AMD Radeon HD 8970M vs NVIDIA GeForce GTX 1060 3 GB Comparison

AMD
RADEON

AMD Radeon HD 8970M

CORE STATE Neptune
VRAM 4 GB
CLOCK SPEED 900 MHz
TDP 100 W
BUS WIDTH 256 bit
ARCHITECTURE GCN 1.0
nm
PROCESS 28 nm
LAUNCH DATE 2013
VS
NVIDIA
GEFORCE

GeForce GTX 1060 3 GB

CORE STATE GP106
VRAM 3 GB
CLOCK SPEED 1708 MHz
TDP 120 W
BUS WIDTH 192 bit
ARCHITECTURE Pascal
nm
PROCESS 16 nm
LAUNCH DATE 2016

PERFORMANCE BENCHMARKS

geekbench_opencl
21,237
27,001
geekbench_metal
N/A
13,527
geekbench_vulkan
N/A
17,475

Analysis: AMD Radeon HD 8970M vs NVIDIA GeForce GTX 1060 3 GB

The AMD Radeon HD 8970M and the NVIDIA GeForce GTX 1060 3 GB represent two very different eras of mobile and desktop graphics. The data shows a clear overall winner in raw compute, but the comparison is more nuanced when considering the specific strengths of each architecture. The GTX 1060 3 GB holds a decisive lead in the only shared benchmark, yet the HD 8970M offers a unique feature set and a different performance profile that may still appeal to specific use cases.

Where Each One Wins

The benchmark results indicate a clear split between the two GPUs, with the NVIDIA GeForce GTX 1060 3 GB taking the sole victory in the head-to-head comparison. The data shows the GTX 1060 3 GB winning the geekbench_opencl test with a score of 27001, compared to the AMD Radeon HD 8970M's score of 21237. This represents a 21.3% advantage for the NVIDIA part in this specific compute workload. The HD 8970M has no wins in the head-to-head benchmarks, meaning that in the single test where both were measured, it came out behind.

However, the win count does not tell the full story. The AMD Radeon HD 8970M's average benchmark score of 21237 is actually higher than the GTX 1060 3 GB's average score of 19334. This is because the NVIDIA part's average is dragged down by its performance in other API tests, such as geekbench_metal (13527) and geekbench_vulkan (17475), which are not available for the AMD part. In terms of the OpenCL test that both share, the NVIDIA part is the clear winner. The GTX 1060 3 GB also has a higher percentile ranking at 64, compared to the HD 8970M's percentile of 66, which is a counterintuitive but data-driven fact. The percentile vs all GPUs places the HD 8970M slightly higher despite the head-to-head loss, indicating that its single score is more competitive against the broader GPU landscape.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The AMD Radeon HD 8970M uses the Neptune chip based on the GCN 1.0 architecture, manufactured on a 28 nm process at TSMC. It features 2,800 million transistors on a 212 mm² die, resulting in a transistor density of 13.2M per mm². In contrast, the NVIDIA GeForce GTX 1060 3 GB uses the GP106 chip based on the Pascal architecture, also built by TSMC but on a much more advanced 16 nm process. This newer node allows it to pack 4,400 million transistors into a slightly smaller 200 mm² die, achieving a much higher transistor density of 22.0M per mm².

The core configurations differ significantly. The AMD part has 1280 shading units, 80 texture mapping units (TMUs), and 32 raster output units (ROPs). The NVIDIA part has fewer shading units at 1152 and fewer TMUs at 72, but it has significantly more ROPs at 48. Clock speeds also diverge sharply. The HD 8970M runs at a base clock of 850 MHz with a boost of 900 MHz, while the GTX 1060 3 GB runs at a base of 1506 MHz and boosts to 1708 MHz. This clock advantage helps the NVIDIA part achieve a much higher pixel rate of 81.98 GPixel/s and texture rate of 123.0 GTexel/s, compared to the AMD part's 28.80 GPixel/s and 72.00 GTexel/s. In terms of raw floating-point performance, the GTX 1060 3 GB delivers 3.935 TFLOPS versus the HD 8970M's 2.304 TFLOPS.

Memory configurations also differ. The HD 8970M has 4 GB of GDDR5 on a 256-bit bus, delivering 153.6 GB/s of bandwidth. The GTX 1060 3 GB has less memory at 3 GB, on a narrower 192-bit bus, but it achieves higher bandwidth at 192.2 GB/s due to its faster 8 Gbps effective memory speed. The NVIDIA part also supports a wider range of modern APIs, including DirectX 12 (12_1) and Vulkan 1.4, while the AMD part is limited to DirectX 12 (11_1) and Vulkan 1.2.170. Both support OpenGL 4.6. The GTX 1060 3 GB has a higher TDP at 120 W compared to the HD 8970M's 100 W, reflecting its higher clock speeds and performance.

Head-to-Head Benchmarks

The only direct benchmark comparison available is the geekbench_opencl test, and the results are unambiguous. The NVIDIA GeForce GTX 1060 3 GB scores 27001, which is 21.3% higher than the AMD Radeon HD 8970M's score of 21237. This is a substantial margin that highlights the generational leap in compute performance. The GTX 1060 3 GB's advantage in this test can be attributed to its significantly higher clock speeds and superior FP32 throughput. The HD 8970M's score of 21237 places it in a competitive position relative to its own nearest rivals, being 0.4% ahead of the AMD Radeon RX Vega M GL (21153) and 1% ahead of the NVIDIA GeForce RTX 5050 (21035), but 0.7% behind the NVIDIA RTX A4000 Mobile (21379) and 1.8% behind the NVIDIA Quadro RTX 5000 (21629).

For the GTX 1060 3 GB, its OpenCL score of 27001 is even more impressive when compared to its nearest rivals. It is 0.4% ahead of the AMD Radeon RX 7900 XTX (19410), 0.8% ahead of the NVIDIA TITAN Xp (19177), and 0.9% ahead of the NVIDIA GeForce GTX 780 (19164). This shows that the GTX 1060 3 GB, despite being a mid-range part from 2016, can outperform much more expensive and higher-tier GPUs in this particular workload. The data also shows that the GTX 1060 3 GB has a much higher pixel rate (81.98 GPixel/s) and texture rate (123.0 GTexel/s) than the HD 8970M, which would translate to better performance in rasterization-heavy tasks, even though no direct benchmark for those is provided.

The Verdict

From the data, the NVIDIA GeForce GTX 1060 3 GB is the superior performer for compute-intensive tasks, as evidenced by its 21.3% lead in the geekbench_opencl benchmark. Its higher clock speeds, greater FP32 throughput, and more advanced Pascal architecture make it the clear choice for users who prioritize raw processing power. The GTX 1060 3 GB also offers a more modern feature set, including support for DirectX 12 (12_1) and Vulkan 1.4, which could be important for newer software titles. Its higher ROP count of 48 also suggests better fill-rate performance, which is beneficial for high-resolution rendering.

The AMD Radeon HD 8970M, however, is not without its merits. It offers a larger 4 GB memory capacity compared to the GTX 1060 3 GB's 3 GB, which could be advantageous in scenarios with high memory usage. Its lower TDP of 100 W makes it more power-efficient, and its MXM module form factor is designed for laptops, making it a more suitable option for mobile workstations. The HD 8970M's percentile ranking of 66 is also higher than the GTX 1060 3 GB's 64, suggesting that its single OpenCL score is more competitive relative to the entire GPU market. For users with a legacy system that requires an MXM module, the HD 8970M is the only viable option here.

FAQ

Q: Which GPU is faster in the geekbench_opencl test?

A: The NVIDIA GeForce GTX 1060 3 GB is faster, scoring 27001 compared to the AMD Radeon HD 8970M's 21237, a difference of 21.3%.

Q: How much memory does each GPU have?

A: The AMD Radeon HD 8970M has 4 GB of GDDR5 memory, while the NVIDIA GeForce GTX 1060 3 GB has 3 GB of GDDR5 memory.

Q: What are the architecture and process node for each GPU?

A: The AMD Radeon HD 8970M uses the GCN 1.0 architecture on a 28 nm process, while the NVIDIA GeForce GTX 1060 3 GB uses the Pascal architecture on a 16 nm process.

Q: Which GPU has a higher transistor count?

A: The NVIDIA GeForce GTX 1060 3 GB has 4,400 million transistors, compared to the AMD Radeon HD 8970M's 2,800 million.

Q: What is the TDP for each GPU?

A: The AMD Radeon HD 8970M has a TDP of 100 W, while the NVIDIA GeForce GTX 1060 3 GB has a higher TDP of 120 W.

Q: Which GPU supports a newer version of Vulkan?

A: The NVIDIA GeForce GTX 1060 3 GB supports Vulkan 1.4, whereas the AMD Radeon HD 8970M supports Vulkan 1.2.170.

Specification Differences

| Specification | AMD Radeon HD 8970M | NVIDIA GeForce GTX 1060 3 GB |

| :--- | :--- | :--- |

| Architecture | GCN 1.0 | Pascal |

| Process Node | 28 nm | 16 nm |

| Transistors | 2,800 million | 4,400 million |

| Die Size | 212 mm² | 200 mm² |

| Base Clock | 850 MHz | 1506 MHz |

| Boost Clock | 900 MHz | 1708 MHz |

| Memory Size | 4 GB | 3 GB |

| Memory Bus | 256 bit | 192 bit |

| Memory Bandwidth | 153.6 GB/s | 192.2 GB/s |

| Shading Units | 1280 | 1152 |

| TMUs | 80 | 72 |

| ROPs | 32 | 48 |

| Pixel Rate | 28.80 GPixel/s | 81.98 GPixel/s |

| Texture Rate | 72.00 GTexel/s | 123.0 GTexel/s |

| FP32 Performance | 2.304 TFLOPS | 3.935 TFLOPS |

| TDP | 100 W | 120 W |

| Slot Width | MXM Module | Dual-slot |

| DirectX Support | 12 (11_1) | 12 (12_1) |

| Vulkan Support | 1.2.170 | 1.4 |

| Release Date | 2013-05-13 | 2016-08-17 |

DETAILED SPECIFICATIONS

SPECIFICATION
HD 8970M
GTX 1060 3 GB
Core Specs
Shading Units
1,280
1,152 -10.0%
Shaders
1,280
1,152 -10.0%
TMUs
80
72 -10.0%
ROPs
32
48 +50.0%
Compute Units
20
—
SM Count
—
9
Clocks
Base Clock
850 MHz
1506 MHz
Boost Clock
900 MHz
1708 MHz
Memory Clock
1200 MHz 4.8 Gbps effective
2002 MHz 8 Gbps effective
Memory
Memory Size
4 GB
3 GB
VRAM (MB)
4,096
3,072 -25.0%
Memory Type
GDDR5
GDDR5
Memory Bus
256 bit
192 bit
Bandwidth
153.6 GB/s
192.2 GB/s
Cache
L1 Cache
16 KB (per CU)
48 KB (per SM)
L2 Cache
512 KB
1536 KB
Performance
Pixel Rate
28.80 GPixel/s
81.98 GPixel/s
Texture Rate
72.00 GTexel/s
123.0 GTexel/s
FP32 (TFLOPS)
2.304 TFLOPS
3.935 TFLOPS
FP64 (TFLOPS)
144.0 GFLOPS (1:16)
123.0 GFLOPS (1:32)
FP16 (TFLOPS)
—
61.49 GFLOPS (1:64)
Power
TDP
100 W
120 W
TDP (W)
100
120 +20.0%
Suggested PSU
—
300 W
Power Connectors
—
1x 6-pin
Architecture
Architecture
GCN 1.0
Pascal
GPU Name
Neptune
GP106
Generation
Solar System (HD 8900M)
GeForce 10
Process Size
28 nm
16 nm
Transistors
2,800 million
4,400 million
Die Size
212 mm²
200 mm²
Foundry
TSMC
TSMC
Density
13.2M / mm²
22.0M / mm²
API Support
DirectX
12 (11_1)
12 (12_1)
OpenGL
4.6
4.6
Vulkan
1.2.170
1.4
OpenCL
2.1 (1.2)
3.0
CUDA
—
6.1
Shader Model
6.5 (5.1)
6.8
Physical
Slot Width
MXM Module
Dual-slot
Length
—
250 mm 9.8 inches
Outputs
Portable Device Dependent
1x DVI1x HDMI 2.03x DisplayPort 1.4a
Bus Interface
PCIe 3.0 x16
PCIe 3.0 x16
Other
Launch Price
—
199 USD
Production
End-of-life
End-of-life
Predecessor
London
GeForce 900
Successor
Gem System
GeForce 20
View Radeon HD 8970M Details View GeForce GTX 1060 3 GB Details