NVIDIA GeForce GT 640 vs NVIDIA GeForce GT 730M Comparison

The NVIDIA GeForce GT 730M and GT 640 share the same GK107 chip, but they are far from identical in practice. The benchmark data shows a clear split: the GT 640 wins both available head-to-head tests, while the GT 730M trails by margins that range from moderate to significant. In Geekbench OpenCL, the GT 640 scores 3736 against the GT 730M’s 3107, a 16.8% deficit for the mobile part. The Vulkan gap is narrower but still favors the desktop card: 3809 versus 3524, a 7.5% difference. These results place the GT 730M at an average benchmark score of 3316, while the GT 640 sits at 3210 — yet the head-to-head numbers tell the real story, as the GT 640’s aggregate is pulled down by its Metal score of 2086, a test the GT 730M does not have listed.

Head-to-Head Benchmarks

The OpenCL result is the most decisive data point between these two cards. The GT 640 delivers 3736 points, which is 629 points higher than the GT 730M’s 3107. That 16.8% delta is substantial for two GPUs built on the same silicon, and it reflects a consistent throughput advantage in compute-heavy workloads. The GT 730M’s nearest rival list includes the GT 640 with a deltaPct of 3.3, meaning the GT 730M’s average score is actually 3.3% higher than the GT 640’s average — but that is an artifact of the GT 640’s Metal result dragging down its aggregate. When both cards run the same OpenCL test, the GT 640 wins outright.

In Vulkan, the gap shrinks but the outcome does not change. The GT 640 scores 3809, while the GT 730M posts 3524. That 7.5% lead for the GT 640 is more modest, but still consistent. For a mobile GPU, the GT 730M’s Vulkan showing is respectable — it beats its own OpenCL score by over 400 points — but it cannot overcome the GT 640’s higher clock-driven performance. The GT 640 also has a Geekbench Metal score of 2086, which is far below its other results; this suggests the card is significantly weaker in Apple-oriented compute APIs, but since the GT 730M has no Metal result, it cannot be compared directly.

The win count is unambiguous: the GT 640 takes both head-to-head tests, giving it a 2-0 record. The GT 730M has zero wins. If you are choosing purely on benchmark performance, the GT 640 is the faster card in every measured scenario. The margins are not trivial either — a 16.8% OpenCL lead and a 7.5% Vulkan lead are enough to feel in real-world tasks like GPU-accelerated encoding or compute offload.

Architecture Differences

Both GPUs are built on NVIDIA’s Kepler architecture using the GK107 chip, fabricated on a 28 nm process at TSMC. Transistor count is identical at 1,270 million, and die size is the same at 118 mm², yielding a transistor density of 10.8M per mm². This is the same physical silicon, but the two cards are configured and clocked differently, which explains the performance gap.

The GT 730M has a base clock of 725 MHz with no boost, meaning it runs at a fixed frequency. The GT 640’s clock speeds are not listed in the data, but the performance figures imply a higher operating frequency — its pixel rate of 7.216 GPixel/s and texture rate of 28.86 GTexel/s are both notably higher than the GT 730M’s 5.800 GPixel/s and 23.20 GTexel/s. The FP32 throughput tells the same story: the GT 640 delivers 692.7 GFLOPS versus the GT 730M’s 556.8 GFLOPS, a 24% advantage. Since both cards have identical shading units (384), TMUs (32), and ROPs (16), the entire difference comes down to clock speed and memory timing.

Memory configurations are nearly the same but not identical. Both use 2 GB of DDR3 on a 128-bit bus. The GT 730M runs its memory at 900 MHz (1800 Mbps effective), yielding 28.80 GB/s of bandwidth. The GT 640 runs at 891 MHz (1782 Mbps effective), giving 28.51 GB/s — a hair less. So the GT 640 wins on compute rates but loses slightly on memory bandwidth, a tradeoff that favors the desktop card in most tasks.

The TDP difference is stark: the GT 730M draws 33 W, while the GT 640 draws 65 W. That is nearly double the power budget, which explains why the GT 640 can sustain higher clocks. The GT 730M is an MXM module with no power connectors, designed for laptops. The GT 640 is a single-slot desktop card, 145 mm long, also with no power connectors, but it lists a suggested PSU of 250 W. Both are end-of-life products, with the GT 730M released in January 2013 and the GT 640 in June 2012.

Where Each One Wins

The GT 640 wins in raw compute performance, period. Its OpenCL score is 16.8% higher, and its Vulkan score is 7.5% higher. If you are running GPU-accelerated tasks that leverage OpenCL or Vulkan — think rendering, physics simulation, or compute shaders — the GT 640 is the clear choice. Its higher pixel rate (7.216 vs 5.800 GPixel/s) and texture rate (28.86 vs 23.20 GTexel/s) also make it better suited for fill-rate-bound workloads like high-resolution texture filtering or basic 3D rendering. The FP32 throughput of 692.7 GFLOPS versus 556.8 GFLOPS further cements its lead in general-purpose compute.

The GT 730M’s wins are more situational. It has a slightly higher memory bandwidth at 28.80 GB/s versus 28.51 GB/s, which could give it a marginal edge in memory-bound scenarios, but the difference is less than 1% and unlikely to matter in practice. Its real advantage is power efficiency: at 33 W TDP, it draws less than half the power of the GT 640. For a laptop user who needs a discrete GPU without killing battery life, the GT 730M is the only option that fits. The GT 640’s 65 W TDP and 250 W suggested PSU make it a desktop-only part. The GT 730M also runs at a fixed 725 MHz with no boost, which is simpler and more predictable for thermal management in a thin chassis.

In terms of API support, both cards are identical: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. There is no feature split there. The GT 640 has a Metal benchmark score of 2086, but the GT 730M has none, so no conclusion can be drawn about Metal performance. The GT 640’s display outputs — 1x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 — are fixed desktop connectors, while the GT 730M’s outputs are portable-device dependent, meaning they vary by laptop design.

Specification Differences

| Specification | GT 730M | GT 640 |

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

| Generation | GeForce 700M | GeForce 600 |

| Base Clock | 725 MHz | Not listed |

| Boost Clock | 725 MHz | Not listed |

| Memory Clock | 900 MHz (1800 Mbps) | 891 MHz (1782 Mbps) |

| Memory Bandwidth | 28.80 GB/s | 28.51 GB/s |

| Pixel Rate | 5.800 GPixel/s | 7.216 GPixel/s |

| Texture Rate | 23.20 GTexel/s | 28.86 GTexel/s |

| FP32 | 556.8 GFLOPS | 692.7 GFLOPS |

| TDP | 33 W | 65 W |

| Slot Width | MXM Module | Single-slot |

| Suggested PSU | Not listed | 250 W |

| Length | Not listed | 145 mm (5.7 inches) |

| Display Outputs | Portable Device Dependent | 1x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 |

| Release Date | 2013-01-19 | 2012-06-04 |

| Predecessor | GeForce 600M | GeForce 500 |

| Successor | GeForce 800M | GeForce 700 |

| Launch MSRP | Not listed | 99 USD |

Everything else — chip, architecture, process node, foundry, transistors, die size, transistor density, shading units, TMUs, ROPs, memory size, memory type, bus width, directX, OpenGL, Vulkan, production status, and power connectors — is identical between the two.

FAQ

Q: Which card is faster in compute workloads?

A: The GT 640 is faster in every measured benchmark. It scores 3736 in OpenCL versus the GT 730M’s 3107, and 3809 in Vulkan versus 3524. That is a 16.8% lead in OpenCL and a 7.5% lead in Vulkan.

Q: Do both cards have the same GPU core?

A: Yes, both use the GK107 chip with 384 shading units, 32 TMUs, and 16 ROPs, built on a 28 nm TSMC process with 1,270 million transistors on a 118 mm² die.

Q: Is the GT 730M more power efficient?

A: Yes, significantly. The GT 730M has a TDP of 33 W, while the GT 640 draws 65 W. The GT 730M is also an MXM module with no power connectors, while the GT 640 is a single-slot desktop card with no power connectors but a 250 W suggested PSU.

Q: Why is the GT 640 faster if the GT 730M has higher memory bandwidth?

A: The GT 640’s compute rates are much higher — 7.216 GPixel/s pixel rate and 28.86 GTexel/s texture rate versus the GT 730M’s 5.800 and 23.20. Its FP32 throughput is 692.7 GFLOPS versus 556.8. The GT 730M’s memory bandwidth advantage (28.80 vs 28.51 GB/s) is under 1% and does not compensate.

Q: Can the GT 730M be used in a desktop?

A: No, its slot width is listed as MXM Module, and its display outputs are portable-device dependent. The GT 640, by contrast, is a single-slot card with DVI, HDMI 1.4a, and DisplayPort 1.2 outputs.

Q: What is the release timeline for these cards?

A: The GT 640 was released on 2012-06-04 as part of the GeForce 600 generation, while the GT 730M came later on 2013-01-19 in the GeForce 700M generation. Both are end-of-life, with the GT 730M’s predecessor being GeForce 600M and successor GeForce 800M, while the GT 640’s predecessor is GeForce 500 and successor GeForce 700.

The Verdict

The data does not support picking the GT 730M on performance grounds. The GT 640 wins both head-to-head benchmarks, with a 16.8% lead in OpenCL and a 7.5% lead in Vulkan. Its pixel rate, texture rate, and FP32 throughput are all substantially higher, and it has a Metal benchmark result (2086) that the GT 730M lacks entirely. If you are building a desktop system and need the best Kepler-based GK107 performance, the GT 640 is the only rational choice.

The GT 730M’s case rests entirely on form factor and power. At 33 W TDP, it is designed for laptops where the GT 640’s 65 W draw and 250 W suggested PSU would be impossible. It is an MXM module with no fixed display outputs, meaning it belongs in a portable machine. Its memory bandwidth is marginally higher (28.80 vs 28.51 GB/s), but that is a rounding error in real workloads.

For a desktop user, the GT 640’s launch MSRP of 99 USD and single-slot design make it a straightforward low-profile add-in card. For a laptop user, the GT 730M is the only option that physically fits. The benchmark results are clear: the GT 640 is the faster card, and the GT 730M is the efficiency-oriented mobile variant. Choose the GT 640 if you want performance, choose the GT 730M only if your platform requires it. The 2-0 win record for the GT 640 in head-to-head tests is the final word.