AMD Radeon 660M vs NVIDIA Quadro M2000 Comparison

AMD
RADEON

AMD Radeon 660M

CORE STATE Rembrandt
VRAM System Shared
CLOCK SPEED 1900 MHz
TDP 40 W
BUS WIDTH System Shared
ARCHITECTURE RDNA 2.0
nm
PROCESS 6 nm
LAUNCH DATE 2022
VS
NVIDIA
GEFORCE

Quadro M2000

CORE STATE GM206
VRAM 4 GB
CLOCK SPEED 1163 MHz
TDP 75 W
BUS WIDTH 128 bit
ARCHITECTURE Maxwell 2.0
nm
PROCESS 28 nm
LAUNCH DATE 2016

PERFORMANCE BENCHMARKS

geekbench_opencl
12,876
14,588
geekbench_vulkan
14,748
14,475

Analysis: AMD Radeon 660M vs NVIDIA Quadro M2000

The NVIDIA Quadro M2000 and AMD Radeon 660M represent two distinct approaches to mobile and workstation graphics, separated by six years of architectural evolution. The data shows a split decision: the Quadro M2000 takes the Geekbench OpenCL test, while the Radeon 660M counters in Vulkan. With one win apiece, the overall average benchmark scores land close—14,532 for the Quadro against 13,812 for the Radeon—but the underlying specifications and use cases diverge sharply.

Head-to-Head Benchmarks

In the Geekbench OpenCL test, the NVIDIA Quadro M2000 posts a score of 14,588, which is 13.3% higher than the AMD Radeon 660M’s 12,876. This is the largest margin in either direction between the two cards. The Quadro’s advantage here is substantial and reflects its dedicated memory subsystem and higher raw FP32 throughput of 1.786 TFLOPS, compared to the Radeon’s 1,459.2 GFLOPS. The OpenCL workload appears to favor the Quadro’s 768 shading units and 32 ROPs, which outnumber the Radeon’s 384 shading units and 16 ROPs.

The Vulkan test flips the result. The AMD Radeon 660M scores 14,748, edging out the Quadro’s 14,475 by 1.9%. While the margin is narrower than the OpenCL gap, it is consistent with the Radeon’s architectural strengths. The Radeon 660M supports DirectX 12 Ultimate (12_2) and includes 6 ray tracing cores, whereas the Quadro M2000 is limited to DirectX 12 (12_1) with no ray tracing hardware. Vulkan’s lower-level API likely leverages the RDNA 2.0 design’s modern feature set more effectively than OpenCL does.

Looking at the nearest rivals provides context for these scores. The Quadro M2000’s 14,532 average sits between the AMD Radeon RX 5500 XT (14,692, 1.1% higher) and the NVIDIA GeForce GTX 965M (14,404, 0.9% lower). The Radeon 660M’s 13,812 average is nearly identical to the NVIDIA RTX A2000 Mobile (13,821, only 0.1% higher) and trails the AMD Radeon RX 570X (13,871) by 0.4%. Interestingly, the Radeon 660M edges out the AMD Radeon RX 7900 XT (13,745) by 0.5%, a notable result for an integrated processor. The Quadro’s percentile ranking is 56, just one point above the Radeon’s 55, indicating they occupy essentially the same tier in the overall GPU landscape despite their different designs.

Where Each One Wins

The NVIDIA Quadro M2000 wins in compute-heavy OpenCL workloads. Its 13.3% lead in that test is the decisive factor in its average score advantage. The card’s dedicated 4 GB of GDDR5 memory on a 128-bit bus delivers 105.8 GB/s of bandwidth, which is fixed and predictable. For tasks that rely on OpenCL—such as certain scientific simulations, financial modeling, or legacy compute applications—the Quadro’s higher pixel rate (37.22 GPixel/s) and texture rate (55.82 GTexel/s) give it a clear edge. Its 56th percentile ranking among all GPUs also reflects slightly broader competitiveness than the Radeon’s 55th.

The AMD Radeon 660M wins in Vulkan-based workloads. Its 1.9% margin may be small, but it signals better compatibility with modern graphics APIs. The Radeon’s support for DirectX 12 Ultimate and its 6 ray tracing cores make it the more future-ready option for gaming or rendering that utilizes these features. The Radeon also has a significant clock speed advantage: its base clock of 1500 MHz and boost of 1900 MHz dwarf the Quadro’s 796 MHz base and 1163 MHz boost. This higher frequency, combined with the RDNA 2.0 architecture’s efficiency, allows the Radeon to compensate for its lower core counts in certain scenarios. The Radeon’s FP16 throughput of 2.918 TFLOPS (2:1) also suggests better performance in workloads that can leverage half-precision math, a feature the Quadro lacks entirely.

The Radeon 660M’s system-shared memory is both a weakness and a strength. It means bandwidth is “System Dependent,” so performance scales with the host system’s memory speed. In a well-configured laptop with fast RAM, the Radeon could potentially close the gap in OpenCL, but the data as tested shows it trailing by 13.3%. Meanwhile, the Quadro’s fixed 105.8 GB/s bandwidth provides consistent performance regardless of the host platform.

The Verdict

For users prioritizing OpenCL compute performance, the NVIDIA Quadro M2000 is the clear choice. It delivers a 13.3% higher score in that specific test and a 5.2% higher average benchmark score overall (14,532 vs. 13,812). The Quadro’s dedicated memory and higher FP32 throughput make it suitable for professional workstation tasks that rely on established compute APIs. Its 56th percentile ranking also indicates it sits slightly higher in the overall performance distribution.

For users working with Vulkan or modern graphics features, the AMD Radeon 660M is the better pick. It wins the Vulkan test by 1.9% and offers architectural advantages—6 ray tracing cores, DirectX 12 Ultimate support, and a 6 nm process node—that the Quadro cannot match. The Radeon’s 1900 MHz boost clock and FP16 capability further enhance its appeal for contemporary workloads. Its 55th percentile ranking is nearly identical to the Quadro’s, but the feature set points forward rather than backward.

The data does not support a universal winner. Instead, the choice hinges on software environment. OpenCL-centric workflows favor NVIDIA; Vulkan-centric workflows favor AMD. The Quadro M2000 is end-of-life with a 2016 release, while the Radeon 660M is also end-of-life but from 2022, meaning the latter has newer architectural foundations. Neither card has a launch MSRP listed, so no pricing comparison is possible from the data.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA Quadro M2000 has an average benchmark score of 14,532, compared to the AMD Radeon 660M’s 13,812. This represents a 5.2% advantage for the Quadro.

Q: How large is the difference in the Geekbench OpenCL test?

A: The Quadro M2000 scores 14,588, while the Radeon 660M scores 12,876. The Quadro wins by 13.3%, which is the largest margin in any benchmark between the two.

Q: Does the AMD Radeon 660M win any benchmark?

A: Yes, the Radeon 660M wins the Geekbench Vulkan test with a score of 14,748, beating the Quadro’s 14,475 by 1.9%.

Q: What are the nearest rivals for each card based on average score?

A: The Quadro M2000’s closest rival is the NVIDIA GeForce GTX 965M (14,404, 0.9% lower), while the Radeon 660M’s closest rival is the NVIDIA RTX A2000 Mobile (13,821, 0.1% higher).

Q: How do the process nodes compare?

A: The Quadro M2000 uses a 28 nm process from TSMC, while the Radeon 660M uses a 6 nm process, also from TSMC. The Radeon’s node is significantly more advanced.

Q: Which card has more shading units?

A: The Quadro M2000 has 768 shading units, while the Radeon 660M has 384. The Quadro doubles the Radeon in this metric, though the Radeon compensates with higher clock speeds.

Architecture Differences

The NVIDIA Quadro M2000 is built on the Maxwell 2.0 architecture using the GM206 chip, fabricated on a 28 nm process at TSMC. It contains 2,940 million transistors on a 228 mm² die, resulting in a transistor density of 12.9M per mm². The Radeon 660M uses the RDNA 2.0 architecture with the Rembrandt chip, also from TSMC but on a 6 nm process. This chip packs 13,100 million transistors into a 208 mm² die, achieving a density of 63.0M per mm²—nearly five times denser than the Quadro.

The Radeon 660M includes 6 ray tracing cores, a feature entirely absent from the Quadro M2000. It also supports DirectX 12 Ultimate (12_2), while the Quadro is limited to DirectX 12 (12_1). Both cards support OpenGL 4.6 and Vulkan 1.4. The Radeon offers FP16 compute at 2.918 TFLOPS (2:1), whereas the Quadro has no FP16 capability listed. The Radeon’s memory is system-shared, with bandwidth dependent on the host platform, while the Quadro uses dedicated GDDR5.

The Quadro’s FP32 throughput is 1.786 TFLOPS, higher than the Radeon’s 1,459.2 GFLOPS. However, the Radeon’s pixel rate (30.40 GPixel/s) and texture rate (45.60 GTexel/s) are lower than the Quadro’s 37.22 GPixel/s and 55.82 GTexel/s. The Radeon compensates with base and boost clocks of 1500 MHz and 1900 MHz, respectively, versus the Quadro’s 796 MHz and 1163 MHz.

Specification Differences

| Specification | NVIDIA Quadro M2000 | AMD Radeon 660M |

|---|---|---|

| Process Node | 28 nm | 6 nm |

| Transistors | 2,940 million | 13,100 million |

| Die Size | 228 mm² | 208 mm² |

| Transistor Density | 12.9M / mm² | 63.0M / mm² |

| Base Clock | 796 MHz | 1500 MHz |

| Boost Clock | 1163 MHz | 1900 MHz |

| Memory | 4 GB GDDR5 | System Shared |

| Memory Bus | 128 bit | System Shared |

| Memory Bandwidth | 105.8 GB/s | System Dependent |

| Shading Units | 768 | 384 |

| TMUs | 48 | 24 |

| ROPs | 32 | 16 |

| Ray Tracing Cores | None | 6 |

| FP32 | 1.786 TFLOPS | 1,459.2 GFLOPS |

| FP16 | None | 2.918 TFLOPS (2:1) |

| TDP | 75 W | 40 W |

| Slot Width | Single-slot | IGP |

| Bus Interface | PCIe 3.0 x16 | PCIe 4.0 x8 |

| Display Outputs | 4x DisplayPort 1.2 | Portable Device Dependent |

| DirectX | 12 (12_1) | 12 Ultimate (12_2) |

| Release Date | 2016-04-07 | 2022-01-03 |

| Production Status | End-of-life | End-of-life |

DETAILED SPECIFICATIONS

SPECIFICATION
660M
Quadro M2000
Core Specs
Shading Units
384
768 +100.0%
Shaders
384
768 +100.0%
TMUs
24
48 +100.0%
ROPs
16
32 +100.0%
Compute Units
6
Clocks
Base Clock
1500 MHz
796 MHz
Boost Clock
1900 MHz
1163 MHz
Memory Clock
System Shared
1653 MHz 6.6 Gbps effective
Memory
Memory Size
System Shared
4 GB
VRAM (MB)
4,096
Memory Type
System Shared
GDDR5
Memory Bus
System Shared
128 bit
Bandwidth
System Dependent
105.8 GB/s
Cache
L1 Cache
128 KB per Array
48 KB (per SMM)
L2 Cache
2 MB
1024 KB
L0 Cache
32 KB per WGP
Performance
Pixel Rate
30.40 GPixel/s
37.22 GPixel/s
Texture Rate
45.60 GTexel/s
55.82 GTexel/s
FP32 (TFLOPS)
1,459.2 GFLOPS
1.786 TFLOPS
FP64 (TFLOPS)
91.20 GFLOPS (1:16)
55.82 GFLOPS (1:32)
FP16 (TFLOPS)
2.918 TFLOPS (2:1)
AI/RT
RT Cores
6
Power
TDP
40 W
75 W
TDP (W)
40
75 +87.5%
Suggested PSU
250 W
Power Connectors
None
None
Architecture
Architecture
RDNA 2.0
Maxwell 2.0
GPU Name
Rembrandt
GM206
Generation
Navi II IGP (Rembrandt Mobile)
Quadro Maxwell (Mx000)
Process Size
6 nm
28 nm
Transistors
13,100 million
2,940 million
Die Size
208 mm²
228 mm²
Foundry
TSMC
TSMC
Density
63.0M / mm²
12.9M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 (12_1)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
2.0
3.0
CUDA
5.2
Shader Model
6.8
6.8
Physical
Slot Width
IGP
Single-slot
Length
201 mm 7.9 inches
Height
111 mm 4.4 inches
Outputs
Portable Device Dependent
4x DisplayPort 1.2
Bus Interface
PCIe 4.0 x8
PCIe 3.0 x16
Other
Production
End-of-life
End-of-life
Predecessor
Vega II IGP
Quadro Kepler
Successor
Navi III IGP
Quadro Pascal
View Radeon 660M Details View Quadro M2000 Details