NVIDIA GeForce GT 735M vs NVIDIA GeForce GTX 650 Comparison

The NVIDIA GeForce GTX 650 and the NVIDIA GeForce GT 735M are both end-of-life Kepler parts, but they serve very different purposes. The GTX 650 is a desktop add-in card built on the GK106 chip, while the GT 735M is a mobile IGP (integrated graphics processor) on the smaller GK208 die. Benchmark data shows the GTX 650 holds a single, decisive win in the one head-to-head test, but the GT 735M’s low power draw and different memory configuration make it a distinct alternative for specific system builds.

Where Each One Wins

The data is clear-cut: the GTX 650 wins the only head-to-head benchmark available. In the Geekbench OpenCL test, the GTX 650 scores 4545 against the GT 735M’s 3616, a 25.7% advantage. This makes the GTX 650 the choice for any workload that leverages raw compute throughput, such as OpenCL-accelerated encoding or general-purpose GPU tasks. Its higher pixel rate of 8.464 GPixel/s versus 5.024 GPixel/s and texture rate of 33.86 GTexel/s versus 20.10 GTexel/s also suggest it handles fill-rate-bound scenarios more effectively.

The GT 735M wins on system integration and power efficiency. Its TDP is 33 W, roughly half the GTX 650’s 65 W, and it requires no power connectors, drawing everything from the PCIe slot. It is listed as an IGP, meaning it is designed to be soldered onto a laptop motherboard or a compact board, where the GTX 650’s 147 mm length and single-slot cooler with a 1x 6-pin connector would be impractical. For a portable or low-profile system, the GT 735M is the only realistic option, despite its lower benchmark scores.

The GTX 650 also offers better API support parity with the GT 735M—both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175—so there is no feature gap there. The GTX 650’s 1024 MB of GDDR5 memory provides 80.00 GB/s of bandwidth, while the GT 735M’s 2 GB of DDR3 tops out at just 14.40 GB/s. That bandwidth difference is why the GTX 650 dominates in compute benchmarks, even though the GT 735M has double the memory capacity.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The GTX 650 has an average benchmark score of 3823 across three tests (Geekbench Metal, OpenCL, and Vulkan), while the GT 735M has an average of 3616 from a single Geekbench OpenCL run. The GTX 650 also sits at the 22nd percentile of all GPUs, compared to the GT 735M’s 21st percentile.

Q: Can the GT 735M match the GTX 650 in compute performance?

A: No. In the Geekbench OpenCL head-to-head, the GTX 650 scores 4545 versus the GT 735M’s 3616, a 25.7% difference. The GTX 650’s FP32 throughput is 812.5 GFLOPS, while the GT 735M manages only 482.3 GFLOPS.

Q: Is the GT 735M more power-efficient than the GTX 650?

A: Yes. The GT 735M has a 33 W TDP and requires no external power connectors, whereas the GTX 650 has a 65 W TDP and needs a 1x 6-pin power connector, with a suggested PSU of 250 W.

Q: Do both GPUs support the same modern APIs?

A: Yes. Both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. There is no difference in API feature levels between the two.

Q: Which GPU has more memory bandwidth?

A: The GTX 650 has significantly more. It uses a 128-bit bus with GDDR5 memory at 80.00 GB/s, while the GT 735M uses a 64-bit bus with DDR3 at 14.40 GB/s. The GT 735M does have 2 GB of memory versus 1 GB on the GTX 650.

Q: Are these GPUs still in production?

A: No. Both are marked as end-of-life. The GTX 650 was released later, on 2013-11-26, while the GT 735M came out earlier on 2013-03-31.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, and it is a decisive win for the GTX 650. The score difference is 4545 to 3616, a 25.7% gap. That is not a marginal lead; it is a substantial performance advantage in a compute-heavy workload. To put it in context, the GTX 650’s nearest rival is the NVIDIA GeForce MX110, which scores 3834 (a 0.3% difference), and the Intel UHD Graphics 710 at 3792 (0.8% above the GTX 650). The GTX 650 sits comfortably above those integrated and entry-level parts, but its OpenCL score is nearly 1.3 times higher than the GT 735M’s.

The GT 735M’s 3616 OpenCL score places it in a different performance tier. Its nearest rivals include the NVIDIA GeForce GTX 1050 at 3629 (a 0.3% difference) and the NVIDIA RTX 5000 Mobile Ada Generation at 3596 (0.6% above the GT 735M). This is notable: the GT 735M’s compute performance is statistically similar to a GTX 1050 in this specific test, despite the GTX 1050 being a much newer and more powerful desktop card. That suggests the GT 735M, while slower than the GTX 650, is not a slouch for its class—it just lacks the memory bandwidth and raw throughput to compete with the GTX 650.

There is no head-to-head data for Geekbench Metal or Vulkan, so the GTX 650’s scores in those tests (2400 and 4524, respectively) cannot be compared directly to the GT 735M. However, the GTX 650’s average benchmark score of 3823, pulled from all three tests, is still 5.7% higher than the GT 735M’s single-test average of 3616.

Specification Differences

The two GPUs diverge sharply on memory and physical design. The GTX 650 uses 1024 MB of GDDR5 across a 128-bit bus, yielding 80.00 GB/s of bandwidth. The GT 735M uses 2 GB of DDR3 on a 64-bit bus, yielding just 14.40 GB/s. That is a 5.5x bandwidth advantage for the GTX 650, which explains its compute dominance. The GT 735M does have double the memory capacity, but the type and bus width severely limit its effective throughput.

Clock speeds also differ. The GT 735M has a base clock of 575 MHz and a boost clock of 628 MHz; the GTX 650’s base and boost clocks are not listed. Memory clocks are listed differently as well: the GTX 650 runs at 1250 MHz with 5 Gbps effective, while the GT 735M runs at 900 MHz with 1800 Mbps effective. The GTX 650 also has more ROPs (16 versus 8) and a higher pixel rate (8.464 GPixel/s versus 5.024 GPixel/s), plus a higher texture rate (33.86 GTexel/s versus 20.10 GTexel/s). Both have 384 shading units and 32 TMUs, so the core count is identical, but the GTX 650’s higher clocks and memory system give it far more output.

Physical requirements are polar opposites. The GTX 650 is a single-slot, 147 mm (5.8 inches) card that needs a 1x 6-pin power connector and a 250 W PSU. The GT 735M is an IGP with no specified dimensions, no power connectors, and a 33 W TDP. The GTX 650 uses a PCIe 3.0 x16 interface, while the GT 735M uses PCIe 3.0 x8. Display outputs also differ: the GTX 650 has 1x DVI and 2x DisplayPort 1.2, whereas the GT 735M’s outputs are listed as “Portable Device Dependent.”

Architecture Differences

Both GPUs are built on TSMC’s 28 nm process, but they use different chips. The GTX 650 uses the GK106 chip, part of the original Kepler architecture, while the GT 735M uses the GK208 chip, labeled as Kepler 2.0. The die sizes reflect a major difference: GK106 measures 221 mm² with 2,540 million transistors, while GK208 is just 87 mm² with 1,020 million transistors. Transistor density is nearly identical (11.5M/mm² versus 11.7M/mm²), but the GTX 650’s larger die houses more functional units, including 16 ROPs versus 8.

The GTX 650 is part of the GeForce 600 generation, while the GT 735M belongs to the GeForce 700M generation. Their predecessors and successors also differ: the GTX 650 follows the GeForce 500 series and leads to GeForce 700, while the GT 735M follows GeForce 600M and leads to GeForce 800M. Despite the architectural naming difference (Kepler vs. Kepler 2.0), both support identical API versions, so there is no software feature advantage on either side. The GT 735M’s lower transistor count and smaller die make it a power-optimized part, but that comes at the cost of raw compute resources.

The Verdict

Pick the GTX 650 if you need compute performance. The data shows a 25.7% lead in OpenCL, backed by 5.5x the memory bandwidth and higher pixel/texture rates. It is the clear choice for any desktop system with a 250 W PSU and room for a 147 mm single-slot card. The GTX 650’s 22nd percentile ranking and average score of 3823 put it in the same league as the MX110 and Radeon R5 Graphics, making it a viable entry-level discrete GPU for light gaming or OpenCL tasks.

Pick the GT 735M if you are building or upgrading a compact or mobile system where power and space are the constraints. Its 33 W TDP, lack of power connectors, and IGP form factor mean it can fit where the GTX 650 cannot. Its 2 GB of memory is also useful for buffer-heavy workloads, even if the 64-bit DDR3 interface limits actual throughput. The GT 735M’s 21st percentile ranking and OpenCL score of 3616 show it is not far behind the GTX 650 in overall standing, but it trades away most of the performance headroom for integration ease.

There is no scenario where the GT 735M beats the GTX 650 in a benchmark, given the available data. The GTX 650 wins the only head-to-head test, and its specifications are uniformly superior for compute and fill-rate tasks. The GT 735M’s only advantages are its lower power draw, smaller footprint, and doubled memory capacity—none of which show up in benchmark scores. For a fixed desktop build, the GTX 650 is the better performer. For a portable or low-power system, the GT 735M is the only one that fits.