Radeon Pro

AMD Radeon Pro
Radeon Pro Logo
Design firmAdvanced Micro Devices
TypeProfessional workstations

Radeon Pro is AMD's brand of professional oriented GPUs. It replaced AMD's FirePro brand in 2016. Compared to the Radeon brand for mainstream consumer/gamer products, the Radeon Pro brand is intended for use in workstations and the running of computer-aided design (CAD), computer-generated imagery (CGI), digital content creation (DCC), high-performance computing/GPGPU applications, and the creation and running of virtual reality programs and games.

The Radeon Pro product line directly competes with Nvidia, i.e. their Quadro (since discontinued) line of professional workstation cards.

Products

Radeon PRO series

Radeon Pro Duo (2016)

The first card to be released under the Radeon Pro name was the dual GPU Radeon Pro Duo in April 2016. The card features 2 liquid cooled R9 Nano cores & was marketed strongly for both the running and creation of virtual reality content with the slogan "For Gamers Who Create and Creators Who Game". The aesthetics and marketing of the Pro Duo follow that of the rest of the Fury products in the 300 series.

Radeon Pro Duo (2017)

In April 2017 AMD announced a new version of the Radeon Pro Duo for release the following month. The newer version of the Pro Duo utilizes dual GPUs from the Polaris architecture, using the same GPUs as in the WX7100. While this results in a smaller number of compute units and lower theoretical performance, it allows for the inclusion of 32 GB GDDR5 SDRAM and a lower board power.

Radeon Pro SSG (Fiji)

Using AMD Radeon's GCN 3 architecture, the Radeon Pro SSG was unveiled in July 2016. SSG stands for Solid State Graphics, and the card will couple AMD's Fiji core with solid-state storage to increase the frame buffer for rendering. This expansion of quick access storage will, therefore, relieve the issue of latency that occurs when a GPU has to retrieve information from a mass storage device via the CPU when a card's limited VRAM is maxed out in heavy workloads. Users will be able to add up to 1 TB of PCIe M.2 NAND flash memory to improve render and scrubbing times. AMD demonstrated a 5.3 fold increase in performance on 8K video scrubbing. This SSD storage space can be made available to the operating system or controlled entirely by the GPU. The Fiji-based Radeon Pro SSG card was available as a beta program.

Radeon Pro SSG (Vega)

In July 2017, AMD released the Vega-based Radeon Pro SSG. The card utilizes 16 GB of second generation ECC high bandwidth memory (HBM2), an upgrade from the Fiji-based card's 4 GB of first generation HBM memory. The Vega card also increased the built in solid-state storage to 2 TB.

Radeon Vega Frontier Edition

AMD announced in May 2017 the Radeon Vega Frontier Edition, for release in June of that year. While not branded as a Pro product, the card is marketed within the Radeon Pro series. The Radeon Vega Frontier Edition uses the new "Next-Gen Compute Unit" and 16 GB of HBM2 memory for an expected 13.1 TFLOPs of single precision and 26.2 TFLOPs of half precision performance. Ultimately, two Frontier Edition products were released with either air or liquid cooling. The liquid cooling part supported a higher TDP, and was able to reach and sustain higher clock speeds, but otherwise the two products have similar hardware specifications.

Radeon Pro V series

The Pro V series was announced in August 2018 with the Vega-based Radeon Pro V340, a dual-GPU flagship card for use in datacenter virtualization, supporting up to 32 virtual machines at a time, as well as several other potential uses for Computer-aided design, general rendering tasks, and Desktop as a Service. It was expected to be available in Q4 of that year.

Radeon Pro WX series

AMD Radeon Pro WX 3100

Radeon Pro WX series are graphics cards designed specifically for professional applications used in engineering, design, content creation, and science. The first Radeon Pro cards with the WX prefix to be announced were the WX 7100, the WX 5100 and the WX 4100 in July 2016. These Polaris based cards are once again aimed at the traditional professional market and are set to replace the FirePro Wx100 series and FirePro Wx300 series. These cards, along with the Pro SSG, used the new, non-toxic and energy efficient YInMn Blue, discovered by Mas Subramanian. This unique aesthetic for the Radeon Pro line distinguished the professional products from the consumer Radeon series.

The smallest card, the half-height WX 4100, was marketed for use in small form factor workstations. Designed for real-time content engines and CAD and CAM manufacturing, the WX 5100 fits in between the WX 4100 and the WX 7100 in terms of performance, with the latter once again marketed with emphasis on the application of VR and other media creation, while claiming to be "The Most Affordable Workstation Solution".

In June 2017, AMD announced the addition of the lower power WX 2100 and WX 3100 cards to the Radeon Pro WX series. Both cards are based on the Polaris GPU and are rated at 1.25 TFLOPS. The WX 2100 has 2 GB of GDDR5 SDRAM, while the WX 3100 has 4 GB of GDDR5 memory.

In September 2017, AMD launched the WX 9100 based on the Vega architecture. The card features 16 GB of ECC HBM2 memory and is rated at 12.29 TFLOPS. As the new flagship of the WX line, it greatly exceeds the performance of the older WX 7100 which is rated at 5.73 TFLOPS. The WX 9100 has ISV (Independent Software Vendor) certified drivers for professional applications including Siemens NX, PTC Creo, Dassault Systèmes CATIA and 3DExperience Platform, Dassault Systèmes SOLIDWORKS, and Autodesk® Revit®. The WX 9100 is particularly well-suited for mission critical workloads and complex scientific modeling because the ECC memory helps correct "single or double bit error as a result of naturally occurring background radiation."

Radeon Pro 400 series

Mobile Radeon Pro parts were first revealed with the release of the 2016 update to the Apple 15" MacBook Pro. These appear to be Polaris 11 derived parts with 10-16 4th generation GCN compute units, providing between 1 and 1.86 TFLOPS of performance.

Radeon Pro 500 series

Released in conjunction with the 2017 Apple iMac refresh, the Radeon Pro 500 series serve as GPUs for the 4K and 5K Retina display iMacs. The 500 series ranges supports 2 to 8 GB of graphics RAM with performance from 1.3 to 5.5 TFLOPS.

Radeon Pro Vega series

The Radeon Pro Vega product line of GPUs were first announced in 2017 as a part of Apple's iMac Pro. The two models, Radeon Pro Vega 56 and 64, support 8 and 16 GB of HBM2 memory, respectively. On October 30, 2018, Apple added graphics upgrade options for their 15-inch MacBook Pro lineup consisting Radeon Pro Vega 16 and 20. Derived from Vega 12 GPU that was only used on Apple laptops, both GPU features a 4 GB HBM2 memory stack and performance up to 3.3 TFLOPS.

The second-generation, 7 nm Radeon Pro Vega II was announced in 2019 as part of Apple's third-generation Mac Pro desktop computer. The Pro Vega II supports 32 GB of HBM2 memory, while the Pro Vega II Duo combines two Vega GPUs and supports 64 GB of HBM2 memory. The Mac Pro supports up to two Pro Vega II or Pro Vega II Duo graphics cards, allowing up to four Vega GPUs and 128 GB of HBM2 memory to be used in a system.

Radeon Pro VII series

The Radeon Pro VII was announced in May 2020, as a professional variant of the Radeon VII.

Radeon Pro 5000/5000M series

Released in conjunction with the 2019 Apple 16-inch MacBook Pro. Two models were announced, the 5300M and the 5500M. Both feature GDDR6 memory interfaces, with 192 GB/s bandwidth. The 5500M supports up to 8 GB of GDDR6 and 4.0 TFLOPS. In June 2020, a new 5600M GPU model with 8 GB of HBM2 memory was quietly released.

Radeon Pro W5000/W5000M series

AMD Radeon Pro W5500

The Radeon Pro W5700, which is based on RDNA Architecture for desktop workstations, was officially released on November 19, 2019. The smaller model Radeon Pro W5500 was released in February 2020.

Radeon Pro W5000X series

The RDNA Radeon Pro W5000X series cards were designed for the 2019 Mac Pro as MPX cards, which interface through PCIe and a second proprietary Apple connector. These cards are the W5700X and W5500X.

Radeon Pro W6000/W6000M series

The first RDNA2 based W6000 series cards were officially announced on June 8, 2021 and launched in Q3 2021, with the AMD Radeon Pro W6800, W6600 and W6600M for mobile. On January 19, 2022, the W6400 was also released.

Radeon Pro W6000X series

The RDNA2 Radeon Pro W6000X series cards were designed for the 2019 Mac Pro as MPX cards. These cards are the W6900X, W6800X, and W6600X.

Radeon Pro W7000 series

The first RDNA 3-based W7000 series cards were officially announced on April 13, 2023, with the AMD Radeon Pro W7900 and W7800. On August 3 of the same year, the W7600 and W7500 were announced. On November 13, the W7700 was announced.

Software

Most professional compute is done with the help of the Radeon Open Compute and the GPUOpen platforms.

Project Loom

At an AMD event in 2016 Project Loom was announced as a collaboration between AMD and Radiant Images. The real-time GPU accelerated photo and video stitching program will complement AMD's virtual reality development platform. While traditional photo stitching is not that much of a complex task, Project Loom aims to improve render times when tasked with the heavy workload of stitching together multiple high resolution angles to form a 360 degree VR experience, either to headsets or mobile devices. Using AMD's Direct GMA protocol, the software allows Radeon Pro graphics cards to work directly with video capture hardware to stitch together a 30 fps, 360 degree 4k resolution video from 24, 1080p cameras at 60 fps.

The software is to be competitive with Nvidia's VRWorks 360 Video SDK, and is reportedly set to be made open-source through GPUOpen.

ProRender

The successor to FireRender, Radeon ProRender works with high-end graphics programs as an OpenCL photorealistic offline 3D renderer and raytracing engine. ProRender aims to compete with programs such as NVIDIA's Iray and other expensive, proprietary solutions. However, AMD is making ProRender free and available for all graphics hardware. ProRender was released by AMD in June 2016 with support for Blender, 3D Studio Max, SolidWorks, and Maya.

Driver

API OpenGL 4.5 is supported and 4.6 is in development. API Vulkan 1.0 is supported for all with GCN Architecture. Vulkan 1.1 (GCN 2nd Gen. and higher) will be supported with actual drivers in 2018.

As with other GPU architectures, the floating-point performance is dependent on the precision and the GCN generation:

  • In 4th Gen GCN, FP64 is 1/16 of FP32. Newer gaming cards have better ratios, which should be reflected on newer derivative "Pro" versions:
    • The gaming card Radeon R9 295X2 has it bumped up to 1/8 FP32.
    • The gaming card Radeon VII has it bumped up to 1/4 FP32.
    • The Radeon Pro Vega 20 has the ratio bumped up to 1/2 FP32.
  • In 5th Gen GCN, FP16 is double of FP32. In 1st Gen to 4th it was equal to FP32.

For those requiring higher FP64 performance, a form of FP64 distinct from the IEEE double-precision can be emulated with the much faster FP32 operations. The cost is around a ~1/3 performance compared to FP32, much better than what the native support could provide.

Chipset table

Workstation

Radeon PRO series

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TBP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro Duo
(Fiji)
Apr 26, 2016
$1,499 USD
GCN 3
TSMC 28 nm
2× 8.9×109
596 mm2
2× 4096:256:64
64 CU
1000 2× 256.0 2× 64.00 2× 8,192 2× 8,192 2× 512.0 2× 4 GB 2× 512 HBM
2× 4096-bit
1000 350 W PCIe 3.0
×16
3× DP 1.2
1× HDMI 1.4a
Radeon Pro Duo
(Polaris 10)
Apr 24, 2017
$999 USD
GCN 4
GloFo 14 nm
2× 5.7×109
232 mm2
2× 2304:128:32
36 CU
1243 2× 179.0 2× 39.78 2× 5,728 2× 5,728 2× 358.0 2× 16 GB 2× 224 GDDR5
2× 256-bit
7000 250 W 3× DP 1.4a
1× HDMI 2.0b
Radeon Pro SSG
(Fiji)
Jul 26, 2016
prototype only
$9,999 USD
GCN 3
TSMC 28 nm
8.9×109
596 mm2
4096:256:64
64 CU
1050 268.8 67.20 8,601 8,601 537.6 4 GB
+1 TB SSD
512 HBM + SSG
4096-bit
1000 200 W 4× DP 1.2
Radeon Pro SSG
(Vega 10)
Aug 8, 2017
$6,999 USD
GCN 5
GloFo 14 nm
12.5×109
495 mm2
4096:256:64
64 CU
1440
1500
368.6
384.0
92.16
96.00
23,593
24,576
11,796
12,288
737.2
768.0
16 GB
+2 TB SSD
484 HBM2 + SSG
2048-bit
1890 230 W 6× miniDP 1.4a
Radeon Vega Frontier Edition
(Vega 10)
Jun 27, 2017
$999 USD
1382
1600
353.8
409.6
88.4
102.4
22,643
26,214
11,321
13,107
707.6
819.2
16 GB HBM2
2048-bit
300 W 3× DP 1.4a
1× HDMI 2.0b
Radeon Vega Frontier Edition
(Liquid-cooled)
(Vega 10)
Jun 27, 2017
$1,499 USD
375 W
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro WX x100 series

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TBP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro WX 2100
(Polaris 12)
Jun 1, 2017
$149 USD
GCN 4
GloFo 14 nm
2.2×109
103 mm2
512:32:16
8 CU
925
1219
29.6
39.0
14.8
19.5
947
1,250
947
1,250
59.2
78
2 48 GDDR5
64-bit
6000 35 W PCIe 3.0
×8
1× DP 1.4a
2× miniDP 1.4a
Radeon Pro WX 3100
(Polaris 12)
Jun 1, 2017
$199 USD
4 96 GDDR5
128-bit
50 W
Radeon Pro WX 4100
(Polaris 11)
Nov 10, 2016
$399 USD
3.0×109
123 mm2
1024:64:16
16 CU
1125
1201
72
76.9
18
19.2
2,304
2,460
2,304
2,460
144
154
96 7000 4× miniDP 1.4a
Radeon Pro WX 5100
(Polaris 10)
Nov 18, 2016
$499 USD
5.7×109
232 mm2
1792:112:32
28 CU
713
1086
79.85
121.6
22.8
34.75
2,555
3,892
2,555
3,892
159.7
243.3
8 160 GDDR5
256-bit
5000 75 W PCIe 3.0
×16
4× DP 1.4a
Radeon Pro WX 7100
(Polaris 10)
Nov 10, 2016
$799 USD
2304:144:32
36 CU
1188
1243
171
179
38
39.78
4,150
5,728
5,474
5,728
342.1
358
224 7000 130 W
Radeon Pro WX 9100
(Vega 10)
Sep 13, 2017
$2,199 USD
GCN 5
GloFo 14 nm
12.5×109
495mm2
4096:256:64
64 CU
1200
1500
307.2
384.0
76.8
96.0
19,660
24,576
9,830
12,288
614.4
768
16 484 HBM2
2048-bit
1890 230 W 6× miniDP 1.4a
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro WX x200 series

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TBP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro WX 3200
(Polaris 23)
Jul 2, 2019
$199 USD
GCN 4
GloFo 14 nm
2.2×109
103mm2
640:32:16
10 CU
1295 41.44 20.72 1,658 1,658 103.6 4 96 GDDR5
128-bit
6000 50 W PCIe 3.0
×8
4× mini-DP 1.4a
Radeon Pro WX 8200
(Vega 10)
Aug 13, 2018
$999 USD
GCN 5
GloFo 14 nm
12.5×109
495mm2
3584:224:64
56 CU
1200
1500
268.8
336.0
76.8
96.00
17,203
21,504
8,601
10,752
537.6
672.0
8 512 HBM2
2048-bit
2000 230 W PCIe 3.0
×16
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro Vega (for Apple Mac Pro)

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TBP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro Vega II
(Vega 20)
2019
$2,800 USD
GCN 5
TSMC 7 nm
13.23×109
331 mm2
4096:256:64
64 CU
1720 440.3 110.1 28,180 14,090 880 32 1024 HBM2
4096-bit
2000 475 W PCIe 3.0
×16
4× Thunderbolt 3
(USB Type-C)
1× HDMI 2.0b
Radeon Pro Vega II Duo
(Vega 20)
2019
$5,600 USD
2× 4096:256:64
64 CU
1720 2× 440.3 2× 110.1 2× 28,180 2× 14,090 2× 880 2× 32 2× 1024 HBM2
2× 4096-bit
2000
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro VII

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TBP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro VII
(Vega 20)
May 13, 2020
$1,899 USD
GCN 5
TSMC 7 nm
13.23×109
331 mm2
3840:240:64
60 CU
1400
1700
336.0
408.0
89.6
108.8
21,504
26,112
10,752
13,056
5,376
6,528
16 1024 HBM2
4096-bit
2000 250 W PCIe 4.0
×16
6× miniDP 1.4a
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro 5000 series (for Apple iMac)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro 5300
(Navi 14)
Aug 4, 2020 RDNA
TSMC N7
6.4×109
158 mm2
1280:80:32
20 CU
1000
1650
80
132
32
52.8
5,120
8,448
2,560
4,224
160
264
4 224 GDDR6
128-bit
14000 130 W PCIe 4.0
×8
Radeon Pro 5500 XT
(Navi 14)
1536:96:32
24 CU
1187
1757
114
168.7
38
56.2
7,292
10,796
3,646
5,398
227.9
337.3
8
Radeon Pro 5700
(Navi 10)
10.3×109
251 mm2
2304:144:64
36 CU
1243
1350
179
194.4
79.6
86.4
11,456
12,442
5,728
6,221
358
388.8
384 GDDR6
256-bit
12000 PCIe 4.0
×16
Radeon Pro 5700 XT
(Navi 10)
2560:160:64
40 CU
1243
1499
198.9
239.8
79.6
95.94
12,728
15,350
6,364
7,675
397.8
479.7
16
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro W5000 series

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W5500
(Navi 14)
Feb 10, 2020
$399 USD
RDNA
TSMC N7
6.4×109
158 mm2
1408:88:32
22 CU
1744
1855
153.4
163.2
55.8
59.36
9,822
10,450
4,911
5,224
306.9
326.5
8 224 GDDR6
128-bit
14000 125 W PCIe 4.0
×8
4× DP 1.4a
Radeon Pro W5700
(Navi 10)
Nov 19, 2019
$799 USD
10.3×109
251 mm2
2304:144:64
36 CU
1400
1880
201.6
270.7
89.6
120.3
12,902
17,330
6,451
8,663
403.2
541.4
448 GDDR6
256-bit
205 W PCIe 4.0
×16
5× miniDP 1.4a
1× USB Type-C
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro W5000X series (for Apple Mac Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W5500X
(Navi 14)
2020 RDNA
TSMC N7
6.4×109
158 mm2
1536:96:32
24 CU

1757

163.2

59.36

11,200

5,600

326.5
8 224 GDDR6
128-bit
14000 130 W PCIe 4.0
×8
2× HDMI 2.0b
Radeon Pro W5700X
(Navi 10)
Dec 11, 2019 10.3×109
251 mm2
2560:160:64
40 CU
1243
1860
198.8
297.6
79.5
119.04
12,728
19,046
6,364
9,523
397.8
595.2
16 448 GDDR6
256-bit
250 W PCIe 4.0
×16
4× Thunderbolt 3
1× HDMI 2.0b
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro W6000 series

Model
(Code name)
Release date
& price
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Infinity
Cache
Memory TDP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W6300
(Navi 24)
Oct 2022
OEM
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)
Jan 19, 2022
$229 USD
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
50 W
Radeon Pro W6600
(Navi 23)
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)
Jun 8, 2021
$2249 USD
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
  1. 1 2 3 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)

Radeon Pro W6000X series (for Apple Mac Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Infinity
Cache
Memory TDP Bus
interface
Graphic output
ports
Config Clock
(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)
Mar 8, 2022 RDNA 2
TSMC N7
11.06×109
237 mm2
2048:124:64:32
32 CU
2068
2479

307.3

158.6

19,673

9,837

614.8
32 MB 8 256 GDDR6
128-bit
16000 120 W PCIe 4.0
×8
2× HDMI 2.1
Radeon Pro W6800X
(Navi 21)
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
300 W PCIe 4.0
×16
4× Thunderbolt 3
1× HDMI 2.1
Radeon Pro W6800X Duo
(Navi 21)
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)
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 GDDR6
256-bit
300 W
  1. 1 2 3 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)

Radeon Pro W7000 series

Model
(Code name)
Release date
& price
Architecture
& fab
Chiplets
(active)
Transistors
& die size
Core Fillrate Processing power
(TFLOPS)
Infinity
Cache
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W7500
(Navi 33)
Aug 3, 2023
$429 USD
RDNA 3
TSMC N6
N/a 13.3×109
204 mm2
1792:112:64
28:56:28 CU
1500
1700
168.0
190.4
96.0
108.8
24.37 12.19 0.381 32 MB 8 288 GDDR6
128-bit
18000 70 W PCIe 4.0
×8
Radeon Pro W7600
(Navi 33)
Aug 3, 2023
$599 USD
2048:128:64
32:64:32 CU
1720
2440
220.1
312.3
110.0
156.2
39.98 19.99 0.625 130 W
Radeon Pro W7700
(Navi 32)
Nov 13, 2023
$999 USD
RDNA 3
TSMC N5
(GCD)
TSMC N6
(MCD)
1 × GCD
4 × MCD
28.1×109
~346 mm2
3072:192:96
48:96:48 CU
1900
2600
364.8
499.2
182.4
249.2
56.54 28.3 0.884 64 MB 16 576 GDDR6
256-bit
190 W PCIe 4.0
×16
Radeon Pro W7800
(Navi 31)
Apr 13, 2023
$2499 USD
57.7×109
~531 mm2
4480:280:128
70:128:70 CU
1855
2499
519.4
699.7
237.4
319.8
90.50 45.25 1.414 32 260 W
Radeon Pro W7800 48GB
(Navi 31)
Dec 2024 1 × GCD
6 × MCD
96 MB 48 864 GDDR6
384-bit
281 W
Radeon Pro W7900 Dual Slot
(Navi 31)
Jun 19, 2024
$3499 USD
6144:384:192
96:192:96 CU
1855
2495
712.3
958.0
356.1
479.0
122.6 61.32 1.916 295 W
Radeon Pro W7900
(Navi 31)
Apr 13, 2023
$3999 USD
  1. ↑ Approximate die size of entire MCM package that consists of single GCD (Graphics Compute Die) and six MCDs (Memory Cache Die).
    Radeon Pro W7800 has only four active MCDs, inactive one is for structural support and heat dissipation.
  2. 1 2 3 Boost values (if available) are stated below the base value in italic.
  3. ↑ Texture fillrate is calculated as the number of Texture Mapping Units multiplied by the base (or boost) core clock speed.
  4. ↑ Pixel fillrate is calculated as the number of Render Output Units multiplied by the base (or boost) core clock speed.
  5. ↑ Precision performance is calculated from the base (or boost) core clock speed based on a FMA operation.
  6. ↑ Unified shaders : Texture mapping units : Render output units : Ray accelerators : AI accelerators and Compute units (CU)
  7. ↑ 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.

Radeon AI Pro R9000 series

Model
(Code name)
Release date Architecture
& fab
Transistors
& die size
Core Fillrate Processing power Infinity
Cache
Memory TBP Bus
interface
(TFLOPS) AI TOPS
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double FP8 INT8 INT4 Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon AI Pro R9600D
(Navi 48)
Dec 11, 2025 RDNA 4
TSMC N4P
53.9 billion
357 mm2
3072:192:96
48:96:48 CU
1080
2020

387.8

193.92

49.6

24.8

0.388

199

199

397
64 MB 32 640 GDDR6
256-bit
20000 150 W PCIe 5.0
×16
Radeon AI Pro R9600
(Navi 48)
2026
Radeon AI Pro R9700S
(Navi 48)
Dec 11, 2025 4096:256:128
64:128:64 CU
1660
2920
425
747.5
212.5
373.76
54.4
95.68
27.2
47.84
0.425
0.747
217.7
383
217.7
383
435.5
766
300 W
Radeon AI Pro R9700
(Navi 48)
Jul 23, 2025
  1. 1 2 3 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)

Mobile workstation

Radeon Pro WX x100 Mobile series

  • Half Precision Power (FP16) is equal single precision power (FP32) in 4th GCN Generation (in 5th Gen: Half Precision (FP16) = 2x SP (FP32))
Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro WX 2100 (Mobile)
(Polaris 12)
Mar 1, 2017 GCN 4
Samsung/GloFo
14LPP
2.2×109
103 mm2
512:32:16
8 CU
925
1219
29.60
39.01
14.80
19.50
947.2
1,250
59.20
78.02
2 48 GDDR5
64-bit
6000 35 W PCIe 3.0
×16
Radeon Pro WX 3100 (Mobile)
(Polaris 12)
4 96 GDDR5
128-bit
50 W
Radeon Pro WX 4130 (Mobile)
(Polaris 11)
3.0×109
123 mm2
640:40:16
10 CU
1002
1053
40.08
42.12
16.03
16.85
1,282
1,348
80.16
84.24
50 W
Radeon Pro WX 4150 (Mobile)
(Polaris 11)
896:56:16
14 CU
1002
1053
56.11
58.97
16.03
16.85
1,796
1,887
112.2
117.9
45-50 W
Radeon Pro WX 4170 (Mobile)
(Polaris 11)
1024:64:16
16 CU
1002
1201
64.12
76.86
16.03
19.22
2,052
2,460
128.3
153.7
112 50-60 W
Radeon Pro WX 7100 (Mobile)
(Polaris 10)
5.7×109
232 mm2
2304:144:32
36 CU
1188
1243
171.1
179.0
38.0
39.78
5,474
5,728
342.1
358.0
8 160 GDDR5
256-bit
5000 100-130 W
Radeon Pro WX 7130 (Mobile)
(Polaris 10)
  1. 1 2 3 Boost values (if available) are stated below the base value in italic.
  2. ↑ Texture fillrate is calculated as the number of Texture Mapping Units (TMUs) multiplied by the base (or boost) core clock speed.
  3. ↑ Pixel fillrate is calculated as the number of Render Output Units (ROPs) 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 and Compute units (CU)

Radeon Pro 400 series (for Apple MacBook Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro 450
(Polaris 11)
Oct 30, 2016 GCN 4
GloFo 14 nm
3.0×109
123 mm2
640:40:16
10 CU
800 32.00 12.80 1,024 1,024 64.00 2 80 GDDR5
128-bit
5000 35 W PCIe 3.0
×8
Radeon Pro 455
(Polaris 11)
768:48:16
12 CU
855 41.04 13.68 1,313 1,313 82.08
Radeon Pro 460
(Polaris 11)
1024:64:16
16 CU
850
907
54.40
58.05
13.60
14.51
1,741
1,858
1,741
1,858
108.8
116.1
4
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro 500 series (for Apple iMac & MacBook Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro 555
(Polaris 21)
Jun 6, 2017 GCN 4
GloFo 14 nm
3.0×109
123 mm2
768:48:16
12 CU
850 40.80 13.60 1,306 1,306 81.60 2 81.60 GDDR5
128-bit
5100 50 W PCIe 3.0
×16
Radeon Pro 555X
(Polaris 21)
Jul 16, 2018 907 43.54 14.51 1,393 1,393 87.07 4 94.08 5900 50 W
Radeon Pro 560
(Polaris 21)
Jun 6, 2017 1024:64:16
16 CU
907 58.05 14.51 1,858 1,858 116.1 81.28 5100 50 W
Radeon Pro 560X
(Polaris 21)
Jul 16, 2018 1004 64.26 16.06 2,056 2,056 128.5 94.08 5900 75 W
Radeon Pro 570
(Polaris 20)
Jun 6, 2017 5.7×109
232 mm2
1792:112:32
28 CU
1000
1105
112.0
123.8
32.00
35.36
3,584
3,960
3,584
3,960
224.0
247.5
217.0 GDDR5
256-bit
6800 120 W
Radeon Pro 570X
(Polaris 20)
Mar 18, 2019 217.6 150 W
Radeon Pro 575
(Polaris 20)
Jun 6, 2017 2048:128:32
32 CU
1096 140.3 35.07 4,489 4,489 280.6 217.0 120 W
Radeon Pro 575X
(Polaris 20)
Mar 18, 2019 217.6 150 W
Radeon Pro 580
(Polaris 10)
Jun 6, 2017 2304:144:32
36 CU
1100
1200
158.4
172.8
35.2
38.4
5,069
5,530
5,069
5,530
316.8
345.6
8 217.0 150 W
Radeon Pro 580X
(Polaris 10)
Mar 18, 2019 218.9 185 W
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro WX x200 Mobile series

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro WX 3200 (Mobile)
(Polaris 23)
Mar 1, 2017 GCN 4
GloFo 14 nm
2.2×109
103 mm2
640:32:16
10 CU
1082
1295
34.62
41.44
17.31
20.72
1,385
1,658
1,385
1,658
86.56
103.6
4 64 GDDR5
128-bit
4000 65 W PCIe 3.0
×8
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro Vega series (for Apple iMac & MacBook Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro Vega 16
(Vega 12)
Nov 14, 2018 GCN 5
GloFo 14 nm
? 1024:64:32
16 CU
815
1190
52.16
76.16
26.08
38.08
3,338
4,874
1,669
2,437
104.3
152.3
4 307.2 HBM2
1024-bit
2400 50 W PCIe 3.0
×16
Radeon Pro Vega 20
(Vega 12)
1280:80:32
20 CU
815
1283
65.20
102.6
26.08
41.06
4,173
6,569
2,086
3,285
130.4
205.3
189.4 1480 50 W
Radeon Pro Vega 48
(Vega 10)
Mar 19, 2019 12.5×109
495 mm2
3072:192:64
48 CU

1200

230.4

76.80

14,746

7,373

460.8
8 402.4 HBM2
2048-bit
1572 ?
Radeon Pro Vega 56
(Vega 10)
Aug 17, 2017 3584:224:64
56 CU
1138
1250
254.9
280.0
72.83
80.00
16,314
17,920
8,157
8,960
509.8
560.0
120 W
Radeon Pro Vega 64
(Vega 10)
Jun 17, 2017 4096:256:64
64 CU
1250
1350
320.0
345.6
80.00
86.40
20,480
22,118
10,240
11,059
640.0
691.2
16 ?
Radeon Pro Vega 64X
(Vega 10)
Mar 19, 2019 4096:256:64
64 CU
1250
1468
320.0
375.8
80.00
93.95
20,480
24,051
10,240
12,026
640.0
751.6
512.0 2000
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro 5000M series (for Apple MacBook Pro)

Model
(Code name)
Release
date
Architecture
& fab
Transistors
 die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro 5300M
(Navi 14)
Nov 13, 2019 RDNA
TSMC N7
6.4×109
158 mm2
1280:80:32
20 CU
1000
1250
80.00
100.0
32.00
40.00
5,120
6,400
2,560
3,200
160.0
200.0
4 192 GDDR6
128-bit
12000 50 W PCIe 4.0
×8
Radeon Pro 5500M
(Navi 14)
1536:96:32
24 CU
1000
1300
96.00
124.8
32.00
41.60
6,144
8,908
3,072
4,454
192.0
278.4
4
8
Radeon Pro 5600M
(Navi 12)
Jun 15, 2020 Unknown 2560:160:64
40 CU
1000
1035
160.0
165.6
64.00
66.24
10,240
10,598
5,120
5,299
320.0
331.2
8 394 HBM2
2048-bit
1540 PCIe 4.0
×16
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro W5000M series

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W5500M
(Navi 14)
Feb 10, 2020 RDNA
TSMC N7
6.4×109
158 mm2
1408:88:32
22 CU
1000
1700
88.00
149.6
32.00
54.40
5,632
9,574
2,816
4,787
176.0
299.2
4 224 GDDR6
128-bit
14000 65-85 W PCIe 4.0
×8
  1. 1 2 3 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 and Compute units (CU)

Radeon Pro W6000M series

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
Infinity
Cache
Memory TDP Bus
interface
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro W6300M
(Navi 24)
Jan 19, 2022 RDNA 2
TSMC N6
5.4×109
107 mm2
768:48:32:12
12 CU

2214

106.3

70.8

6,801

3,401

212.5
8 MB 2 64 GDDR6
32-bit
14000 25 W PCIe 4.0
×4
Radeon Pro W6500M
(Navi 24)
1024:64:32:16
16 CU

2588

165.6

82.8

10,478

5,239

327.4
16 MB 4 128 GDDR6
64-bit
35–50 W
Radeon Pro W6600M
(Navi 23)
Jun 8, 2021 RDNA 2
TSMC N7
11.06×109
237 mm2
1792:112:64:28
28 CU
2200
2900
246.4
324.8
140.8
185.6
15,770
20,787
7,885
10,394
492.8
649.6
32 MB 8 224 GDDR6
128-bit
65–90 W PCIe 4.0
×16
  1. 1 2 3 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)

Data center GPUs

Radeon Pro V series

Model
(Code name)
Release
date
Architecture
& fab
Transistors
& die size
Core Fillrate Processing power
(GFLOPS)
L3
cache
Memory TDP Bus
interface
Graphic output
ports
Config Clock
(MHz)
Texture
(GT/s)
Pixel
(GP/s)
Half Single Double Size
(GB)
Bandwidth
(GB/s)
Bus type
& width
Clock
(MT/s)
Radeon Pro V320
(Vega 10)
Jun 29, 2017
(Custom SKU)
GCN 5
GF 14LPP
12.5×109
495 mm2
4096:256:64:-
64 CU
852
1000
218.1
256
54.52
64
13,959
16,384
6,979
8,192
436.2
512
— 8 or 16 484 HBM2
2048-bit
1890 230 W PCIe 3.0
×16
4× miniDP 1.4a
Radeon Pro V340
(Vega 10)
Aug 26, 2018
2× 12.5×109
495 mm2
2× 3584:224:64:-
56 CU
852
1500
2× 190.8
336
2× 54.52
96
2× 12,214
21,504
2× 6,107
10,752
2× 381.7
672
— 2× 16 2× 484 HBM2
2× 2048-bit
300 W 1× miniDP 1.4a
Radeon Pro V520
(Navi 12)
Dec 1, 2020 RDNA
TSMC N7
Unknown 2304:144:64:-
36 CU
1000
1600
144
230.4
64
102.4
9,216
14,746
4,608
7,373
288
460.8
— 8 512 HBM2
2048-bit
2000 225 W PCIe 4.0
×16
N/a
Radeon Pro V620
(Navi 21)
Nov 4, 2021 RDNA 2
TSMC N7
26.8×109
520 mm2
4608:288:128:72
72 CU
1825
2200
525.6
633.6
233.6
281.6
33,638
40,550
16,819
20,275
1,051
1,267
128 MB 32 512 GDDR6
256-bit
16000 300 W PCIe 4.0
×16
N/a
Radeon Pro V710
(Navi 32)
Oct 3, 2024 RDNA 3
TSMC N5 (GCD)
TSMC N6 (MCD)
28.1×109
384 mm2
3456:216:96:54
54 CU
1900
2000

432

192

55,296

28,314

1,728
54 MB 28 448 GDDR6
224-bit
16000 158 W PCIe 4.0
×16
N/a
  1. 1 2 3 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)

See also