AMD Radeon 860M vs NVIDIA GeForce RTX 3090 Comparison

AMD
RADEON

AMD Radeon 860M

CORE STATE Krackan Point
VRAM System Shared
CLOCK SPEED 3000 MHz
TDP 15 W
BUS WIDTH System Shared
ARCHITECTURE RDNA 3.5
nm
PROCESS 4 nm
LAUNCH DATE 2025
VS
NVIDIA
GEFORCE

GeForce RTX 3090

CORE STATE GA102
VRAM 24 GB
CLOCK SPEED 1695 MHz
TDP 350 W
BUS WIDTH 384 bit
ARCHITECTURE Ampere
nm
PROCESS 8 nm
LAUNCH DATE 2020

PERFORMANCE BENCHMARKS

geekbench_opencl
22,759
172,758
geekbench_vulkan
30,043
53,927
3dmark_3dmark_steel_nomad_dx12
N/A
5,118
passmark_directx_10
N/A
182
passmark_directx_11
N/A
220
passmark_directx_12
N/A
110
passmark_directx_9
N/A
268
passmark_g2d
N/A
1,063
passmark_g3d
N/A
26,645
passmark_gpu_compute
N/A
15,356

Analysis: AMD Radeon 860M vs NVIDIA GeForce RTX 3090

The benchmark data presents a stark generational and performance-class divide. The NVIDIA GeForce RTX 3090 is a former flagship discrete GPU, while the AMD Radeon 860M is a modern integrated graphics processor. Across the shared benchmarks, the RTX 3090 dominates, but the 860M's efficiency and modern architecture make it a compelling option for a completely different use case. The data shows that the RTX 3090 wins 2 out of 2 head-to-head benchmarks, with a significant margin in compute workloads.

The Verdict

The verdict is clear: these are products for entirely different systems and users. The NVIDIA GeForce RTX 3090 is the unequivocal choice for any workload requiring maximum raw compute and memory bandwidth. Its average benchmark score of 27,565 places it in the 73rd percentile of all GPUs, trading blows with the NVIDIA GeForce RTX 4070 Mobile (0.5% delta) and AMD Radeon RX 6700 XT (0.5% delta). With 24 GB of GDDR6X memory and a 384-bit bus, it is designed to handle massive datasets and high-resolution rendering without compromise.

The AMD Radeon 860M is a different beast entirely. As an integrated GPU with a 15 W TDP, it is not a competitor to the RTX 3090. Its average benchmark score of 26,401 puts it in the 72nd percentile, remarkably close to its rival, the NVIDIA GeForce MX550 (-0.1% delta). This indicates that for its power envelope, it delivers exceptional performance. The data suggests it is the correct choice for thin-and-light laptops where battery life and low heat are paramount, and where the user needs a capable iGPU for light gaming and general productivity. Benchmark results indicate the 860M is a leader in its class, but that class is fundamentally not the same as the RTX 3090's.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce RTX 3090 has a higher average benchmark score of 27,565, compared to the AMD Radeon 860M's 26,401. This represents a performance gap of about 4.4%.

Q: How do the two compare in the Geekbench OpenCL test?

A: The RTX 3090 is vastly superior, scoring 172,758 points compared to the 860M's 22,759 points. This is a delta of 659.1% in favor of the NVIDIA GPU.

Q: What is the difference in power consumption?

A: The RTX 3090 has a TDP of 350 W, while the AMD Radeon 860M has a TDP of just 15 W. This makes the 860M over 23 times more power-efficient on paper.

Q: What memory configurations do they use?

A: The RTX 3090 uses 24 GB of dedicated GDDR6X memory on a 384-bit bus with 936.2 GB/s of bandwidth. The Radeon 860M uses System Shared memory, meaning its size, type, and bus width are all "System Shared," with bandwidth being "System Dependent."

Q: Which GPU has higher clock speeds?

A: The AMD Radeon 860M has a higher boost clock of 3000 MHz, compared to the RTX 3090's boost clock of 1695 MHz. However, the RTX 3090 has a higher base clock at 1395 MHz versus 600 MHz.

Architecture Differences

The architectural divide between these two GPUs is fundamental. The NVIDIA GeForce RTX 3090 is built on the Ampere architecture, using the GA102 chip manufactured on an 8 nm process at Samsung. This is a massive, discrete GPU with 28,300 million transistors on a 628 mm² die. It is equipped with 82 RT cores and 328 Tensor cores, making it a powerhouse for hardware-accelerated ray tracing and AI workloads. Its feature set is anchored by support for DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

In contrast, the AMD Radeon 860M is based on the RDNA 3.5 architecture, using the Krackan Point chip. This is a modern, highly efficient integrated GPU built on a 4 nm process at TSMC. The data notes its generation as "Navi III IGP (Strix Point Mobile)," and its transistor count and die size are listed as unknown, reflecting its integrated nature. It has 8 RT cores but no Tensor cores listed, and it supports the same API levels: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The 860M's architecture focuses on delivering solid performance within a minimal power envelope, while the RTX 3090's architecture is built for maximum throughput.

Specification Differences

The specification differences are extensive, highlighting their disparate roles. The RTX 3090 has 10,496 shading units, 328 TMUs, and 112 ROPs, while the 860M has 512 shading units, 32 TMUs, and 16 ROPs. This 20x difference in shading units directly translates to raw pixel and texture processing capabilities. The RTX 3090’s pixel rate is 189.8 GPixel/s and texture rate is 556.0 GTexel/s, whereas the 860M manages 48.00 GPixel/s and 96.00 GTexel/s respectively. The FP32 compute performance is similarly lopsided: 35.58 TFLOPS for the RTX 3090 versus 3.072 TFLOPS for the 860M. The RTX 3090 uses a PCIe 4.0 x16 interface, while the 860M uses a narrower PCIe 4.0 x8. The RTX 3090 is a triple-slot card requiring a 750 W power supply and a 1x 12-pin connector, while the 860M is an IGP with no power connectors and a 15 W TDP. The RTX 3090 has dedicated display outputs (1x HDMI 2.1, 3x DisplayPort 1.4a), whereas the 860M's outputs are portable device dependent.

Head-to-Head Benchmarks

The head-to-head data is a clear demonstration of the performance hierarchy. In the Geekbench OpenCL test, the NVIDIA GeForce RTX 3090 scores 172,758, a massive 659.1% higher than the AMD Radeon 860M's 22,759. This test stresses general-purpose compute, and the RTX 3090's sheer number of cores and high memory bandwidth allow it to crush the integrated part. This is the single biggest win for the RTX 3090 in the data.

In the Geekbench Vulkan test, the gap narrows but remains decisive. The RTX 3090 scores 53,927, which is 79.5% higher than the 860M's 30,043. While the 860M's score is respectable for an iGPU, the RTX 3090's advantage in raw hardware resources is insurmountable. These results confirm that for any compute-intensive task, the RTX 3090 is in a different league. The RTX 3090 wins 2 of 2 head-to-head benchmarks, while the 860M wins 0.

Where Each One Wins

Based on the benchmark wins, the use cases are clear. The NVIDIA GeForce RTX 3090 wins in every scenario requiring maximum performance. This includes professional 3D rendering, high-resolution video editing, complex scientific simulations, and demanding AI model training, where its 24 GB memory and 936.2 GB/s bandwidth are critical. Its performance in the 73rd percentile, rivaling the RTX 4070 Mobile, makes it a viable option for high-end desktop workstations, though its 350 W TDP and triple-slot size demand a large chassis and robust cooling.

The AMD Radeon 860M wins in scenarios defined by efficiency and portability. Its 15 W TDP is its defining feature. It is the clear choice for ultraportable laptops where battery life is critical. While it loses to the RTX 3090 in raw compute, its average score of 26,401 and 72nd percentile ranking place it in the same performance tier as the NVIDIA GeForce MX550 (-0.1% delta), making it a strong performer for an integrated solution. It is the winner for users who need a device that can handle everyday tasks, light content creation, and esports-level gaming without needing a discrete GPU. The data shows that while the RTX 3090 is the undisputed performance king, the 860M is the efficiency champion for a mobile-first world.

DETAILED SPECIFICATIONS

SPECIFICATION
860M
RTX 3090
Core Specs
Shading Units
512
10,496 +1950.0%
Shaders
512
10,496 +1950.0%
TMUs
32
328 +925.0%
ROPs
16
112 +600.0%
Compute Units
8
SM Count
82
Clocks
Base Clock
600 MHz
1395 MHz
Boost Clock
3000 MHz
1695 MHz
Memory Clock
System Shared
1219 MHz 19.5 Gbps effective
Memory
Memory Size
System Shared
24 GB
VRAM (MB)
24,576
Memory Type
System Shared
GDDR6X
Memory Bus
System Shared
384 bit
Bandwidth
System Dependent
936.2 GB/s
Cache
L1 Cache
128 KB per Array
128 KB (per SM)
L2 Cache
1024 KB
6 MB
L0 Cache
32 KB per WGP
Performance
Pixel Rate
48.00 GPixel/s
189.8 GPixel/s
Texture Rate
96.00 GTexel/s
556.0 GTexel/s
FP32 (TFLOPS)
3.072 TFLOPS
35.58 TFLOPS
FP64 (TFLOPS)
192.0 GFLOPS (1:16)
556.0 GFLOPS (1:64)
FP16 (TFLOPS)
3.072 TFLOPS (1:1)
35.58 TFLOPS (1:1)
AI/RT
RT Cores
8
82 +925.0%
Tensor Cores
328
Power
TDP
15 W
350 W
TDP (W)
15
350 +2233.3%
Suggested PSU
750 W
Power Connectors
None
1x 12-pin
Architecture
Architecture
RDNA 3.5
Ampere
GPU Name
Krackan Point
GA102
Generation
Navi III IGP (Strix Point Mobile)
GeForce 30
Process Size
4 nm
8 nm
Transistors
unknown
28,300 million
Die Size
unknown
628 mm²
Foundry
TSMC
Samsung
Density
45.1M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
2.1
3.0
CUDA
8.6
Shader Model
6.8
6.8
Physical
Slot Width
IGP
Triple-slot
Length
336 mm 13.2 inches
Height
140 mm 5.5 inches
Outputs
Portable Device Dependent
1x HDMI 2.13x DisplayPort 1.4a
Bus Interface
PCIe 4.0 x8
PCIe 4.0 x16
Other
Launch Price
1,499 USD
Production
Active
End-of-life
Predecessor
Navi II IGP
GeForce 20
Successor
GeForce 40
View Radeon 860M Details View GeForce RTX 3090 Details