NVIDIA GeForce GT 635M vs NVIDIA GeForce GT 730M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GT 635M has a higher average benchmark score of 3740, compared to the GT 730M's average of 3316. This places the GT 635M in the 22nd percentile of all GPUs, while the GT 730M sits in the 20th percentile.

Q: How do the two GPUs compare in the Geekbench OpenCL test?

A: In the Geekbench OpenCL test, the GT 635M scores 3740, while the GT 730M scores 3107. The GT 635M wins this head-to-head by a delta of 20.4%, representing a substantial performance gap in this specific workload.

Q: What are the architectural differences between the two chips?

A: The GT 635M is built on the Fermi architecture using the GF108 chip, manufactured on a 40 nm process at TSMC. The GT 730M uses the Kepler architecture with the GK107 chip, built on a 28 nm process, also at TSMC. The GT 730M has a significantly higher transistor density of 10.8M / mm² versus 5.0M / mm² for the GT 635M.

Q: Do both GPUs support the same graphics APIs?

A: Both GPUs support DirectX 12 (11_0) and OpenGL 4.6. However, the GT 730M additionally supports Vulkan 1.2.175, while the GT 635M has no Vulkan support listed in the data.

Q: What is the difference in shading units between the two?

A: The GT 730M has 384 shading units, which is four times the 96 shading units found on the GT 635M. This architectural advantage contributes to the GT 730M's higher theoretical compute throughput.

Q: Which GPU has a higher pixel rate and texture rate?

A: The GT 730M has a pixel rate of 5.800 GPixel/s and a texture rate of 23.20 GTexel/s. The GT 635M has a pixel rate of 1.900 GPixel/s and a texture rate of 7.600 GTexel/s. The GT 730M leads in both metrics by a wide margin.

---

Where Each One Wins

The benchmark data presents a clear but counterintuitive split. The GT 635M wins the only head-to-head benchmark available, the Geekbench OpenCL test, with a 20.4% advantage. This suggests that in raw compute workloads that resemble OpenCL tasks, the older Fermi-based GT 635M holds a definitive edge over the Kepler-based GT 730M.

However, the GT 730M wins in nearly every architectural specification category. It has four times the shading units (384 vs. 96), double the texture mapping units (32 vs. 16), and four times the raster output pipelines (16 vs. 4). Its pixel rate of 5.800 GPixel/s is over three times higher than the GT 635M's 1.900 GPixel/s, and its texture rate of 23.20 GTexel/s is likewise over three times higher than the 7.600 GTexel/s of the GT 635M.

The GT 730M also wins on process technology, using a 28 nm node versus the GT 635M's 40 nm node, and it supports the newer PCIe 3.0 x16 interface compared to the GT 635M's PCIe 2.0 x16. In terms of raw fill-rate and geometry throughput, the GT 730M is the clear victor. The data suggests the GT 730M is better positioned for tasks that rely on pixel and texture throughput, while the GT 635M appears optimized for compute-style workloads in this specific benchmark.

---

Architecture Differences

The two GPUs belong to different architectural generations. The GT 635M is based on Fermi, a design that predates the Kepler architecture used in the GT 730M. The chip identifiers reflect this: GF108 for the GT 635M and GK107 for the GT 730M. Both are manufactured by TSMC, but on different process nodes: 40 nm for Fermi and 28 nm for Kepler.

The transistor counts differ dramatically. The GT 730M packs 1,270 million transistors into a die size of 118 mm², achieving a transistor density of 10.8M / mm². The GT 635M has only 585 million transistors on a 116 mm² die, yielding a density of 5.0M / mm². This means the GT 730M nearly doubles the transistor density despite having a similar physical die size.

The most significant architectural divergence is in the execution resources. The GT 730M features 384 shading units, 32 TMUs, and 16 ROPs. In contrast, the GT 635M has 96 shading units, 16 TMUs, and only 4 ROPs. This gives the GT 730M a 4x advantage in shader count and ROP count, and a 2x advantage in TMUs.

Memory subsystems are identical in configuration: both use 2 GB of DDR3 with a 128-bit bus, resulting in the same 28.80 GB/s bandwidth. The memory clock is also identical at 900 MHz (1800 Mbps effective). However, the GT 730M has a base and boost clock of 725 MHz, while the GT 635M's core clock is not specified in the data.

The GT 730M also supports the newer PCIe 3.0 x16 bus interface, whereas the GT 635M is limited to PCIe 2.0 x16. In terms of API support, the GT 730M adds Vulkan 1.2.175, which the GT 635M lacks entirely. Both support DirectX 12 (11_0) and OpenGL 4.6.

---

Specification Differences

| Specification | NVIDIA GeForce GT 635M | NVIDIA GeForce GT 730M |

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

| Architecture | Fermi | Kepler |

| Chip | GF108 | GK107 |

| Process Node | 40 nm | 28 nm |

| Transistors | 585 million | 1,270 million |

| Die Size | 116 mm² | 118 mm² |

| Transistor Density | 5.0M / mm² | 10.8M / mm² |

| Base Clock | Not specified | 725 MHz |

| Boost Clock | Not specified | 725 MHz |

| Shading Units | 96 | 384 |

| TMUs | 16 | 32 |

| ROPs | 4 | 16 |

| Pixel Rate | 1.900 GPixel/s | 5.800 GPixel/s |

| Texture Rate | 7.600 GTexel/s | 23.20 GTexel/s |

| FP32 Performance | 182.4 GFLOPS | 556.8 GFLOPS |

| TDP | 35 W | 33 W |

| Slot Width | IGP | MXM Module |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Vulkan Support | None | 1.2.175 |

| Release Date | 2012-03-21 | 2013-01-19 |

| Generation | GeForce 600M | GeForce 700M |

| Predecessor | GeForce 500M | GeForce 600M |

| Successor | GeForce 700M | GeForce 800M |

The two GPUs share identical memory specifications: 2 GB DDR3, 128-bit bus, 28.80 GB/s bandwidth, and 900 MHz memory clock. Both have no power connectors, no RT cores, no tensor cores, and no FP16 support. Their display outputs are both "Portable Device Dependent," and neither has a launch MSRP listed.

---

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test. In this benchmark, the GT 635M scores 3740, while the GT 730M scores 3107. The delta is 20.4% in favor of the GT 635M. This is a decisive win for the older Fermi chip, which outperforms the newer Kepler chip by a significant margin in this compute-oriented test.

Looking at the nearest rivals for each GPU provides additional context. The GT 635M's closest competitor is the NVIDIA Quadro 3000M, which scores 3718, a delta of just 0.6% behind. The GT 740M scores 3717, also 0.6% behind, and the GeForce 825M scores 3694, 1.2% behind. The Intel UHD Graphics 710 scores 3792, which is 1.4% ahead of the GT 635M. This places the GT 635M in a tight performance cluster around the 3700-3800 score range.

The GT 730M's nearest rivals paint a different picture. Its closest competitor is the Intel HD Graphics 530, which scores 3332, a delta of -0.5% (meaning the GT 730M is slightly behind). The GeForce 920M scores 3287, which is 0.9% behind the GT 730M. The Intel HD Graphics P4600 scores 3389, 2.2% ahead, and the GeForce GT 640 scores 3210, 3.3% behind. The GT 730M's average benchmark score of 3316 is notably lower than the GT 635M's average of 3740.

The GT 730M does have a Vulkan benchmark score of 3524, but no comparable Vulkan score exists for the GT 635M, so a direct comparison in that API is not possible from the data. Similarly, the GT 635M has no Vulkan support at all, making this a moot point for cross-GPU comparison.

---

The Verdict

The data presents a clear verdict despite the architectural differences. The NVIDIA GeForce GT 635M wins the only head-to-head benchmark available, scoring 3740 versus the GT 730M's 3107 in Geekbench OpenCL, a 20.4% advantage. Anyone prioritizing compute performance in OpenCL workloads should choose the GT 635M based on this result.

However, the GT 730M is not without merits. Its Kepler architecture brings a 4x increase in shading units (384 vs. 96), a 2x increase in TMUs (32 vs. 16), and a 4x increase in ROPs (16 vs. 4). Its pixel rate of 5.800 GPixel/s and texture rate of 23.20 GTexel/s are roughly three times higher than the GT 635M's respective figures of 1.900 GPixel/s and 7.600 GTexel/s. The GT 730M also has significantly higher FP32 performance at 556.8 GFLOPS versus 182.4 GFLOPS for the GT 635M.

The GT 730M is built on a more advanced 28 nm process with a higher transistor density of 10.8M / mm², and it supports Vulkan 1.2.175, which the GT 635M does not. It also uses the newer PCIe 3.0 x16 interface and has a slightly lower TDP of 33 W versus 35 W.

For users whose workloads depend on fill-rate, texture processing, or modern API support, the GT 730M's specification sheet is the stronger option. For users running OpenCL compute tasks, the GT 635M's benchmark result is the deciding factor. The data shows a nuanced picture: the GT 635M wins the compute benchmark, while the GT 730M wins on architectural specifications. The choice ultimately depends on whether the workload is compute-bound or graphics-bound, with the benchmark data favoring the GT 635M in compute and the specifications favoring the GT 730M in graphics throughput.