NVIDIA GeForce GTX 660 Ti vs NVIDIA GeForce GTX 950 Comparison

NVIDIA
GEFORCE

NVIDIA GeForce GTX 660 Ti

CORE STATE GK104
VRAM 2 GB
CLOCK SPEED 980 MHz
TDP 150 W
BUS WIDTH 192 bit
ARCHITECTURE Kepler
nm
PROCESS 28 nm
LAUNCH DATE 2012
VS
NVIDIA
GEFORCE

GeForce GTX 950

CORE STATE GM206
VRAM 2 GB
CLOCK SPEED 1188 MHz
TDP 90 W
BUS WIDTH 128 bit
ARCHITECTURE Maxwell 2.0
nm
PROCESS 28 nm
LAUNCH DATE 2015

PERFORMANCE BENCHMARKS

geekbench_opencl
15,113
15,662
geekbench_vulkan
15,012
16,734
geekbench_metal
N/A
7,172

Analysis: NVIDIA GeForce GTX 660 Ti vs NVIDIA GeForce GTX 950

Head-to-Head Benchmarks

The recorded data shows two benchmark comparisons between the NVIDIA GeForce GTX 660 Ti and the NVIDIA GeForce GTX 950. The GTX 950 wins both tests, establishing a clear performance advantage in the measured workloads.

In the Geekbench OpenCL test, the GTX 950 scores 15,662 points, while the GTX 660 Ti scores 15,113 points. This yields a delta of -3.5% for the GTX 660 Ti, meaning the GTX 950 is approximately 3.5% faster in this compute-oriented test. While the margin is not enormous, it is consistent and reproducible in the data.

The Geekbench Vulkan test shows a more pronounced difference. The GTX 950 scores 16,734 points, compared to the GTX 660 Ti's 15,012 points. The delta here is -10.3% for the GTX 660 Ti, indicating the GTX 950 holds a double-digit lead in this API workload. This suggests that the newer architecture handles Vulkan operations with greater efficiency, a finding supported by the API version support listed in the specifications.

The head-to-head tally shows 0 wins for the GTX 660 Ti and 2 wins for the GTX 950. Neither benchmark favors the older card. When looking at the average benchmark scores across all recorded tests, the GTX 660 Ti posts an average of 15,063, while the GTX 950 posts an average of 13,189. However, this average is influenced by the GTX 950's additional Metal benchmark result of 7,172, which is not part of the head-to-head comparison. For the two tests where both cards are measured, the GTX 950 is ahead in both.

The Verdict

Based strictly on the measured benchmark data, the NVIDIA GeForce GTX 950 is the faster card in the two workloads tested. It wins the OpenCL test by 3.5% and the Vulkan test by 10.3%. The GTX 660 Ti does not win either benchmark in the head-to-head comparison.

Users seeking better raw compute performance in modern API workloads should favor the GTX 950. The Vulkan gap is particularly notable, as it suggests the GTX 950 handles modern graphics APIs with significantly more headroom.

However, the GTX 660 Ti is not entirely without merit. It carries a higher transistor count at 3,540 million versus 2,940 million, and a larger die size at 294 mm² versus 228 mm². It also has more shading units (1,344 versus 768), more texture mapping units (112 versus 48), and a higher FP32 compute rating at 2.634 TFLOPS versus 1.825 TFLOPS. In raw theoretical throughput, the GTX 660 Ti looks stronger. Yet the actual benchmark results show the GTX 950 pulling ahead, indicating that architectural efficiency trumps raw specification advantages in practice.

For the average user looking at these two end-of-life cards, the GTX 950 is the better choice for running modern software that leverages Vulkan or OpenCL. The data does not support recommending the GTX 660 Ti over the GTX 950 for any benchmark test that was recorded.

Where Each One Wins

The GTX 950 wins in both recorded head-to-head benchmarks. It holds the advantage in Geekbench OpenCL with a score of 15,662 versus 15,113, and in Geekbench Vulkan with a score of 16,734 versus 15,012. These are the only two tests where both cards are measured directly.

The GTX 660 Ti does not win any head-to-head benchmark. However, it does have a higher average benchmark score when considering all its recorded tests (15,063 versus 13,189 for the GTX 950). This is because the GTX 950's average is dragged down by its additional Metal benchmark score of 7,172, which is a platform-specific test that the GTX 660 Ti does not have recorded. This is not a meaningful advantage for the GTX 660 Ti in a head-to-head comparison.

In terms of theoretical specifications, the GTX 660 Ti leads in several categories: pixel rate (27.44 GPixel/s versus 38.02 GPixel/s actually favors the GTX 950), texture rate (109.8 GTexel/s versus 57.02 GTexel/s favors the GTX 660 Ti), FP32 (2.634 TFLOPS versus 1.825 TFLOPS favors the GTX 660 Ti), and memory bandwidth (144.2 GB/s versus 105.8 GB/s favors the GTX 660 Ti). The GTX 950 wins in pixel rate (38.02 GPixel/s versus 27.44 GPixel/s) and ROP count (32 versus 24).

FAQ

Q: Which card has the higher average benchmark score?

A: The GTX 660 Ti has an average benchmark score of 15,063, while the GTX 950 has an average of 13,189. However, this average for the GTX 950 includes a Metal benchmark where it scores 7,172, which is not part of the head-to-head comparison.

Q: What is the performance difference in Vulkan workloads?

A: The GTX 950 scores 16,734 in Geekbench Vulkan, while the GTX 660 Ti scores 15,012. The GTX 950 is 10.3% faster in this test.

Q: How do the two cards compare in OpenCL performance?

A: The GTX 950 scores 15,662 in Geekbench OpenCL, while the GTX 660 Ti scores 15,113. The GTX 950 is 3.5% faster in this test.

Q: Which card has a higher transistor count?

A: The GTX 660 Ti has 3,540 million transistors, while the GTX 950 has 2,940 million transistors. The GTX 660 Ti has a larger die size as well at 294 mm² versus 228 mm².

Q: Which card has a lower power consumption rating?

A: The GTX 950 has a lower TDP of 90 W, while the GTX 660 Ti has a TDP of 150 W. The GTX 950 also requires a 250 W power supply, versus 450 W for the GTX 660 Ti.

Q: Which card supports a newer version of DirectX?

A: The GTX 950 supports DirectX 12 (12_1), while the GTX 660 Ti supports DirectX 12 (11_0). The GTX 950 also supports Vulkan 1.4, while the GTX 660 Ti supports Vulkan 1.2.175.

Architecture Differences

The GTX 660 Ti is built on the Kepler architecture, using the GK104 chip. The GTX 950 uses the Maxwell 2.0 architecture, based on the GM206 chip. Both are manufactured on a 28 nm process at TSMC, but they represent different generations: the GTX 660 Ti is from the GeForce 600 series, while the GTX 950 is from the GeForce 900 series.

The transistor density differs slightly, with the GTX 660 Ti at 12.0M transistors per mm and the GTX 950 at 12.9M per mm. The GTX 660 Ti has a larger die (294 mm versus 228 mm) and more transistors (3,540 million versus 2,940 million), but the GTX 950 packs its transistors more densely.

The shading units differ significantly: the GTX 660 Ti has 1,344 shading units, while the GTX 950 has 768. Texture mapping units also differ (112 versus 48), but the GTX 950 has more ROPs (32 versus 24). The GTX 660 Ti has a higher texture rate (109.8 GTexel/s versus 57.02 GTexel/s) and higher FP32 compute (2.634 TFLOPS versus 1.825 TFLOPS), but the GTX 950 has a higher pixel rate (38.02 GPixel/s versus 27.44 GPixel/s).

The memory subsystem differs as well. The GTX 660 Ti uses a 192-bit memory bus with bandwidth of 144.2 GB/s, while the GTX 950 uses a 128-bit bus with bandwidth of 105.8 GB/s. Both have 2 GB of GDDR5 memory, but the GTX 950 runs at a higher effective memory speed of 6.6 Gbps versus 6 Gbps.

The API support shows the GTX 950 as more modern: it supports DirectX 12 (12_1) versus the GTX 660 Ti's 12 (11_0), and Vulkan 1.4 versus 1.2.175. OpenGL 4.6 is supported by both. The GTX 950 also has more display outputs, including 3x DisplayPort 1.2 and 1x HDMI 2.0. The GTX 660 Ti has 1x DisplayPort 1.2 and 1x HDMI 1.4a.

Specification Differences

The two cards differ in several specification fields. The GTX 660 Ti has a base clock of 915 MHz and a boost clock of 980 MHz, while the GTX 950 has a base clock of 1,024 MHz and a boost clock of 1,188 MHz. The GTX 950 runs at higher clock speeds, which helps explain its benchmark advantage.

Memory configuration differs: both have 2 GB of GDDR5, but the GTX 660 Ti uses a 192-bit bus (144.2 GB/s bandwidth) while the GTX 950 uses a 128-bit bus (105.8 GB/s bandwidth). The GTX 950 has a higher effective memory speed of 6.6 Gbps versus 6 Gbps.

The power requirements are notably different. The GTX 660 Ti has a TDP of 150 W and requires a 450 W power supply with 2x 6-pin connectors. The GTX 950 has a TDP of 90 W, requires a 250 W power supply, and uses a single 6-pin connector. This makes the GTX 950 significantly more power-efficient.

Physical dimensions also differ: the GTX 660 Ti is 241 mm (9.5 inches) long, while the GTX 950 is 202 mm (8 inches) long. Both are dual-slot cards.

The display outputs are different: the GTX 660 Ti has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The GTX 950 has 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2.

The release dates differ by three years: the GTX 660 Ti was released in August 2012, while the GTX 950 was released in August 2015. The GTX 660 Ti has a launch MSRP of 299 USD, while the GTX 950 has a launch MSRP of 159 USD. The GTX 660 Ti belongs to the GeForce 600 series, while the GTX 950 belongs to the GeForce 900 series.

DETAILED SPECIFICATIONS

SPECIFICATION
GTX 660 Ti
GTX 950
Core Specs
Shading Units
1,344
768 -42.9%
Shaders
1,344
768 -42.9%
TMUs
112
48 -57.1%
ROPs
24
32 +33.3%
Clocks
Base Clock
915 MHz
1024 MHz
Boost Clock
980 MHz
1188 MHz
Memory Clock
1502 MHz 6 Gbps effective
1653 MHz 6.6 Gbps effective
Memory
Memory Size
2 GB
2 GB
VRAM (MB)
2,048
2,048 0.0%
Memory Type
GDDR5
GDDR5
Memory Bus
192 bit
128 bit
Bandwidth
144.2 GB/s
105.8 GB/s
Cache
L1 Cache
16 KB (per SMX)
48 KB (per SMM)
L2 Cache
384 KB
1024 KB
Performance
Pixel Rate
27.44 GPixel/s
38.02 GPixel/s
Texture Rate
109.8 GTexel/s
57.02 GTexel/s
FP32 (TFLOPS)
2.634 TFLOPS
1.825 TFLOPS
FP64 (TFLOPS)
109.8 GFLOPS (1:24)
57.02 GFLOPS (1:32)
Power
TDP
150 W
90 W
TDP (W)
150
90 -40.0%
Suggested PSU
450 W
250 W
Power Connectors
2x 6-pin
1x 6-pin
Architecture
Architecture
Kepler
Maxwell 2.0
GPU Name
GK104
GM206
Generation
GeForce 600
GeForce 900
Process Size
28 nm
28 nm
Transistors
3,540 million
2,940 million
Die Size
294 mm²
228 mm²
Foundry
TSMC
TSMC
Density
12.0M / mm²
12.9M / mm²
API Support
DirectX
12 (11_0)
12 (12_1)
OpenGL
4.6
4.6
Vulkan
1.2.175
1.4
OpenCL
3.0
3.0
CUDA
3.0
5.2
Shader Model
6.5 (5.1)
6.8
Physical
Slot Width
Dual-slot
Dual-slot
Length
241 mm 9.5 inches
202 mm 8 inches
Outputs
2x DVI1x HDMI 1.4a1x DisplayPort 1.2
1x DVI1x HDMI 2.03x DisplayPort 1.2
Bus Interface
PCIe 3.0 x16
PCIe 3.0 x16
Other
Launch Price
299 USD
159 USD
Production
End-of-life
End-of-life
Predecessor
GeForce 500
GeForce 700
Successor
GeForce 700
GeForce 10
View GeForce GTX 660 Ti Details View GeForce GTX 950 Details