NVIDIA GeForce GTS 450 vs NVIDIA GeForce GTX 670MX Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 670MX leads with an average score of 5721, while the NVIDIA GeForce GTS 450 scores 4893. That puts the GTX 670MX in the 33rd percentile of all GPUs versus the GTS 450's 28th percentile.

Q: How do the two compare in the only shared benchmark test?

A: In Geekbench OpenCL, the GTX 670MX scores 6125 versus the GTS 450's 4893, a 20.1% difference in favor of the GTX 670MX. This is the sole head-to-head benchmark recorded in the database.

Q: What are the nearest rivals to each card, and how close are they?

A: The GTS 450's closest rival is the AMD Radeon R7 M265 at 4929 (0.7% ahead), followed by the AMD FirePro W5130M at 4904 (0.2% ahead). The GTX 670MX's nearest rival is the NVIDIA Quadro M500M at 5604, which is 2.1% behind it, while the Intel Iris Pro Graphics P6300 sits only 0.2% behind at 5712.

Q: Which card has the higher transistor count and density?

A: The GTX 670MX packs 3,540 million transistors on a 294 mm² die, giving a density of 12.0M per mm². The GTS 450 has 1,170 million transistors on a 238 mm² die, yielding just 4.9M per mm².

Q: Do these cards support different API levels?

A: Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 670MX additionally supports Vulkan 1.2.175, while the GTS 450 has no Vulkan support listed.

Q: What is the power draw difference between the two?

A: The GTX 670MX is rated at 75 W TDP with no power connectors, while the GTS 450 draws 106 W and requires a single 6-pin connector. The GTS 450 also lists a suggested PSU of 300 W.

Architecture Differences

The two GPUs represent two distinct NVIDIA architectures. The GTS 450 is built on Fermi, the architecture that powered the GeForce 400 generation, while the GTX 670MX uses Kepler, from the GeForce 600M series. This architectural shift is fundamental to nearly every performance and efficiency difference between them.

The manufacturing process tells part of the story. The GTS 450 uses a 40 nm process from TSMC, while the GTX 670MX uses a newer 28 nm process, also from TSMC. This node shrink is a major reason the GTX 670MX achieves far higher transistor density: 12.0M per mm² versus 4.9M per mm². The GTX 670MX crams 3,540 million transistors onto a 294 mm² die, while the GTS 450 fits only 1,170 million onto a 238 mm² die. That is three times the transistor count on a die only about 24% larger.

The compute resources differ dramatically. The GTX 670MX has 960 shading units, 80 texture mapping units, and 24 raster output units. The GTS 450, by contrast, has 192 shading units, 32 TMUs, and 16 ROPs. In raw throughput terms, the GTX 670MX delivers 1,153.9 GFLOPS of FP32 compute versus 601.3 GFLOPS for the GTS 450. Pixel fill rate is 12.02 GPixel/s versus 6.264 GPixel/s, and texture fill rate is 48.08 GTexel/s versus 25.06 GTexel/s.

Memory architecture also diverges. The GTS 450 has 1024 MB of GDDR5 on a 128-bit bus, with bandwidth of 57.73 GB/s. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus, reaching 67.20 GB/s. Clock speeds differ as well: the GTS 450's memory runs at 902 MHz (3.6 Gbps effective), while the GTX 670MX's memory runs at 700 MHz (2.8 Gbps effective), yet the wider bus more than compensates.

The bus interface and power delivery also differ. The GTS 450 uses PCIe 2.0 x16 and requires a 6-pin power connector, while the GTX 670MX uses PCIe 3.0 x16 and needs no external power connector. The GTX 670MX is also more power-efficient despite its larger feature set, with a 75 W TDP versus 106 W for the GTS 450. Both are end-of-life products, but the GTX 670MX supports Vulkan 1.2.175 in addition to the shared DirectX 12 (11_0) and OpenGL 4.6 APIs.

Head-to-Head Benchmarks

The database records exactly one shared benchmark test between these two cards: Geekbench OpenCL. The result is decisive. The GTX 670MX scores 6125, while the GTS 450 manages 4893. That is a 20.1% margin in favor of the GTX 670MX, which is a substantial gap for a single compute workload.

This 20.1% lead in OpenCL is consistent with the underlying hardware differences. The GTX 670MX has five times the shading units (960 versus 192), roughly double the FP32 throughput (1,153.9 GFLOPS versus 601.3 GFLOPS), and nearly double the pixel and texture rates. The wider 192-bit memory bus and larger 3 GB frame buffer also give it a bandwidth advantage of 67.20 GB/s versus 57.73 GB/s, which helps in memory-heavy compute tasks.

The GTX 670MX also has a second benchmark result in the database: a Geekbench Vulkan score of 5316. The GTS 450 has no Vulkan score listed, and since it lacks Vulkan support entirely, this is not a head-to-head comparison. Still, the Vulkan result shows the GTX 670MX is competitive in modern API workloads, posting a score that is only about 13% below its own OpenCL score.

In terms of overall standing, the GTX 670MX averages 5721 across its two benchmark results, placing it at the 33rd percentile of all GPUs. The GTS 450's single score of 4893 places it at the 28th percentile. The GTX 670MX thus sits five percentile points higher, a modest but real separation in the overall distribution.

When placed against their nearest rivals, the two cards occupy similar competitive tiers. The GTS 450 trades blows with the AMD Radeon R7 M265 (0.7% behind the GTS 450) and the AMD FirePro W5130M (0.2% behind the GTS 450), while also sitting 0.2% behind the NVIDIA GeForce RTX 5060 Ti 8 GB. The GTX 670MX is 0.2% ahead of the Intel Iris Pro Graphics P6300 and 0.5% ahead of the AMD Radeon HD 8790M, while trailing the NVIDIA GeForce GTX 550 Ti by 0.2%. Both cards are thus clustered tightly among their peers, but the GTX 670MX's cluster sits at a higher absolute performance level.

Specification Differences

The two cards diverge on nearly every major specification. Here are the fields where they differ:

• Architecture: Fermi (GTS 450) versus Kepler (GTX 670MX)
• Generation: GeForce 400 versus GeForce 600M
• Process node: 40 nm versus 28 nm
• Transistors: 1,170 million versus 3,540 million
• Die size: 238 mm² versus 294 mm²
• Transistor density: 4.9M / mm² versus 12.0M / mm²
• Base clock: not listed versus 601 MHz
• Boost clock: not listed versus 601 MHz
• Memory clock: 902 MHz (3.6 Gbps effective) versus 700 MHz (2.8 Gbps effective)
• Memory size: 1024 MB versus 3 GB
• Memory bus width: 128 bit versus 192 bit
• Memory bandwidth: 57.73 GB/s versus 67.20 GB/s
• Shading units: 192 versus 960
• TMUs: 32 versus 80
• ROPs: 16 versus 24
• Pixel rate: 6.264 GPixel/s versus 12.02 GPixel/s
• Texture rate: 25.06 GTexel/s versus 48.08 GTexel/s
• FP32: 601.3 GFLOPS versus 1,153.9 GFLOPS
• TDP: 106 W versus 75 W
• Slot width: Dual-slot versus not listed
• Power connectors: 1x 6-pin versus none
• Suggested PSU: 300 W versus not listed
• Bus interface: PCIe 2.0 x16 versus PCIe 3.0 x16
• Display outputs: 2x DVI, 1x mini-HDMI 1.3a versus portable device dependent
• Vulkan support: not listed versus 1.2.175
• Dimensions: 210 mm length, 111 mm height versus not listed
• Release date: 2010-09-12 versus 2012-09-30
• Predecessor: GeForce 200 versus GeForce 500M
• Successor: GeForce 500 versus GeForce 700M
• Launch MSRP: 129 USD versus not listed

Fields that are the same include manufacturer (NVIDIA), foundry (TSMC), memory type (GDDR5), DirectX support (12 with 11_0 feature level), OpenGL support (4.6), and production status (end-of-life).

The Verdict

The data points to a clear winner for raw performance: the NVIDIA GeForce GTX 670MX. It beats the GTS 450 by 20.1% in Geekbench OpenCL, has nearly double the FP32 compute throughput, and delivers higher pixel and texture rates across the board. Its 33rd percentile standing versus the GTS 450's 28th percentile confirms the hierarchy.

But the GTX 670MX is not just faster; it is also more efficient. It achieves all of this at a 75 W TDP with no external power connector, whereas the GTS 450 needs 106 W and a 6-pin connector. The GTX 670MX also brings 3 GB of memory versus 1 GB, a wider 192-bit bus, PCIe 3.0 support, and Vulkan compatibility. The GTS 450 offers no compensating advantages except a lower launch MSRP of 129 USD and a dual-slot desktop form factor with fixed display outputs.

For a desktop user in 2010, the GTS 450 was a reasonable entry point. For anyone comparing these two today, the GTX 670MX is the superior part in every measured category except price, and it even has a lower power draw. The GTS 450's only edge is that it was priced at 129 USD at launch, but the GTX 670MX has no recorded launch MSRP to compare against.

The verdict is straightforward: the GTX 670MX wins the performance contest decisively, and it does so while consuming 31 W less power. The GTS 450 remains a historical footnote from the Fermi era, outclassed by its Kepler successor in compute, memory, and efficiency.

Where Each One Wins

NVIDIA GeForce GTX 670MX wins on:

• Compute performance: 20.1% ahead in Geekbench OpenCL, with 1,153.9 GFLOPS versus 601.3 GFLOPS FP32.
• Memory capacity and bandwidth: 3 GB versus 1 GB, with 67.20 GB/s versus 57.73 GB/s over a wider 192-bit bus.
• Pixel and texture throughput: 12.02 GPixel/s versus 6.264 GPixel/s, and 48.08 GTexel/s versus 25.06 GTexel/s.
• Power efficiency: 75 W TDP with no power connector versus 106 W with a 6-pin connector.
• Modern API support: Vulkan 1.2.175 support, which the GTS 450 lacks entirely.
• System integration: PCIe 3.0 x16 versus PCIe 2.0 x16, and a portable-device-friendly design with no fixed display outputs.
• Overall standing: 33rd percentile versus 28th percentile, and a higher average benchmark score of 5721 versus 4893.

NVIDIA GeForce GTS 450 wins on:

• Launch price: A recorded launch MSRP of 129 USD, which the GTX 670MX does not have.
• Desktop form factor: Dual-slot design with a fixed 210 mm length and 111 mm height, plus 2x DVI and 1x mini-HDMI 1.3a outputs, making it a drop-in card for desktop systems.
• Nearest rival cushion: The GTS 450 is 0.7% ahead of its closest rival (AMD Radeon R7 M265), while the GTX 670MX is only 0.2% ahead of its closest rival (Intel Iris Pro Graphics P6300). This means the GTS 450 has a slightly larger buffer over its immediate competitor, though at a lower absolute level.

In practical terms, the GTX 670MX is the pick for anyone prioritizing performance, efficiency, or modern feature support. The GTS 450 is only preferable for a legacy desktop build where the fixed form factor and 129 USD launch pricing matter more than compute capability. The benchmark data shows no scenario where the GTS 450 outperforms the GTX 670MX in raw speed; its wins are purely contextual, not performance-based.