). "First Details About AMD’s Next Generation Video Engine Revealed" (in en-US). https://www.tomshardware.com/news/next-amd-video-engine-may-lack-av1. 
  16. ↑ Klotz, Aaron (December 12, 2022). "AMD's Radeon RX 7900 AV1 encoder is almost on par with Intel Arc and Nvidia's RTX 40 series" (in en-US). https://www.techspot.com/news/96945-amd-radeon-rx-7900-av1-encoder-almost-par.html. 
  17. ↑ Taylor, Adam (December 14, 2022). "Tested: With RDNA 3, AMD Radeon is finally useful for content creators" (in en-US). https://www.pcworld.com/article/1434166/amd-rdna-3-radeon-rx-7900-xtx-content-creation-review.html. 
  18. ↑ Sag, Anshel (November 14, 2022). "AMD's New Radeon RX 7900XTX And 7900XT Put The Pressure On NVIDIA" (in en-US). https://www.forbes.com/sites/moorinsights/2022/11/14/amds-new-radeon-rx-7900xtx-and-7900xt-put-the-pressure-on-nvidia/?sh=21776d571aa3. 



Retrieved from "https://handwiki.org/wiki/index.php?title=Engineering:RDNA_3&oldid=3236998"