NVIDIA GeForce 945M vs NVIDIA GeForce GTX 1660 Ti Comparison

NVIDIA
GEFORCE

NVIDIA GeForce 945M

CORE STATE GM107
VRAM 2 GB
CLOCK SPEED 1020 MHz
TDP 75 W
BUS WIDTH 128 bit
ARCHITECTURE Maxwell
nm
PROCESS 28 nm
LAUNCH DATE 2015
VS
NVIDIA
GEFORCE

GeForce GTX 1660 Ti

CORE STATE TU116
VRAM 6 GB
CLOCK SPEED 1770 MHz
TDP 120 W
BUS WIDTH 192 bit
ARCHITECTURE Turing
nm
PROCESS 12 nm
LAUNCH DATE 2019

PERFORMANCE BENCHMARKS

geekbench_opencl
8,099
51,029
3dmark_3dmark_steel_nomad_dx12
N/A
1,061
geekbench_vulkan
N/A
5,827
passmark_directx_10
N/A
66
passmark_directx_11
N/A
103
passmark_directx_12
N/A
51
passmark_directx_9
N/A
190
passmark_g2d
N/A
804
passmark_g3d
N/A
12,871
passmark_gpu_compute
N/A
5,226

Analysis: NVIDIA GeForce 945M vs NVIDIA GeForce GTX 1660 Ti

The NVIDIA GeForce 945M and the NVIDIA GeForce GTX 1660 Ti represent two distinct eras of mobile graphics, separated by a chasm in architectural design and performance capability. The data shows a decisive generational leap, with the newer Turing-based GPU outperforming its Maxwell-based predecessor by a staggering margin in compute workloads. This analysis dissects the benchmark results, architectural shifts, and specification gaps to determine where each card stands.

FAQ

Q: What is the primary benchmark score difference between the two GPUs?

A: In the Geekbench OpenCL test, the GeForce 945M scores 8,099, while the GeForce GTX 1660 Ti scores 51,029. This represents a delta of -84.1% for the older card, meaning the GTX 1660 Ti is roughly 6.3 times faster in this specific compute test.

Q: Which GPU has a higher average benchmark score across all tests?

A: Despite the GTX 1660 Ti having a much higher single-test score, its average benchmark score is 7,723, which is actually lower than the 945M's average of 8,099. This is because the GTX 1660 Ti's average includes many other tests (like Passmark DX9, DX10, and G2D) where its scores are relatively low.

Q: How do the two cards compare in terms of overall GPU percentile ranking?

A: The GeForce 945M sits at the 42nd percentile of all GPUs, while the GeForce GTX 1660 Ti sits at the 41st percentile. Despite the massive performance gap in the OpenCL test, their overall percentile rankings are nearly identical, indicating that the 945M's limited test pool skews its relative position.

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

A: The 945M's closest rivals include the NVIDIA GRID K2 (0.2% ahead), AMD Radeon R9 M360 (0.4% behind), and NVIDIA GeForce GTX 650 Ti Boost (0.4% ahead). The GTX 1660 Ti's rivals include the AMD Radeon 540 (0.7% behind) and AMD Radeon Pro WX 3100 (1.9% behind).

Q: What are the respective release dates for these GPUs?

A: The NVIDIA GeForce 945M was released on October 26, 2015, while the NVIDIA GeForce GTX 1660 Ti was released on February 21, 2019. This gap of roughly 3.5 years explains the significant architectural advancements.

Q: What is the launch MSRP for the GTX 1660 Ti?

A: The GTX 1660 Ti has a launch MSRP of 279 USD. The 945M has no listed launch MSRP in the data.

Architecture Differences

The architectural chasm between these two GPUs is profound, stemming from a generational shift in design philosophy. The 945M is built on the Maxwell architecture using the GM107 chip, fabricated on a 28 nm process at TSMC. In stark contrast, the GTX 1660 Ti employs the Turing architecture with the TU116 chip, manufactured on a much more advanced 12 nm process, also at TSMC. This process shrink alone allows for a massive increase in transistor count: the 945M houses 1,870 million transistors on a 148 mm² die, while the GTX 1660 Ti packs 6,600 million transistors onto a 284 mm² die. The transistor density correspondingly jumps from 12.6M / mm² to 23.2M / mm², showcasing the efficiency gains of the newer node.

The compute core configurations differ wildly. The 945M features 640 shading units, 40 texture mapping units (TMUs), and 16 raster output units (ROPs). The GTX 1660 Ti, on the other hand, scales this up to 1,536 shading units, 96 TMUs, and 48 ROPs. This 2.4x increase in shading units and 6x increase in TMUs fundamentally changes the card's ability to handle complex geometry and texture-heavy scenes. Neither GPU features dedicated ray tracing or tensor cores, as both are from lines that predate or omit those specific Turing features.

Clock speeds also tell a story of improvement. The 945M has a base clock of 928 MHz and a boost clock of 1020 MHz. The GTX 1660 Ti operates at a significantly higher frequency, with a base clock of 1500 MHz and a boost clock of 1770 MHz. This higher frequency, combined with the architectural improvements, leads to dramatically different compute throughput: the 945M delivers 1,305.6 GFLOPS of FP32 performance, while the GTX 1660 Ti delivers 5.437 TFLOPS. The newer card also supports FP16 at 10.87 TFLOPS (2:1), a feature entirely absent from the older Maxwell chip.

Head-to-Head Benchmarks

The only direct head-to-head benchmark available in the data is the Geekbench OpenCL test, and the results are lopsided. The GeForce 945M scores 8,099, while the GeForce GTX 1660 Ti scores 51,029. This is a delta of -84.1% for the 945M, meaning the GTX 1660 Ti outperforms it by a factor of approximately 6.3. This single data point strongly suggests that the newer card is in a completely different performance class for general-purpose GPU compute tasks.

However, the broader benchmark suite paints a more nuanced picture. The GTX 1660 Ti has multiple benchmark scores, including Passmark tests for various DirectX versions. Its Passmark G3D score is 12,871, which is a strong result for gaming workloads. The 945M lacks these gaming-specific benchmarks in the provided data, making a direct gaming comparison impossible. The GTX 1660 Ti's average benchmark score of 7,723 is dragged down by its low scores in legacy tests like Passmark DirectX 10 (66) and DirectX 9 (190), which are likely not representative of real-world gaming performance but do affect the aggregate statistics.

When looking at the nearest rivals, the context becomes clearer. The 945M's closest competitor is the NVIDIA GRID K2, which is only 0.2% faster in average score. This indicates that the 945M is positioned at a very specific performance tier. The GTX 1660 Ti's nearest rival is the AMD Radeon 540, which is 0.7% slower. This suggests that despite its massive OpenCL lead over the 945M, the GTX 1660 Ti's averaged performance places it in a lower tier than its raw compute power might suggest.

Specification Differences

The specification sheet reveals a stark contrast in every major category. The memory subsystem is one of the most significant differences. The 945M uses 2 GB of DDR3 memory on a 128-bit bus, yielding a bandwidth of just 28.80 GB/s. The GTX 1660 Ti, however, uses 6 GB of GDDR6 memory on a 192-bit bus, providing a massive 288.0 GB/s of bandwidth—exactly 10 times higher. This bandwidth difference is critical for modern games and high-resolution textures.

The power and physical characteristics also differ substantially. The 945M has a TDP of 75 W and uses an MXM Module form factor with no power connectors, making it suitable for laptops. The GTX 1660 Ti has a TDP of 120 W, is a dual-slot card, and requires a single 8-pin power connector. Its suggested PSU is 300 W, and it has physical dimensions of 229 mm in length, 111 mm in height, and 35 mm in width. The 945M's display outputs are listed as "Portable Device Dependent," while the GTX 1660 Ti offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a.

The bus interface also differs: the 945M uses MXM-B (3.0), which is a laptop-specific connector, while the GTX 1660 Ti uses the standard PCIe 3.0 x16 interface. DirectX support is also more advanced on the newer card: the 945M supports DirectX 12 (11_0), while the GTX 1660 Ti supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. The 945M has a transistor count of 1,870 million, while the GTX 1660 Ti has 6,600 million, and the die sizes are 148 mm² and 284 mm² respectively.

Where Each One Wins

Based on the available data, the GeForce GTX 1660 Ti is the clear winner in raw compute performance. Its Geekbench OpenCL score of 51,029 crushes the 945M's 8,099, making it the obvious choice for any workload that leverages GPU compute, such as video encoding, machine learning inference, or scientific calculations. The 945M's performance in this test is only slightly better than its nearest rivals, such as the NVIDIA GRID K2 (8,080) and AMD Radeon R9 M360 (8,129), indicating it is firmly a low-tier performer.

The 945M's wins are more subtle. Its average benchmark score of 8,099 is actually higher than the GTX 1660 Ti's average of 7,723, which could indicate that it is more consistent across a limited set of tests. However, this is likely an artifact of the test pool: the 945M has only one benchmark result, while the GTX 1660 Ti's average is pulled down by low scores in legacy DirectX tests. The 945M also has a slightly higher overall percentile ranking (42 vs 41), but this is marginal and likely not meaningful given the different test distributions.

In terms of efficiency, the 945M's 75 W TDP is significantly lower than the GTX 1660 Ti's 120 W, making it a more power-efficient option for mobile devices. However, this comes at the cost of massive performance loss. The GTX 1660 Ti's higher transistor density (23.2M / mm² vs 12.6M / mm²) suggests it is more efficient per square millimeter of silicon, but the data does not provide a direct performance-per-watt comparison.

The Verdict

The data overwhelmingly favors the NVIDIA GeForce GTX 1660 Ti for anyone prioritizing performance. Its 6.3x lead in the Geekbench OpenCL test, combined with its superior memory bandwidth (288.0 GB/s vs 28.80 GB/s) and higher shading unit count (1,536 vs 640), makes it the obvious choice for demanding applications. The GTX 1660 Ti's support for DirectX 12 (12_1) and FP16 compute also future-proofs it better than the 945M's DirectX 12 (11_0) and lack of FP16 support. With a launch MSRP of 279 USD, it is a desktop-oriented card that offers substantial compute capability.

The NVIDIA GeForce 945M, on the other hand, is a relic of the Maxwell era designed for thin-and-light laptops. Its 75 W TDP and MXM module form factor make it suitable for portable devices, but its DDR3 memory and 28 nm process node are severely outdated. The data shows that its nearest rivals are other low-end mobile and workstation GPUs like the GRID K2 and R9 M360, placing it firmly in the entry-level segment. If the use case is strictly legacy laptop upgrades or power-constrained environments, the 945M could serve, but benchmark results indicate it is vastly outclassed.

For most users, the GTX 1660 Ti is the only logical choice. The 945M's sole advantage—a lower power draw—does not compensate for the 84.1% performance deficit in compute tasks. The percentile rankings of 42 and 41 are misleadingly similar, but the raw scores tell the real story: the GTX 1660 Ti is a modern, capable GPU, while the 945M is an obsolete part that should be avoided for any new system build.

DETAILED SPECIFICATIONS

SPECIFICATION
945M
GTX 1660 Ti
Core Specs
Shading Units
640
1,536 +140.0%
Shaders
640
1,536 +140.0%
TMUs
40
96 +140.0%
ROPs
16
48 +200.0%
SM Count
—
24
Clocks
Base Clock
928 MHz
1500 MHz
Boost Clock
1020 MHz
1770 MHz
Memory Clock
900 MHz 1800 Mbps effective
1500 MHz 12 Gbps effective
Memory
Memory Size
2 GB
6 GB
VRAM (MB)
2,048
6,144 +200.0%
Memory Type
DDR3
GDDR6
Memory Bus
128 bit
192 bit
Bandwidth
28.80 GB/s
288.0 GB/s
Cache
L1 Cache
64 KB (per SMM)
64 KB (per SM)
L2 Cache
2 MB
1536 KB
Performance
Pixel Rate
16.32 GPixel/s
84.96 GPixel/s
Texture Rate
40.80 GTexel/s
169.9 GTexel/s
FP32 (TFLOPS)
1,305.6 GFLOPS
5.437 TFLOPS
FP64 (TFLOPS)
40.80 GFLOPS (1:32)
169.9 GFLOPS (1:32)
FP16 (TFLOPS)
—
10.87 TFLOPS (2:1)
Power
TDP
75 W
120 W
TDP (W)
75
120 +60.0%
Suggested PSU
—
300 W
Power Connectors
None
1x 8-pin
Architecture
Architecture
Maxwell
Turing
GPU Name
GM107
TU116
Generation
GeForce 900M
GeForce 16
Process Size
28 nm
12 nm
Transistors
1,870 million
6,600 million
Die Size
148 mm²
284 mm²
Foundry
TSMC
TSMC
Density
12.6M / mm²
23.2M / mm²
API Support
DirectX
12 (11_0)
12 (12_1)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
5.0
7.5
Shader Model
6.7 (5.1)
6.8
Physical
Slot Width
MXM Module
Dual-slot
Length
—
229 mm 9 inches
Height
—
111 mm 4.4 inches
Outputs
Portable Device Dependent
1x DVI1x HDMI 2.01x DisplayPort 1.4a
Bus Interface
MXM-B (3.0)
PCIe 3.0 x16
Other
Launch Price
—
279 USD
Production
End-of-life
End-of-life
Predecessor
GeForce 800M
GeForce 10
Successor
GeForce 10 Mobile
GeForce 20
View GeForce 945M Details View GeForce GTX 1660 Ti Details