Radeon 300 series

The Radeon 300 series is a series of graphics processors developed by AMD. All of the GPUs of the series are produced in 28 nm format and use the Graphics Core Next (GCN) micro-architecture.

Radeon 300 series

Release date	June 16, 2015; 9 years ago (June 16, 2015)
Codename	Caribbean Islands[1] Sea Islands Volcanic Islands
Architecture	GCN 1st gen GCN 2nd gen GCN 3rd gen
Transistors	690M (Exo) 28 nm 950M (Oland) 28 nm 1.500M (Cape Verde) 28 nm 1.550M (Meso) 28 nm 2.080M (Bonaire) 28 nm 2.800M (Pitcairn) 28 nm 5.000M (Tonga) 28 nm 6.200M (Grenada) 28 nm 8.900M (Fiji) 28 nm
Cards
Entry-level	Radeon R5 310 Radeon R5 330 Radeon R5 340 Radeon R5 340X Radeon R7 340 Radeon R7 350 Radeon R7 350X
Mid-range	Radeon R7 360 Radeon R7 370 Radeon R9 360 Radeon R9 370 Radeon R9 370X Radeon R9 380 Radeon R9 380X
High-end	Radeon R9 390 Radeon R9 390X
Enthusiast	Radeon R9 390 X2 Radeon R9 Nano Radeon R9 Fury Radeon R9 Fury X Radeon Pro Duo
API support
DirectX	Direct3D 12.0 (feature level 12_0) [2] Shader Model 6.5
OpenCL	OpenCL 2.1
OpenGL	OpenGL 4.5 OpenGL 4.6 (Windows 7+ and Adrenalin 18.4.1+, Linux)[3][4][5][6][7]
Vulkan	Vulkan 1.2[8] (Windows) Vulkan 1.3 (Linux) SPIR-V
History
Predecessor	Radeon 200 series
Successor	Radeon 400 series
Support status
Unsupported

The series includes the Fiji and Tonga GPU dies based on AMD's GCN 3 or "Volcanic Islands" architecture, which had originally been introduced with the Tonga based (though cut-down) R9 285 slightly earlier. Some of the cards in the series include the Fiji based flagship AMD Radeon R9 Fury X, cut-down Radeon R9 Fury and small form factor Radeon R9 Nano,^[9] which are the first GPUs to feature High Bandwidth Memory (HBM) technology, which AMD co-developed in partnership with SK Hynix. HBM is faster and more power efficient than GDDR5 memory, though also more expensive.^[10] However, the remaining GPUs in the series outside the Tonga based R9 380 and R9 380X are based on previous generation GPUs with revised power management, and therefore only feature GDDR5 memory (something Tonga does as well). The Radeon 300 series cards including the R9 390X were released on June 18, 2015. The flagship device, the Radeon R9 Fury X, was released on June 24, 2015, with the dual-GPU variant, the Radeon Pro Duo, being released on April 26, 2016.^[11]

Micro-architecture and instruction set

The R9 380/X along with the R9 Fury & Nano series were AMD's first cards (after the earlier R9 285) to use the third iteration of their GCN instruction set and micro-architecture. The other cards in the series feature first and second gen iterations of GCN. The table below details which GCN-generation each chip belongs to.

AMD Fiji with HBM

Ancillary ASICs

Any ancillary ASICs present on the chips are being developed independently of the core architecture and have their own version name schemes.

Multi-monitor support

Main article: AMD Eyefinity

The AMD Eyefinity branded on-die display controllers were introduced in September 2009 in the Radeon HD 5000 series and have been present in all products since.^[12]

AMD TrueAudio

Main article: AMD TrueAudio

AMD TrueAudio was introduced with the AMD Radeon Rx 200 series, but can only be found on the dies of GCN 2nd gen and later products.

Video acceleration

AMD's SIP core for video acceleration, Unified Video Decoder and Video Coding Engine, are found on all GPUs and are supported by AMD Catalyst and by the open-source Radeon graphics driver.

Frame limiter

A new feature to the lineup allows users to reduce power consumption by not rendering unnecessary frames. It is user configurable.

LiquidVR support

LiquidVR is a technology that improves the smoothness of virtual reality. The aim is to reduce latency between hardware so that the hardware can keep up with the user's head movement, eliminating the motion sickness. A particular focus is on dual GPU setups where each GPU now renders for one eye individually of the display.

Virtual super resolution support

Originally introduced with the previous generation R9 285 and R9 290 series graphics cards, this feature allows users to run games with higher image quality by rendering frames at above native resolution. Each frame is then downsampled to native resolution. This process is an alternative to supersampling which is not supported by all games. Virtual super resolution is similar to Dynamic Super Resolution, a feature available on competing Nvidia graphics cards, but trades flexibility for increased performance.^[13]

OpenCL (API)

OpenCL accelerates many scientific Software Packages against CPU up to factor 10 or 100 and more. Open CL 1.0 to 1.2 are supported for all chips with Terascale and GCN Architecture. OpenCL 2.0 is supported with GCN 2nd Gen. and higher. ^[14] For OpenCL 2.1 and 2.2 only Driver Updates are necessary with OpenCL 2.0 conformant Cards.

Vulkan (API)

API Vulkan 1.0 is supported for all GCN architecture cards. Vulkan 1.2 requires GCN 2nd gen or higher with the Adrenalin 20.1 and Linux Mesa 20.0 drivers and newer.

Chipset tables

Desktop models

Model (Codename)	Release Date & Price	Architecture (Fab)	Transistors Die Size	Core		Fillrate[a][b][c]		Processing power[a][d] (GFLOPS)		Memory				TBP (W)	Bus interface
				Config[e]	Clock[a] (MHz)	Texture (GT/s)	Pixel (GP/s)	Single	Double	Size (MiB)	Bus type & width	Clock (MT/s)	Band- width (GB/s)
Radeon R5 330 (Oland Pro)	May 2015 OEM	GCN 1st gen (28 nm)	1040×106 90 mm2	320:20:8	Un­known 855	17.1	6.84	547.2	34.2	1024 2048	DDR3 128-bit	1800	28.8	30	PCIe 3.0 x4 x8 ×16
Radeon R5 340 (Oland XT)	May 2015 OEM			384:24:8	Un­known 825	19.8	6.6	633.6	39.6	1024 2048	DDR3 GDDR5 128-bit	1800 4500	28.8 72	75
Radeon R7 340 (Oland XT)	May 2015 OEM			384:24:8	730 780	17.5 18.7	5.8 6.2	560.6 599	32.7 35	1024 2048 4096	DDR3 GDDR5 128-bit	1800 4500	28.8 72	75
Radeon R5 340X[15] (Oland XT)	May 2015 OEM			384:24:8	1050	25.2	8.4	806	50.4	2048	DDR3 64-bit	2000	16	65
Radeon R7 350 (Oland XT)	May 2015 OEM			384:24:8	1000 1050	24 25.2	8 8.4	768 806.4	48 50.4	1024 2048	DDR3 GDDR5 128-bit	1800 4500	28.8 72	75
Radeon R7 350 [16] (Cape Verde XTL)	February 2016 $89 USD		1500×106 123 mm2	512:32:16	925	29.6	14.8	947.2	59.2	2048	GDDR5 128-bit	4500	72	75
Radeon R7 350X[17] (Oland XT)	May 2015 OEM		1040×106 90 mm2	384:24:8	1050	25.2	8.4	806	50.4	4096	DDR3 128-bit	2000	32	30
Radeon R7 360[18][19] (Bonaire Pro)	June 2015 $109 USD	GCN 2nd gen (28 nm)	2080×106 160 mm2	768:48:16	1050	50.4	16.8	1612.8	100.8	2048	GDDR5 128-bit	6500	104	100
Radeon R9 360 (Bonaire Pro)	May 2015 OEM			768:48:16	1000 1050	48 50.4	16 16.8	1536 1612.8	96 100.8	2048	GDDR5 128-bit	6500	104	85
Radeon R7 370[18] (Pitcairn Pro)	June 2015 $149 USD	GCN 1st gen (28 nm)	2800×106 212 mm2	1024:64:32	975	62.4	31.2	1996.8	124.8	2048 4096	GDDR5 256-bit	5600	179.2	110
Radeon R9 370 (Pitcairn Pro)	May 2015 OEM			1024:64:32	950 975	60.8 62.4	30.4 31.2	1945.6 1996.8	121.6 124.8	2048 4096	GDDR5 256-bit	5600	179.2	150
Radeon R9 370X (Pitcairn XT)	August 2015 $179 USD			1280:80:32	1000	80	32	2560	160	2048 4096	GDDR5 256-bit	5600	179.2	185
Radeon R9 380 (Tonga Pro)	May 2015 OEM	GCN 3rd gen (28 nm)	5000×106 359 mm2	1792:112:32	918	102.8	29.4	3290	206.6	4096	GDDR5 256-bit	5500	176	190
Radeon R9 380[20] (Tonga Pro)	June 2015 $199 USD			1792:112:32	970	108.6	31.0	3476.5	217.3	2048 4096	GDDR5 256-bit	5700	182.4[f]	190
Radeon R9 380X[20] (Tonga XT)	November 2015 $229 USD			2048:128:32	970	124.2	31.0	3973.1	248.3	4096	GDDR5 256-bit	5700	182.4	190
Radeon R9 390[20] (Grenada Pro)	June 2015 $329 USD	GCN 2nd gen (28 nm)	6200×106 438 mm2	2560:160:64	1000	160	64	5120	640	8192	GDDR5 512-bit	6000	384	275
Radeon R9 390X[20] (Grenada XT)	June 2015 $429 USD			2816:176:64	1050	184.8	67.2	5913.6	739.2	8192	GDDR5 512-bit	6000	384	275
Radeon R9 Fury[21] (Fiji Pro)	July 2015 $549 USD	GCN 3rd gen (28 nm)	8900×106 596 mm2	3584:224:64	1000	224	64	7168	448	4096	HBM 4096-bit	1000	512	275
Radeon R9 Nano[22] (Fiji XT)	August 2015 $649 USD			4096:256:64	1000	256	64	8192	512					175
Radeon R9 Fury X[20][23] (Fiji XT)	June 2015 $649 USD			4096:256:64	1050	268.8	67.2	8601.6	537.6					275
Radeon Pro Duo[24][25][26][27] (Fiji XT)	April 2016 $1499 USD		2× 8900×106 2× 596 mm2	2× 4096:256:64	1000	512	128	16384	1024	2× 4096	HBM 4096-bit	1000	2× 512	350
Model (Codename)	Release Date & Price	Architecture (Fab)	Transistors Die Size	Config[e]	Clock[a] (MHz)	Texture (GT/s)	Pixel (GP/s)	Single	Double	Size (MiB)	Bus type & width	Clock (MT/s)	Band- width (GB/s)	TBP (W)	Bus interface
				Core		Fillrate[a][b][c]		Processing power[a][d] (GFLOPS)		Memory

• 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. Double precision performance of Hawaii cards is 1/8 of single precision performance, for the other it is 1/16 of single precision performance.
5. Unified Shaders : Texture Mapping Units : Render Output Units
6. The R9 380 utilizes loss-less color compression which can increase effective memory performance (relative to GCN 1^st gen and 2^nd gen cards) in certain situations.^{[citation needed]}

Mobile models

Model (Codename)	Launch	Architecture (Fab)	Core		Fillrate[a][b][c]		Processing power[a][d] (GFLOPS)	Memory				TDP
			Config[e]	Clock[a] (MHz)	Texture (GT/s)	Pixel (GP/s)		Size (GiB)	Bus type & width	Clock (MT/s)	Band- width (GB/s)
Radeon R5 M330[28] (Exo Pro)	2015	GCN 1st gen (28 nm)	320:20:8	Un­known 1030	8.2	20.6	659.2	2 4	DDR3 64-bit	1800 2000	14.4 16	18 W
Radeon R5 M335[28] (Exo Pro)	2015		320:20:8	Un­known 1070	8.6	21.4	684.8	2 4	DDR3 64-bit	2200	17.6	Un­known
Radeon R7 M360[29] (Meso XT)	2015		384:24:8	Un­known 1125	9	27	864	2 4	DDR3 64-bit	2000	16	Un­known
Radeon R9 M365X[30] (Strato Pro)	2015		640:40:16	Un­known 925	14.8	37	1184	4	GDDR5 128-bit	4500	72	50 W
Radeon R9 M370X[30] (Strato Pro)	May 2015		640:40:16	800	12.8	32	1024	2	GDDR5 128-bit	4500	72	40–45 W
Radeon R9 M375[30] (Strato Pro)	2015		640:40:16	Un­known 1015	16.2	40.6	1299.2	4	GDDR5 128-bit	4400	35.2	Un­known
Radeon R9 M375X[30] (Strato Pro)	2015		640:40:16	Un­known 1015	16.2	40.6	1299.2	4	GDDR5 128-bit	4500	72	Un­known
Radeon R9 M380[30] (Strato Pro)	2015		640:40:16	Un­known 900	14.4	36	1152	4	GDDR5 128-bit	6000	96	Un­known
Radeon R9 M385X[30] (Strato)	2015	GCN 2nd gen (28 nm)	896:56:16	Un­known 1100	17.6	61.6	1971.2	4	GDDR5 128-bit	6000	96	~75 W
Radeon R9 M390[30] (Pitcairn)	June 2015	GCN 1st gen (28 nm)	1024:64:32	Un­known 958	30.7	61.3	1962	2	GDDR5 256-bit	5460	174.7	~100 W
Radeon R9 M390X[30] (Amethyst XT)	2015	GCN 3rd gen (28 nm)	2048:128:32	Un­known 723	23.1	92.5	2961.4	4	GDDR5 256-bit	5000	160	125 W
Radeon R9 M395[30] (Amethyst Pro)	2015		1792:112:32	Un­known 834	26.6	93.4	2989.0	2	GDDR5 256-bit	5460	174.7	125 W
Radeon R9 M395X[30] Amethyst XT)	2015		2048:128:32	Un­known 909	29.1	116.3	3723.3	4	GDDR5 256-bit	5460	174.7	125 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. Unified Shaders : Texture Mapping Units : Render Output Units

Radeon Feature Matrix

The following table shows features of AMD/ATI's GPUs (see also: List of AMD graphics processing units).

[ VisualEditor ]

• view
• talk
• edit

Name of GPU series	Wonder	Mach	3D Rage	Rage Pro	Rage 128	R100	R200	R300	R400	R500	R600	RV670	R700	Evergreen	Northern Islands		Southern Islands	Sea Islands	Volcanic Islands	Arctic Islands/Polaris	Vega	Navi 1x	Navi 2x	Navi 3x	Navi 4x
Released	1986	1991	Apr 1996	Mar 1997	Aug 1998	Apr 2000	Aug 2001	Sep 2002	May 2004	Oct 2005	May 2007	Nov 2007	Jun 2008	Sep 2009	Oct 2010	Dec 2010	Jan 2012	Sep 2013	Jun 2015	Jun 2016, Apr 2017, Aug 2019	Jun 2017, Feb 2019	Jul 2019	Nov 2020	Dec 2022	Feb 2025
Marketing Name	Wonder	Mach	3D Rage	Rage Pro	Rage 128	Radeon 7000	Radeon 8000	Radeon 9000	Radeon X700/X800	Radeon X1000	Radeon HD 2000	Radeon HD 3000	Radeon HD 4000	Radeon HD 5000	Radeon HD 6000		Radeon HD 7000	Radeon 200	Radeon 300	Radeon 400/500/600	Radeon RX Vega, Radeon VII	Radeon RX 5000	Radeon RX 6000	Radeon RX 7000	Radeon RX 9000
AMD support
Kind	2D		3D
Instruction set architecture	Not publicly known										TeraScale instruction set						GCN instruction set					RDNA instruction set
Microarchitecture	Not publicly known					GFX1		GFX2			TeraScale 1 (VLIW5) (GFX3)			TeraScale 2 (VLIW5) (GFX4)	TeraScale 2 (VLIW5) up to 68xx (GFX4)	TeraScale 3 (VLIW4) in 69xx [31][32] (GFX5)	GCN 1st gen (GFX6)	GCN 2nd gen (GFX7)	GCN 3rd gen (GFX8)	GCN 4th gen (GFX8)	GCN 5th gen (GFX9)	RDNA (GFX10.1)	RDNA 2 (GFX10.3)	RDNA 3 (GFX11)	RDNA 4 (GFX12)
Type	Fixed pipeline[a]						Programmable pixel & vertex pipelines				Unified shader model
Direct3D	—		5.0	6.0		7.0	8.1	9.0 11 (9_2)	9.0b 11 (9_2)	9.0c 11 (9_3)	10.0 11 (10_0)	10.1 11 (10_1)		11 (11_0)			11 (11_1) 12 (11_1)	11 (12_0) 12 (12_0)			11 (12_1) 12 (12_1)		11 (12_1) 12 (12_2)
Shader model	—						1.4	2.0+	2.0b	3.0	4.0	4.1		5.0			5.1	5.1 6.5			6.7				6.8
OpenGL	—			1.1	1.2	1.3		2.1[b][33]			3.3			4.5[34][35][36][c]			4.6
Vulkan	—															1.1[c][d]	1.3[37]		1.4[38]
OpenCL	—										Close to Metal		1.1 (not supported by Mesa)	1.2+ (on Linux: 1.1+ (no Image support on clover, with by rustiCL) with Mesa, 1.2+ on GCN 1.Gen)				2.0+ (Adrenalin driver on Win7+) (on Linux ROCM, Mesa 1.2+ (no Image support in clover, but in rustiCL with Mesa, 2.0+ and 3.0 with AMD drivers or AMD ROCm), 5th gen: 2.2 win 10+ and Linux RocM 5.0+				2.2+ and 3.0 windows 8.1+ and Linux ROCM 5.0+ (Mesa rustiCL 1.2+ and 3.0 (2.1+ and 2.2+ wip))[39][40][41]
HSA / ROCm	—																					?
Video decoding ASIC	—										Avivo/UVD	UVD+	UVD 2	UVD 2.2	UVD 3		UVD 4	UVD 4.2	UVD 5.0 or 6.0	UVD 6.3	UVD 7 [42][e]	VCN 2.0 [42][e]	VCN 3.0 [43]	VCN 4.0	VCN 5.0
Video encoding ASIC	—																VCE 1.0	VCE 2.0	VCE 3.0 or 3.1	VCE 3.4	VCE 4.0 [42][e]
Fluid Motion [f]																								?
Power saving	?										PowerPlay				PowerTune	PowerTune & ZeroCore Power					?
TrueAudio	—																	Via dedicated DSP		Via shaders
FreeSync	—																	1 2
HDCP[g]	—										?							1.4		2.2		2.3 [44]
PlayReady[g]	—																			3.0		3.0
Supported displays[h]						1–2	2							2–6								?			4
Max. resolution	?													2–6 × 2560×1600			2–6 × 4096×2160 @ 30 Hz			2–6 × 5120×2880 @ 60 Hz	3 × 7680×4320 @ 60 Hz [45]	7680×4320 @ 60 Hz PowerColor		7680x4320 @165 Hz	7680x4320
/drm/radeon[i]																			—
/drm/amdgpu[i]	—																	Optional [46]

1. The Radeon 100 Series has programmable pixel shaders, but do not fully comply with DirectX 8 or Pixel Shader 1.0. See article on R100's pixel shaders.
2. R300, R400 and R500 based cards do not fully comply with OpenGL 2+ as the hardware does not support all types of non-power of two (NPOT) textures.
3. OpenGL 4+ compliance requires supporting FP64 shaders and these are emulated on some TeraScale chips using 32-bit hardware.
4. Vulkan support is theoretically possible but has not been implemented in a stable driver.
5. The UVD and VCE were replaced by the Video Core Next (VCN) ASIC in the Raven Ridge APU implementation of Vega.
6. Video processing for video frame rate interpolation technique. In Windows it works as a DirectShow filter in your player. In Linux, there is no support on the part of drivers and / or community.
7. To play protected video content, it also requires card, operating system, driver, and application support. A compatible HDCP display is also needed for this. HDCP is mandatory for the output of certain audio formats, placing additional constraints on the multimedia setup.
8. More displays may be supported with native DisplayPort connections, or splitting the maximum resolution between multiple monitors with active converters.
9. DRM (Direct Rendering Manager) is a component of the Linux kernel. AMDgpu is the Linux kernel module. Support in this table refers to the most current version.

Graphics device drivers

Proprietary graphics device driver Catalyst

Main article: AMD Catalyst

AMD Catalyst is being developed for Microsoft Windows and Linux. As of July 2014, other operating systems are not officially supported. This may be different for the AMD FirePro brand, which is based on identical hardware but features OpenGL-certified graphics device drivers.

AMD Catalyst supports all features advertised for the Radeon brand.

Free and open-source graphics device driver radeon

Main article: free and open-source "Radeon" graphics device driver

The free and open-source drivers are primarily developed on and for Linux, but have been ported to other operating systems as well. Each driver is composed out of five parts:

1. Linux kernel component DRM
2. Linux kernel component KMS driver: basically the device driver for the display controller
3. user-space component libDRM
4. user-space component in Mesa 3D
5. a special and distinct 2D graphics device driver for X.Org Server, which is finally about to be replaced by Glamor

The free and open-source radeon kernel driver supports most of the features implemented into the Radeon line of GPUs.^[6]

The radeon kernel driver is not reverse engineered, but based on documentation released by AMD.^[47] This driver still requires proprietary microcode to operate DRM functions and some GPUs may fail to launch the X server if not available.

Free and open-source graphics device driver amdgpu

This new kernel driver is directly supported and developed by AMD. It is available on various Linux distributions, and has been ported to some other operating systems as well. Only GCN GPUs are supported.^[6]

Proprietary graphics device driver AMDGPU-PRO

This new driver by AMD was still undergoing development in 2018, but could be used on a few supported Linux distributions already (AMD officially supports Ubuntu, RHEL/CentOS).^[48] The driver has been experimentally ported to ArchLinux^[49] and other distributions. AMDGPU-PRO is set to replace the previous AMD Catalyst driver and is based on the free and open source amdgpu kernel driver. Pre-GCN GPUs are not supported.

See also

• Graphics Core Next
• AMD FirePro
• AMD FireMV
• AMD FireStream
• List of AMD graphics processing units