RDNA 2

RDNA 2
Launched	November 18, 2020
Designed by	AMD
Manufactured by	TSMC;
Fabrication process	TSMC N6; TSMC N7;
Codename(s)	Navi 2x; "Big Navi";
Product Series
Desktop	Radeon RX 6000 series;
Professional/workstation	Radeon Pro W6000 series;
Specifications
L0 cache	32 KB (per WGP): ; 16 KB vector data; 16 KB scalar data;
L1 cache	128 KB (per array)
L2 cache	1 MB to 4 MB
L3 cache	16 MB to 128 MB
Memory support	GDDR6
Memory clock rate	14–18 Gbps
PCIe support	PCIe 4.0
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 1.3
Media Engine
Encode codecs	H.264; H.265;
Decode codecs	H.264; H.265; AV1;
Color bit-depth	8-bit; 10-bit;
Encoder(s) supported	AMF; VCE;
Display outputs	DisplayPort 1.4a; HDMI 2.1; USB-C;
History
Predecessor	RDNA
Variant	CDNA 2 (datacenter)
Successor	RDNA 3

RDNA 2 is a GPU microarchitecture designed by AMD, released with the Radeon RX 6000 series on November 18, 2020. Alongside powering the RX 6000 series, RDNA 2 is also featured in the SoCs designed by AMD for the PlayStation 5, Xbox Series X/S, and Steam Deck consoles.

Background

On July 7, 2019, AMD released the first iteration of the RDNA microarchitecture, a new graphics architecture designed specifically for gaming that replaced the aging Graphics Core Next (GCN) microarchitecture. With RDNA, AMD sought to reduce latency and improve power efficiency over their previous Vega series based on GCN 5th gen and Nvidia's competing Turing microarchitecture.^[1]

RDNA 2 was first publicly announced in January 2020 with AMD initially calling RDNA 2 a "refresh" of the original RDNA architecture from the previous year.^[2] At AMD's Financial Analysts Day held on March 5, 2020, AMD showed a client GPU roadmap that gave details on RDNA's successor, RDNA 2, that it would again be built using TSMC's 7 nm process and would be coming in 2020.^[3] AMD told their investors that they were targeting a 50% uplift in performance-per-watt and increased IPC with the RDNA 2 microarchitecture.^[4]

On October 28, 2020, AMD held an online unveiling event for the RDNA 2 architecture and Radeon RX 6000 series.^[5]^[6] The event came 20 days after AMD's unveiling event for Ryzen 5000 series processors built on the Zen 3 microarchitecture.

Architectural details

Compute Unit

RDNA 2 contains a significant increase in the number of Compute Units (CUs) with a maximum of 80, a doubling from the maximum of 40 in the Radeon RX 5700 XT.^[1] Each Compute Unit contains 64 shader cores.^[7] CUs are organized into groups of two named Work Group Processors with 32 KB of shared L0 cache per WGP. Each CU contains two sets of an SIMD32 vector unit, an SISD scalar unit, textures units, and a stack of various caches.^[8] New low precision data types like INT4 and INT8 are new supported data types for RDNA 2 CUs.^[8]

The RDNA 2 graphics pipeline has been reconfigured and reordered for greater performance-per-watt and more efficient rendering by moving the caches closer to the shader engines. A new mesh shaders model allows shader rendering to be done in parallel using smaller batches of primitives called "meshlets". As a result, the mesh shaders feature enables greater control of the GPU geometry pipeline.^[9]

Ray tracing

Real-time hardware accelerated ray tracing is a new feature for RDNA 2 which is handled by a dedicated ray accelerator inside each CU.^[10] Ray tracing on RDNA 2 relies on the more open DirectX Raytracing protocol rather than the Nvidia RTX protocol.^[11]

In February 2023, it was reported that driver updates had boosted ray tracing performance by up to 40% using DirectX Raytracing.^[12]

Clock speeds

With RDNA 2 using the same 7 nm node as RDNA, AMD claims that RDNA 2 achieves a 30% frequency increase over its predecessor while using the same power.^[13]

Cache and memory subsystem

In addition to the traditional L1 and L2 caches that GPUs possess, RDNA 2 adds a new global L3 cache that AMD calls "Infinity Cache".^[14] This was done to avoid the use of a wider memory bus while still being able to maintain the same data bandwidth. Product technology architect Sam Naffziger said that, without Infinity Cache, "We were looking at the daunting prospect of having to put a 512-bit interface and all the power, area and expense associated with that".^[15] Using a wider memory bus requires more power which is in conflict with AMD's increased performance-per-watt goals for RDNA 2. AMD engineers ran tests comparing RDNA 2 silicon featuring a large on-die cache and with wider memory buses. They discovered that having such a cache would aid in the re-use of temporal and spatial data when the GPU is rendering a complex image. It is beneficial for the GPU's compute units to have fast access to a physically close cache rather than searching for data in video memory. AMD claims that RDNA 2's 128 MB of on-die Infinity Cache "dramatically reduces latency and power consumption".^[16] The GPU having access to a large L2 or L3 cache allows it to more quickly access necessary data compared to accessing VRAM or system RAM. The Infinity Cache is made up of two sets of 64 MB cache that can run on its own clock rate independent from the GPU cores. The Infinity Cache has a peak internal transfer bandwidth of 1986.6 GB/s and results in less reliance being placed on the GPU's GDDR6 memory controllers.^[8] Each Shader Engine now has two sets of L1 caches. The large cache of RDNA 2 GPUs give them a higher overall memory bandwidth compared to Nvidia's GeForce RTX 30 series GPUs.

Power efficiency

AMD claims that RDNA 2 achieves up to a 54% increase in performance-per-watt over the first RDNA microarchitecture.^[16] 21% of that 54% improvement is attributed to performance-per-clock enhancements, in part due to the addition of Infinity Cache.^[17]

Media engine

RDNA 2 uses the VCN 3.0, VCN 3.1, and VCN 3.1.2 video decoding blocks in its media engine.^[18]^[19] It adds support for AV1 decoding at up to 8K resolution, though AV1 hardware encoding support would not come until RDNA 3 in 2022.^[20]^[21] However, the low-end Navi 24 die and iGPUs based on RDNA 2.0 do not contain any media encoders and cannot decode AV1 as a result.^[18]

Navi 2x dies

		Navi 21^[22]	Navi 22^[23]	Navi 23^[24]	Navi 24^[25]
Launch		Nov 18, 2020	Mar 18, 2021	Aug 11, 2021	Jan 19, 2022
Codename		Sienna Cichlid	Navy Flounder	Dimgrey Cavefish	Beige Goby
Compute units (Stream processors) [FP32 cores]		80 (5120) [10240]	40 (2560) [5120]	32 (2048) [4096]	16 (1024) [2048]
Process		TSMC N7			TSMC N6
Transistors		26.8B	17.2B	11.06B	5.4B
Transistor density		51.5 MTr/mm²	51.3 MTr/mm²	46.7 MTr/mm²	50.5 MTr/mm²
Die size		520 mm²	335 mm²	237 mm²	107 mm²
Products	Desktop	RX 6800 RX 6800 XT RX 6900 XT RX 6950 XT	RX 6700 RX 6700 XT RX 6750 XT	RX 6600 RX 6600 XT RX 6650 XT	RX 6400 RX 6500 XT
	Mobile	—	RX 6800M RX 6850M XT	RX 6600S RX 6600M RX 6650M RX 6650M XT RX 6700S	RX 6300M RX 6450M RX 6500M RX 6550S RX 6550M
	Workstation (desktop)	W6800 W6800X W6800X Duo W6900X	—	W6600 W6600X	W6300 W6400
	Workstation (mobile)	—	—	W6600M	W6300M W6500M

Products

Desktop

Model (Code name)	Release Date & Price	Architecture & fab	Transistors & die size	Core		Fillrate^[a]^[b]^[c]		Processing power^[a]^[d] (GFLOPS)			Infinity Cache		Memory				TBP	Bus interface
Model (Code name)	Release Date & Price	Architecture & fab	Transistors & die size	Config^[e]	Clock^[a] (MHz)	Texture (GT/s)	Pixel (GP/s)	Half	Single	Double	Size	Bandwidth (GB/s)	Size	Bandwidth (GB/s)	Bus type & width	Clock (MT/s)	TBP	Bus interface
Radeon RX 6400 (Navi 24)^[26]^[27]^[f]	Jan 19, 2022 $159 USD	RDNA 2 TSMC N6	5.4×10⁹ 107 mm²	768:48:32:12 12 CU	1923 2321	92.3 111.4	61.5 74.3	5,907 7,130	2,954 3,565	184.6 222.8	16 MB	208	4 GB	128	GDDR6 64-bit	16000	53 W	PCIe 4.0 ×4
Radeon RX 6500 XT (Navi 24)^[28]^[29]^[30]^[f]	Jan 19, 2022 $199 USD (4GB) $219 USD (8GB)	RDNA 2 TSMC N6	5.4×10⁹ 107 mm²	1024:64:32:16 16 CU	2310 2815	147.8 180.2	73.9 90.1	9,462 11,530	4,731 5,765	295.6 360.3	16 MB	232	4 GB 8 GB	144	GDDR6 64-bit	18000	107 W 113 W	PCIe 4.0 ×4
Radeon RX 6600 (Navi 23)^[31]^[32]	Oct 13, 2021 $329 USD	RDNA 2 TSMC N7	11.06×10⁹ 237 mm²	1792:112:64:28 28 CU	1626 2491	182.3 279	104.1 159.4	11,658 17,860	5,828 8,928	364.2 558	32 MB	412.9	8 GB	224	GDDR6 128-bit	14000	132 W	PCIe 4.0 ×8
Radeon RX 6600 XT (Navi 23)^[33]^[34]^[35]^[36]	Aug 11, 2021 $379 USD			2048:128:64:32 32 CU	1968 2589	251.9 331.4	126 165.7	16,122 21,209	8,061 10,605	503.8 662.8		444.9		256		16000	160 W
Radeon RX 6650 XT (Navi 23)^[37]^[38]	May 10, 2022 $399 USD			2048:128:64:32 32 CU	2055 2635	263 337.2	131.5 168.6	16,835 21,586	8,417 10,793	526.1 674.6		468.9		280		17500	180 W
Radeon RX 6700 (Navi 22)^[39]^[40]^[41]^[42]	Jun 9, 2021		17.2×10⁹ 335 mm²	2304:144:64:36 36 CU	1941 2450	279.5 352.8	124.2 156.8	17,888 22,579	8,944 11,290	559 705.6	80 MB	1065	10 GB	320	GDDR6 160-bit	16000	175 W	PCIe 4.0 ×16
Radeon RX 6750 GRE 10GB (Navi 22)^[43]^[44]^[45]	Oct 18, 2023 $269 USD			2304:144:64:36 36 CU	1941 2450	279.5 352.8	124.2 156.8	17,888 22,579	8,944 11,290	559 705.6	80 MB	1065	10 GB	320	GDDR6 160-bit		170 W
Radeon RX 6700 XT (Navi 22)^[46]^[47]^[48]	Mar 18, 2021 $479 USD			2560:160:64:40 40 CU	2321 2581	371.4 413	148.5 165.2	23,767 26,429	11,884 13,215	742.7 825.9	96 MB	1278	12 GB	384	GDDR6 192-bit		230 W
Radeon RX 6750 GRE 12GB (Navi 22)^[49]^[50]^[45]	Oct 18, 2023 $289 USD				2321 2581	371.4 413	148.5 165.2	23,767 26,429	11,884 13,215	742.7 825.9		1278		384			230 W
Radeon RX 6750 XT (Navi 22)^[51]^[52]	May 10, 2022 $549 USD				2150 2600	344 416	137.6 166.4	22,016 26,624	11,008 13,312	688 832		1326		432		18000	250 W
Radeon RX 6800 (Navi 21)^[53]^[54]^[55]^[56]	Nov 18, 2020 $579 USD		26.8×10⁹ 520 mm²	3840:240:96:60 60 CU	1700 2105	408 505.2	163.2 202.1	26,112 32,333	13,056 16,166	816 1,010	128 MB	1432.6	16 GB	512	GDDR6 256-bit	16000	250 W
Radeon RX 6800 XT (Navi 21)^[53]^[57]^[55]^[56]	Nov 18, 2020 $649 USD			4608:288:128:72 72 CU	1825 2250	525.6 648	233.6 288	33,638 41,472	16,819 20,736	1,051 1,296		1664.2					300 W
Radeon RX 6900 XT (Navi 21)^[53]^[58]^[55]^[56]	Dec 8, 2020 $999 USD			5120:320:128:80 80 CU	1825 2250	584 720	233.6 288	37,376 46,080	18,688 23,040	1,168 1,440		1664.2					300 W
Radeon RX 6950 XT (Navi 21)^[59]^[60]	May 10, 2022 $1,099 USD			5120:320:128:80 80 CU	1890 2310	604.8 739.2	241.9 295.7	38,707 47,309	19,354 23,654	1,210 1,478		1793.5		576		18000	335 W

^ ^a ^b ^c Boost values (if available) are stated below the base value in italic.
^ Texture fillrate is calculated as the number of Texture Mapping Units multiplied by the base (or boost) core clock speed.
^ Pixel fillrate is calculated as the number of Render Output Units multiplied by the base (or boost) core clock speed.
^ Precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
^ Unified shaders : Texture mapping units : Render output units : Ray accelerators and Compute units (CU)
^ ^a ^b Navi 24 lacks hardware video encoder.

Mobile

Model (Code name)	Release date	Architecture & fab	Transistors & die size	Core		Fillrate^[a]^[b]^[c]		Processing power^[a]^[d] (GFLOPS)			Infinity Cache	Memory				HW Decoder			HW Encoder			TDP	Bus interface
Model (Code name)	Release date	Architecture & fab	Transistors & die size	Config^[e]	Clock^[a] (MHz)	Texture (GT/s)	Pixel (GP/s)	Half	Single	Double	Infinity Cache	Size	Bandwidth (GB/s)	Bus type & width	Clock (MT/s)	AV1	H265	4K H264	AV1	H265	4K H264	TDP	Bus interface
Radeon RX 6300M (Navi 24)^[f]^[61]	Jan 4, 2022	RDNA 2 TSMC N6	5.4×10⁹ 107 mm²	768:64:32:12 12 CU	1512	97.76	48.38	6,270	3,130	195.6	8 MB	2 GB	64	GDDR6 32-bit	16000							25 W	PCIe 4.0 ×4
Radeon RX 6450M (Navi 24)^[62]	Jan 4, 2023			768:64:32:12 12 CU	2220	118.10	71.04	7,600	3,780	236.3	16 MB	4 GB	128	GDDR6 64-bit								50 W
Radeon RX 6550S (Navi 24)^[63]	Jan 4, 2023			1024:64:32:16 16 CU	2170	154.20	69.44	9,900	4,900	306.3
Radeon RX 6500M (Navi 24)^[f]^[64]	Jan 4, 2022				2191	155.7	70.11	9,970	4,980	311.2
Radeon RX 6550M (Navi 24)^[65]	Jan 4, 2023				2560	182.10	81.92	11,600	5,800	362.5			144		18000							80 W
Radeon RX 6600S (Navi 23)^[66]	Jan 4, 2022	RDNA 2 TSMC N7	11.06×10⁹ 237 mm²	1792:128:64:28 28 CU	1881	244.2	120.3	15,630	7,810	448.1	32 MB		224	GDDR6 128-bit	14000								PCIe 4.0 ×8
Radeon RX 6700S (Navi 23)^[67]	Jan 4, 2022				1890	247.5	120.9	15,840	7,920	495.0		8 GB
Radeon RX 6600M (Navi 23)^[68]	May 31, 2021				2177	274.2	139.3	17,550	7,800	487.5												100 W
Radeon RX 6650M (Navi 23)^[69]	Jan 4, 2022				2222	276.6	139.3	17,700	8,850	553.1			256		16000							120 W
Radeon RX 6800S (Navi 23)^[70]				2048:128:64:32 32 CU	1975	288.0	134.4	18,430	9,220	576.5												100 W
Radeon RX 6650M XT (Navi 23)^[71]				2048:128:64:32 32 CU	2162	311.5	142.2	19,940	9,970	623.1												120 W
Radeon RX 6700M (Navi 22)^[72]	May 31, 2021		17.2×10⁹ 335 mm²	2304:144:64:32 36 CU	2300	331.4	147.2	21,209	10,605	662.1	80 MB	10 GB	320	GDDR6 160-bit								135 W	PCIe 4.0 ×16
Radeon RX 6800M (Navi 22)^[73]	May 31, 2021			2560:160:64:40 40 CU	2300	368.0	147.2	23,550	11,780	736.2	96 MB	12 GB	384	GDDR6 192-bit								145+ W
Radeon RX 6850M XT (Navi 22)^[74]	Jan 4, 2022			2560:160:64:40 40 CU	2463	415.6	157.6	26,430	13,210	825.6	96 MB	12 GB	432	GDDR6 192-bit	18000							165 W

^ ^a ^b ^c Boost values (if available) are stated below the base value in italic.
^ Texture fillrate is calculated as the number of Texture Mapping Units multiplied by the base (or boost) core clock speed.
^ Pixel fillrate is calculated as the number of Render Output Units multiplied by the base (or boost) core clock speed.
^ Precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
^ Unified shaders : Texture mapping units : Render output units : Ray accelerators and Compute units CU
^ ^a ^b Lacks hardware video encoder.

Workstation

Desktop Workstation

Model (Code name)	Release Date & Price	Architecture & fab	Transistors & die size	Core		Fillrate^[a]^[b]^[c]		Processing power^[a]^[d] (GFLOPS)			Infinity Cache	Memory				TDP	Bus interface	Graphic output ports
Model (Code name)	Release Date & Price	Architecture & fab	Transistors & die size	Config^[e]	Clock^[a] (MHz)	Texture (GT/s)	Pixel (GP/s)	Half	Single	Double	Infinity Cache	Size (GB)	Bandwidth (GB/s)	Bus type & width	Clock (MT/s)	TDP	Bus interface	Graphic output ports
Radeon Pro W6300 (Navi 24)^[75]^[76]	Oct 2022 OEM^[77]	RDNA 2 TSMC N6	5.4×10⁹ 107 mm²	768:48:32:12 12 CU	1512 2040	72.58 97.92	48.38 65.28	4,644 6,267	2,322 3,133	145.1 195.8	8 MB	2	64	GDDR6 32-bit	16000	25 W	PCIe 4.0 ×4	2× DP 1.4a
Radeon Pro W6400 (Navi 24)^[78]^[79]	Jan 19, 2022 $229 USD	RDNA 2 TSMC N6	5.4×10⁹ 107 mm²	768:48:32:12 12 CU	2039 2331	97.87 111.9	65.25 74.59	6,264 7,161	3,132 3,580	195.7 223.8	16 MB	4	128	GDDR6 64-bit	16000	50 W	PCIe 4.0 ×4	2× DP 1.4a
Radeon Pro W6600 (Navi 23)^[80]^[81]	Jun 8, 2021 $649 USD	RDNA 2 TSMC N7	11.06×10⁹ 237 mm²	1792:112:64:28 28 CU	2331 2903	261.1 325.1	149.2 185.8	16,709 20,809	8,354 10,404	522.1 650.3	32 MB	8	224	GDDR6 128-bit	14000	130 W	PCIe 4.0 ×8	4× DP 1.4a
Radeon Pro W6800 (Navi 21)^[82]^[83]	Jun 8, 2021 $2249 USD	RDNA 2 TSMC N7	26.8×10⁹ 520 mm²	3840:240:96:60 60 CU	2075 2320	498.0 556.8	199.2 222.7	31,872 35,635	15,936 17,818	996.0 1,114	128 MB	32	512	GDDR6 256-bit	16000	250 W	PCIe 4.0 ×16	6× miniDP 1.4a

^ ^a ^b ^c Boost values (if available) are stated below the base value in italic.
^ Texture fillrate is calculated as the number of Texture Mapping Units multiplied by the base (or boost) core clock speed.
^ Pixel fillrate is calculated as the number of Render Output Units multiplied by the base (or boost) core clock speed.
^ Precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
^ Unified shaders : Texture mapping units : Render output units : Ray accelerators and Compute units (CU)

Model
(Code name)

Release date

Architecture
& fab

Transistors
& die size

Core

Fillrate^[a]^[b]^[c]

Processing power^[a]^[d]
(GFLOPS)

Infinity
Cache

Memory

TDP

Bus
interface

Graphic output
ports

Config^[e]

Clock^[a]
(MHz)

Texture
(GT/s)

Pixel
(GP/s)

Half

Single

Double

Size
(GB)

Bandwidth
(GB/s)

Bus type
& width

Clock
(MT/s)

Radeon Pro W6600X
(Navi 23)^[84]^[85]

Mar 8, 2022

RDNA 2
TSMC N7

11.06×10⁹
237 mm²

2048:124:64:32
32 CU

2479

307.3

158.6

19,673

9,837

614.8

32 MB

8

256

GDDR6
128-bit

14000

?

PCIe 4.0
×8

2× HDMI 2.1

Radeon Pro W6800X
(Navi 21)^[84]^[86]^[87]

Aug 3, 2021

26.8×10⁹
520 mm²

3840:240:96:60
60 CU

1800
2087

432.0
500.8

172.8
200.3

27,648
32,056

13,824
16,028

864.0
1,002

128 MB

32

512

GDDR6
256-bit

16000

300 W

PCIe 4.0
×16

4× Thunderbolt 3
1× HDMI 2.1

Radeon Pro W6800X Duo
(Navi 21)^[84]^[88]

2×	26.8×10⁹ 520 mm²

2×	3840:240:96:60 60 CU

1800
1979

2×	432.0 474.9

2×	172.8 189.9

2×	27,648 30,397

2×	13,824 15,199

2×	864.0 949.9

2× 32

2× 512

GDDR6
2× 256-bit

400 W

Radeon Pro W6900X
(Navi 21)^[84]^[89]

26.8×10⁹
520 mm²

5120:320:128:80
80 CU

1825
2171

584.0
694.7

233.6
277.8

37,376
44,462

18,688
22,231

1,168
1,389

32

512