RDNA 4

RDNA 4 is a GPU microarchitecture designed by AMD, released with the Radeon RX 9000 series on February 28, 2025.^[1]

RDNA 4

Launched	February 28, 2025; 54 days ago (2025-02-28)
Designed by	AMD
Manufactured by	TSMC
Fabrication process	TSMC N4P
Product Series
Desktop	Radeon RX 9000 series
Specifications
Compute	Up to 97.3 TFLOPS (FP16) Up to 48.7 TFLOPS (FP32) Up to 0.76 TFLOPS (FP64)
Clock rate	2070 MHz to 2400 MHz
L0 cache	64 KB (per WGP)
L1 cache	128 KB (per array)
L2 cache	8 MB
L3 cache	64 MB
Memory support	GDDR6
Memory clock rate	up to 20 Gbps
PCIe support	PCIe 5.0
Supported Graphics APIs
Direct3D	Direct3D 12.0 Ultimate (feature level 12_2)
Shader Model	Shader Model 6.7
OpenCL	OpenCL 2.1
OpenGL	OpenGL 4.6
Vulkan	Vulkan 1.4
Media Engine
Encode codecs	H.264 H.265 AV1
Decode codecs	H.264 H.265 AV1
Color bit-depth	8-bit 10-bit 12-bit
Encoder(s) supported	AMF VCE
Display outputs	DisplayPort 2.1 HDMI 2.1a USB-C
History
Predecessor	RDNA 3
Successor	UDNA
Support status
Supported

Navi 4x dies

Navi 48
			Navi 48
Launch			Mar 2025
Codename
Compute units (Stream processors) [FP32 cores]			64 (4096) [8192]
Process			TSMC N4P
Transistors			53.9 bn.
Transistor density			151.1 MTr/mm2
Die size			356.5 mm2
Products	Consumer	Desktop	RX 9070 XT RX 9070
		Mobile	—
	Workstation	Desktop	—
		Mobile	—

Products

Gaming

Desktop

Model (Code name)	Release Date & Price	Architecture & fab	Transistors & die size	Core		Fillrate[i][ii][iii]		Processing power					Infinity Cache		Memory				TBP	Bus interface
								TFLOPS[i][iv]			AI TOPS[v]
				Config[vi]	Clock[i] (MHz)	Texture (GT/s)	Pixel (GP/s)	Half[vii]	Single	Double	INT8	INT4	Size	Bandwidth (GB/s)	Size	Bandwidth (GB/s)	Bus type & width	Clock (MT/s)
Radeon RX 9070 (Navi 48)[2]	Mar 6, 2025 $549 USD	RDNA 4 TSMC N4C	53.9 billion 356.5 mm2	3584:224:128:56:112 56 CU	2070 2520	463.6 564.4	265.0 322.6	59.35 72.25	29.67 36.12	0.463 0.564	237 289	475 578	64 MB		16 GB	640	GDDR6 256-bit	20000	220 W	PCIe 5.0 ×16
Radeon RX 9070 XT (Navi 48)[3]	Mar 6, 2025 $599 USD			4096:256:128:64:128 64 CU	2400 2970	614.4 760.3	307.2 380.2	78.64 97.32	39.32 48.66	0.614 0.760	314 389	629 778							304 W

• v
• t
• e

1. 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. Officially declared performance is twice the one shown here due to the use of sparsity.
6. Unified shaders : Texture mapping units : Render output units : Ray accelerators : AI accelerators and Compute Units (CU)
7. Officially declared half-precision performance is twice of the one shown here due to being based on different operation (a×b+c×d+e).