NVIDIA GeForce GT 640 vs NVIDIA GeForce GTX 750 Ti Comparison

NVIDIA
GEFORCE

NVIDIA GeForce GT 640

CORE STATE GK107
VRAM 2 GB
CLOCK SPEED —
TDP 65 W
BUS WIDTH 128 bit
ARCHITECTURE Kepler
nm
PROCESS 28 nm
LAUNCH DATE 2012
VS
NVIDIA
GEFORCE

GeForce GTX 750 Ti

CORE STATE GM107
VRAM 2 GB
CLOCK SPEED 1085 MHz
TDP 60 W
BUS WIDTH 128 bit
ARCHITECTURE Maxwell
nm
PROCESS 28 nm
LAUNCH DATE 2014

PERFORMANCE BENCHMARKS

geekbench_metal
2,086
5,182
geekbench_opencl
3,736
10,827
geekbench_vulkan
3,809
10,065
3dmark_3dmark_steel_nomad_dx12
N/A
117
passmark_directx_10
N/A
17
passmark_directx_11
N/A
26
passmark_directx_12
N/A
18
passmark_directx_9
N/A
60
passmark_g2d
N/A
528
passmark_g3d
N/A
3,903
passmark_gpu_compute
N/A
1,744

Analysis: NVIDIA GeForce GT 640 vs NVIDIA GeForce GTX 750 Ti

The NVIDIA GeForce GT 640 and the NVIDIA GeForce GTX 750 Ti represent two distinct generations of budget-oriented graphics cards. While both are end-of-life products from NVIDIA, the data reveals a significant performance gulf between them, with the GTX 750 Ti emerging as the overwhelmingly faster option. This analysis compares their benchmark scores, architectural differences, and specification gaps to provide a clear picture of their respective capabilities.

Head-to-Head Benchmarks

The benchmark data presents a decisive victory for the GTX 750 Ti across all three shared tests. In Geekbench Metal, the GTX 750 Ti scores 5,182 compared to the GT 640’s 2,086, resulting in a delta of -59.7% for the older card. This means the GTX 750 Ti delivers roughly 2.5 times the Metal performance, a margin that dwarfs the modest gains typically seen between adjacent product tiers.

The gap widens further in Geekbench OpenCL, where the GTX 750 Ti posts 10,827 against the GT 640’s 3,736. That is a delta of -65.5%, indicating the Maxwell-based card is nearly three times faster in this compute workload. The absolute difference of 7,091 points is the largest of any shared test, suggesting the GTX 750 Ti’s architectural advantages are especially pronounced in general-purpose GPU compute.

Geekbench Vulkan shows a similar trend, with the GTX 750 Ti scoring 10,065 versus 3,809 for the GT 640, a delta of -62.2%. While Vulkan is a more modern API than either card was originally designed for, the GTX 750 Ti still manages to maintain a consistent lead. Across all three tests, the GTX 750 Ti wins outright, with the GT 640 failing to secure a single head-to-head victory. The data is unambiguous: the GTX 750 Ti is not just faster, it is in a completely different performance class.

Where Each One Wins

The GTX 750 Ti wins every benchmark category where both cards have been tested. Its dominance is most apparent in OpenCL compute, where its 640 shading units and higher clock speeds translate into a 190% advantage over the GT 640. For users running GPU-accelerated applications, video encoding, or scientific workloads, the GTX 750 Ti is the only viable choice of the two.

The GT 640, by contrast, has no wins in the head-to-head data. However, its lower power draw and simpler design could be seen as advantages in very specific scenarios. With a TDP of 65 W versus 60 W for the GTX 750 Ti, the older card actually consumes slightly more power, so that potential benefit does not materialize. The GT 640’s only remaining edge is its DisplayPort 1.2 output, which the GTX 750 Ti lacks, but this is a minor connectivity feature rather than a performance metric.

In practical terms, the GTX 750 Ti is the superior card for gaming, compute, and any modern workload. The GT 640 remains functional for basic display output and legacy applications, but its 20th percentile ranking among all GPUs places it firmly in entry-level territory. The GTX 750 Ti’s 19th percentile ranking is similar, but its raw scores show it handles tasks like DirectX 11 (score 26 in Passmark) and DirectX 9 (score 60) far more effectively than the older card could manage.

Architecture Differences

The two cards are built on fundamentally different architectures. The GT 640 uses the GK107 chip based on Kepler, while the GTX 750 Ti uses the GM107 chip based on Maxwell. Both are manufactured on a 28 nm process at TSMC, but the similarities end there. The Kepler architecture in the GT 640 is older, with a transistor count of 1,270 million spread across a 118 mm² die, yielding a transistor density of 10.8M per mm². The Maxwell chip in the GTX 750 Ti packs 1,870 million transistors into a 148 mm² die, achieving a higher density of 12.6M per mm².

This generational leap is reflected in the core configurations. The GTX 750 Ti has 640 shading units, 40 texture mapping units (TMUs), and 16 render output units (ROPs). The GT 640 offers only 384 shading units, 32 TMUs, and 16 ROPs. While the ROP count is identical, the GTX 750 Ti has 67% more shading units and 25% more TMUs, directly contributing to its higher fill rates and compute throughput.

Clock speeds also differ substantially. The GTX 750 Ti has a base clock of 1020 MHz and a boost clock of 1085 MHz, whereas the GT 640’s clock data is not provided in the fact pack. However, the derived performance metrics tell the story: the GTX 750 Ti achieves a pixel rate of 17.36 GPixel/s and a texture rate of 43.40 GTexel/s, compared to 7.216 GPixel/s and 28.86 GTexel/s for the GT 640. The FP32 compute rating is 1,388.8 GFLOPS for the GTX 750 Ti versus 692.7 GFLOPS for the GT 640, confirming that Maxwell’s efficiency gains are substantial.

Memory architecture is another key differentiator. The GT 640 uses DDR3 memory with a 128-bit bus, yielding 28.51 GB/s of bandwidth. The GTX 750 Ti uses GDDR5 with the same 128-bit bus but achieves 86.40 GB/s, a threefold improvement. Both cards have 2 GB of memory, but the GTX 750 Ti’s faster memory is critical for modern textures and high-resolution assets.

Specification Differences

The most striking specification difference is memory bandwidth, where the GTX 750 Ti’s 86.40 GB/s eclipses the GT 640’s 28.51 GB/s by a factor of three. This is driven by the memory type: GDDR5 on the newer card versus DDR3 on the older one, despite both using a 128-bit interface. The memory clock also differs, with the GTX 750 Ti running at 1350 MHz (5.4 Gbps effective) versus 891 MHz (1782 Mbps effective) for the GT 640.

Shading units are another major differentiator, with the GTX 750 Ti offering 640 versus 384 on the GT 640. TMUs follow suit at 40 versus 32, while ROPs remain equal at 16. The GTX 750 Ti also has explicit base and boost clocks (1020 MHz and 1085 MHz), while the GT 640 lacks this data in the fact pack, though its derived rates are consistently lower.

Power consumption is nearly identical, with the GTX 750 Ti rated at 60 W TDP versus 65 W for the GT 640. Both cards are single-slot designs with no power connectors and a suggested PSU of 250 W. Display outputs differ: the GT 640 offers 1x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the GTX 750 Ti offers 2x DVI and 1x mini-HDMI 1.4a. The GTX 750 Ti supports Vulkan 1.4, whereas the GT 640 is limited to Vulkan 1.2.175, though both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 750 Ti launched at 149 USD, while the GT 640 launched at 99 USD, reflecting its higher performance tier.

FAQ

Q: Which card is faster in Geekbench OpenCL?

A: The GTX 750 Ti is significantly faster, scoring 10,827 versus 3,736 for the GT 640, a delta of -65.5% in favor of the newer card.

Q: Do both cards have the same amount of memory?

A: Yes, both feature 2 GB of memory, but the GTX 750 Ti uses faster GDDR5 while the GT 640 uses DDR3, resulting in 86.40 GB/s versus 28.51 GB/s bandwidth.

Q: What is the power consumption difference?

A: The GTX 750 Ti is rated at 60 W TDP, slightly lower than the GT 640’s 65 W, and both have a suggested PSU of 250 W.

Q: Which card has better compute performance for non-gaming tasks?

A: The GTX 750 Ti dominates in compute, with 1,388.8 GFLOPS FP32 performance compared to 692.7 GFLOPS for the GT 640, and it wins all three shared Geekbench tests.

Q: Are there any display output differences?

A: Yes, the GT 640 includes a DisplayPort 1.2 output, while the GTX 750 Ti offers two DVI ports and one mini-HDMI 1.4a, with no DisplayPort.

Q: How do their average benchmark scores compare?

A: The GT 640 has a higher average benchmark score of 3,210 versus 2,953 for the GTX 750 Ti, but this is skewed by the GTX 750 Ti’s inclusion of low-scoring Passmark tests; in direct head-to-head comparisons, the GTX 750 Ti wins every time.

DETAILED SPECIFICATIONS

SPECIFICATION
GT 640
GTX 750 Ti
Core Specs
Shading Units
384
640 +66.7%
Shaders
384
640 +66.7%
TMUs
32
40 +25.0%
ROPs
16
16 0.0%
Clocks
Base Clock
—
1020 MHz
Boost Clock
—
1085 MHz
GPU Clock
902 MHz
—
Memory Clock
891 MHz 1782 Mbps effective
1350 MHz 5.4 Gbps effective
Memory
Memory Size
2 GB
2 GB
VRAM (MB)
2,048
2,048 0.0%
Memory Type
DDR3
GDDR5
Memory Bus
128 bit
128 bit
Bandwidth
28.51 GB/s
86.40 GB/s
Cache
L1 Cache
16 KB (per SMX)
64 KB (per SMM)
L2 Cache
256 KB
2 MB
Performance
Pixel Rate
7.216 GPixel/s
17.36 GPixel/s
Texture Rate
28.86 GTexel/s
43.40 GTexel/s
FP32 (TFLOPS)
692.7 GFLOPS
1,388.8 GFLOPS
FP64 (TFLOPS)
28.86 GFLOPS (1:24)
43.40 GFLOPS (1:32)
Power
TDP
65 W
60 W
TDP (W)
65
60 -7.7%
Suggested PSU
250 W
250 W
Power Connectors
None
None
Architecture
Architecture
Kepler
Maxwell
GPU Name
GK107
GM107
Generation
GeForce 600
GeForce 700
Process Size
28 nm
28 nm
Transistors
1,270 million
1,870 million
Die Size
118 mm²
148 mm²
Foundry
TSMC
TSMC
Density
10.8M / mm²
12.6M / mm²
API Support
DirectX
12 (11_0)
12 (11_0)
OpenGL
4.6
4.6
Vulkan
1.2.175
1.4
OpenCL
3.0
3.0
CUDA
3.0
5.0
Shader Model
6.5 (5.1)
6.7 (5.1)
Physical
Slot Width
Single-slot
Single-slot
Length
145 mm 5.7 inches
145 mm 5.7 inches
Outputs
1x DVI1x HDMI 1.4a1x DisplayPort 1.2
2x DVI1x mini-HDMI 1.4a
Bus Interface
PCIe 3.0 x16
PCIe 3.0 x16
Other
Launch Price
99 USD
149 USD
Production
End-of-life
End-of-life
Predecessor
GeForce 500
GeForce 600
Successor
GeForce 700
GeForce 900
View GeForce GT 640 Details View GeForce GTX 750 Ti Details