AMD Radeon 680M

AMD Radeon 680M
AMD Radeon 680M graphics card review

AMD Radeon 680M: Specifications, Game Tests, and Comparison with Radeon 660M and 780M

The AMD Radeon 680M has become one of the major leaps in the development of AMD's integrated graphics. Twelve RDNA 2 compute units enable it to run competitive games and many older AAA titles at Full HD without a discrete graphics card. However, performance varies by device: in tested laptops and mini-PCs, the difference in Time Spy Graphics scores can reach up to 29%.

Architecture and Position in AMD's Lineup

The Radeon 680M features 12 RDNA 2 compute units, or 768 shading units. The maximum frequency depends on the processor: in the Ryzen 7 6800U and Ryzen 7 7735HS, the graphics operate at frequencies of up to 2.2 GHz, while in the Ryzen 9 6900HS and 6900HX, it can go up to 2.4 GHz.

This iGPU is found in the Ryzen 7 6800U, Ryzen 7 6800H, Ryzen 7 PRO 6850U, Ryzen 9 6900HS, Ryzen 9 6900HX, Ryzen 7 7735U, and Ryzen 7 7735HS.

The Radeon 680M does not have its own video memory. It uses system DDR5 or LPDDR5 alongside the processor cores, making it sensitive to both memory bandwidth and available power. Beyond a certain level, further increases in the APU’s limits yield diminishing returns, as the limitation comes from the 12 compute units themselves.

The entry-level Radeon 660M is based on the same architecture but has only six compute units and a frequency of up to 1.9 GHz. This is not a slightly throttled version of the 680M, but rather a lower-class graphics card.

Radeon 680M in Specific Devices

Device Processor APU and Memory Limit 3DMark Time Spy Graphics 3DMark Fire Strike Graphics
Beelink SER6 Pro Ryzen 7 6800H around 45 W, DDR5-4800 2452 7026
Lenovo Yoga 7 14ARB7 Ryzen 7 6800U 25 W, 16 GB LPDDR5-6400 2449 7279
Lenovo ThinkPad T14 Gen 3 AMD Ryzen 7 PRO 6850U -, 16 GB LPDDR5-6400 2289 7050
ASUS Zenbook S 13 OLED UM5302TA Ryzen 7 6800U -, 16 GB LPDDR5-6400 2287 6683
HP EliteBook 835 G9 Ryzen 7 PRO 6850U -, 16 GB DDR5-4800 1896 5654

The Yoga 7 with a 25-Watt Ryzen 7 6800U essentially matched the mini-PC with a Ryzen 7 6800H: 2449 versus 2452 points in Time Spy Graphics. This illustrates that the mobile U-series, with favorable power settings and fast memory, can perform on par with more powerful processors using the same Radeon 680M.

The ThinkPad T14 and Zenbook S 13 score around 2290 points-approximately 7% lower than the leaders. Scores in the range of 2300 to 2450 points can be considered typical for fast devices with the Radeon 680M.

The HP EliteBook 835 G9 is limited to 1896 points, trailing the Beelink SER6 Pro by almost 29%. This could be due to the DDR5-4800, power limits, and manufacturer settings, but it’s impossible to determine the contribution of each factor based on a single comparison.

Power increases also affect results. In the same device, lowering the APU limit from about 50 to 25 W reduced the Time Spy Graphics score by approximately 19%. Moreover, further increases in power consumption yield less noticeable gains than moving to a graphics card with a higher number of compute units.

What Radeon 680M Can Handle in Games

The Radeon 680M is most suitable for screens with a resolution of 1920 × 1080 or 1920 × 1200. In lighter games, high settings can be used, while heavier projects require switching to low presets and using FSR.

Game and Settings Beelink SER6 Pro Lenovo Yoga 7
Cyberpunk 2077, 1080p, low about 43 FPS about 34 FPS
Cyberpunk 2077, 1080p, medium about 33 FPS about 26 FPS
The Witcher 3, 1080p, high - about 43 FPS
Dota 2, 1080p, high - about 97 FPS

In Cyberpunk 2077, the Beelink mini-PC maintains over 40 frames per second on low settings and around 33 frames on medium. The Yoga 7 is noticeably slower, but after lowering the settings or enabling FSR, the game remains playable.

In Dota 2 and other undemanding competitive games, the Radeon 680M allows for Full HD with high settings. The Witcher 3 and similarly demanding titles typically run at Full HD on low or medium presets.

Modern heavy titles require low settings, FSR, and sometimes frame rate limitations around 30 FPS. In the most demanding games, it is necessary to reduce the internal resolution to 1600 × 900 or 1280 × 720.

On screens with resolutions of 2560 × 1600 and higher, running heavy games at native resolution places too much strain on the Radeon 680M. Scaling can help increase frame rates, but it significantly reduces image sharpness.

Hardware ray tracing is supported; however, there is virtually no headroom for it. Even standard rendering in modern games can fully load all compute units.

Radeon 680M vs. Radeon 660M

The difference is clearly visible in two versions of the Morefine M600 mini-PC. The configuration with the Radeon 680M scored 2408 points in Time Spy Graphics, while the model with the Radeon 660M scored 1430 points. The advantage of the higher-end graphics was 68%.

In Fire Strike Graphics, the gap was smaller but still reached 44%: 6644 versus 4621 points.

This is not an entirely clean comparison as the devices use different processors. However, the results clearly illustrate the difference between the classes. The Radeon 680M often allows for maintaining Full HD and using medium settings, while the Radeon 660M quickly hits minimum presets and lower resolutions.

With only a slight price difference, a laptop with the Radeon 680M is significantly more advantageous for gaming.

Radeon 680M vs. Radeon 780M

The Radeon 780M has transitioned to the RDNA 3 architecture but retains 12 compute units. It operates at a higher frequency and is generally faster than the 680M; however, the performance increase is far more modest than that observed when moving from 660M to 680M.

A fast Radeon 680M with LPDDR5-6400 and a normal power limit can come close to the Radeon 780M in a slim laptop where the new iGPU is more power-constrained.

The transition from the Radeon 680M to the 780M yields a moderate boost and rarely justifies the replacement of a functioning laptop. When purchasing a new device, the Radeon 780M appears more attractive, but the overall speed still depends on memory, power, and cooling.

Conclusion

The AMD Radeon 680M is still suitable for laptops and mini-PCs without a discrete graphics card, which plan to run games regularly. It handles competitive titles, many older AAA games, and some modern releases at low settings while utilizing FSR.

The main advantage over the Radeon 660M is the doubled number of compute units. In Time Spy, the difference between them can reach almost 70%, so they are truly different classes of integrated graphics.

However, the name Radeon 680M does not guarantee a specific speed. In the tested devices, the Time Spy Graphics scores range from 1896 to 2452 points. Therefore, when selecting a model, it is essential to consider not only the processor but also the memory, power limits, and specific device results.

Basic

Label Name
AMD
Platform
Integrated
Launch Date
January 2022
Model Name
Radeon 680M
Generation
Navi II IGP
Base Clock
2000MHz
Boost Clock
2200MHz
Bus Interface
PCIe 4.0 x8
Transistors
13,100 million
RT Cores
12
Compute Units
12
TMUs
?
Texture Mapping Units (TMUs) serve as components of the GPU, which are capable of rotating, scaling, and distorting binary images, and then placing them as textures onto any plane of a given 3D model. This process is called texture mapping.
48
Foundry
TSMC
Process Size
6 nm
Architecture
RDNA 2.0

Memory Specifications

Memory Size
System Shared
Memory Type
System Shared
Memory Bus
?
The memory bus width refers to the number of bits of data that the video memory can transfer within a single clock cycle. The larger the bus width, the greater the amount of data that can be transmitted instantaneously, making it one of the crucial parameters of video memory. The memory bandwidth is calculated as: Memory Bandwidth = Memory Frequency x Memory Bus Width / 8. Therefore, when the memory frequencies are similar, the memory bus width will determine the size of the memory bandwidth.
System Shared
Memory Clock
SystemShared
Bandwidth
?
Memory bandwidth refers to the data transfer rate between the graphics chip and the video memory. It is measured in bytes per second, and the formula to calculate it is: memory bandwidth = working frequency × memory bus width / 8 bits.
System Dependent

Display and Media

Outputs
Portable Device Dependent

Theoretical Performance

Pixel Rate
?
Pixel fill rate refers to the number of pixels a graphics processing unit (GPU) can render per second, measured in MPixels/s (million pixels per second) or GPixels/s (billion pixels per second). It is the most commonly used metric to evaluate the pixel processing performance of a graphics card.
70.40 GPixel/s
Texture Rate
?
Texture fill rate refers to the number of texture map elements (texels) that a GPU can map to pixels in a single second.
105.6 GTexel/s
FP16 (half)
?
An important metric for measuring GPU performance is floating-point computing capability. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy.
6.758 TFLOPS
FP64 (double)
?
An important metric for measuring GPU performance is floating-point computing capability. Double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy, while single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
211.2 GFLOPS
FP32 (float)
?
An important metric for measuring GPU performance is floating-point computing capability. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
3.311 TFLOPS

Miscellaneous

Shading Units
?
The most fundamental processing unit is the Streaming Processor (SP), where specific instructions and tasks are executed. GPUs perform parallel computing, which means multiple SPs work simultaneously to process tasks.
768
L1 Cache
128 KB per Array
L2 Cache
2MB
TDP
50W
Vulkan Version
?
Vulkan is a cross-platform graphics and compute API by Khronos Group, offering high performance and low CPU overhead. It lets developers control the GPU directly, reduces rendering overhead, and supports multi-threading and multi-core processors.
1.3
OpenCL Version
2.0
OpenGL
4.6
DirectX
12 Ultimate (12_2)
Power Connectors
None
ROPs
?
The Raster Operations Pipeline (ROPs) is primarily responsible for handling lighting and reflection calculations in games, as well as managing effects like anti-aliasing (AA), high resolution, smoke, and fire. The more demanding the anti-aliasing and lighting effects in a game, the higher the performance requirements for the ROPs; otherwise, it may result in a sharp drop in frame rate.
32
Shader Model
6.7

Benchmarks

FP32 (float)
Score
3.311 TFLOPS
3DMark Steel Nomad
Score
361
3DMark Time Spy
Score
2399
Blender
Score
249

Compared to Other GPU

FP32 (float) / TFLOPS
3.508 +5.9%
3.363 +1.6%
3.311
3.231 -2.4%
3.07 -7.3%
3DMark Time Spy
4844 +101.9%
1705 -28.9%
733 -69.4%
3DMark Steel Nomad
328 -9.1%
Blender
1408.56 +465.7%
802 +222.1%
391 +57%
45.58 -81.7%