NVIDIA GeForce GTX 465 vs NVIDIA GeForce GTX 675MX Comparison

The NVIDIA GeForce GTX 465 and the NVIDIA GeForce GTX 675MX are two End-of-life NVIDIA GPUs from different eras and different segments: one a desktop card from the GeForce 400 generation, the other a mobile MXM module from the GeForce 600M line. The database places them surprisingly close in overall standing, with the GTX 465 at the 46th percentile versus all GPUs and the GTX 675MX at the 43rd, yet the recorded benchmark data favors the newer mobile part in compute. What follows is a step-by-step walk through where each card wins, how their architectures differ, and what the head-to-head numbers say.

Where Each One Wins

The head-to-head record is lopsided but narrow in scope. There is only one shared benchmark between the two in the database, geekbench_opencl, and the GTX 675MX takes it: 10,723 points against 9,294, a 13.3 percent lead. That makes the wins tally 1 for the GTX 675MX and 0 for the GTX 465 in direct comparison.

Outside the head-to-head, the GTX 465 does record one benchmark of its own, the same geekbench_opencl score of 9,294, which also serves as its average benchmark score. The GTX 675MX carries two entries: 6,131 in geekbench_metal and 10,723 in geekbench_opencl, averaging to 8,427. Note that the averages are not directly comparable since the GTX 675MX's average is dragged down by the metal test, a workload the older card has no entry for.

So where does each one win? Strictly by the data, the GTX 675MX wins compute. The GTX 465 wins nowhere in recorded benchmarks, but it holds structural advantages that benchmarks here do not capture: it is a desktop card with fixed display outputs (2x DVI, 1x mini-HDMI 1.3a) versus the GTX 675MX's "Portable Device Dependent" outputs, and it carries a formally suggested power supply rating rather than relying on a host laptop's design.

Architecture Differences

These two cards sit two architecture generations apart. The GTX 465 is built on Fermi, the GF100 chip, fabricated at TSMC on a 40 nm process. The GTX 675MX uses Kepler, the GK104 chip, also from TSMC but on 28 nm. That node jump shows up clearly in the density figures: the GTX 465 packs 3,100 million transistors into 529 mm², a density of 5.9M per mm², while the GTX 675MX fits 3,540 million transistors into just 294 mm², more than double the density at 12.0M per mm².

The shader counts diverge sharply. The GTX 465 has 352 shading units, 44 texture mapping units, and 32 render output units. The GTX 675MX nearly triples the shading units at 960 and nearly doubles TMUs at 80, while matching the ROP count at 32. That shader advantage is why the raw FP32 compute favors the Kepler part: 1,255.7 GFLOPS against 855.4 GFLOPS.

Texture throughput follows the same pattern, 52.32 GTexel/s for the GTX 675MX versus 26.75 GTexel/s for the GTX 465, while pixel fill rate actually tilts slightly the other way, 13.38 GPixel/s for the Fermi card against 13.08 GPixel/s for the Kepler module. Both support DirectX 12 with feature level 11_0 and OpenGL 4.6, but only the GTX 675MX lists Vulkan support, at version 1.2.175; the GTX 465 has no Vulkan entry.

Head-to-Head Benchmarks

The single shared test is geekbench_opencl, and it goes to the GTX 675MX by a clear margin: 10,723 versus 9,294, a 13.3 percent gap in the Kepler card's favor. Interpreted against each card's rival set, this is consistent with the database's broader picture. The GTX 465's opencl score sits within a fraction of a percent of cards like the GeForce GTX 850M (9,302, 0.1 percent ahead of it), the Radeon R7 M380 (9,313, 0.2 percent ahead), and the GeForce GTX 960 (9,273, 0.2 percent behind it), with the Radeon Vega 8 trailing at 9,221.

The GTX 675MX's rival set clusters around its 8,427 average rather than its opencl score: the Radeon 880M at 8,436 (0.1 percent ahead), the GeForce MX330 at 8,458 (0.4 percent ahead), the Radeon HD 8870M at 8,462 (0.4 percent ahead), and the Radeon R9 M375X at 8,325 (1.2 percent behind). In other words, by average standing the GTX 675MX ranks marginally below all four of its nearest rivals except the R9 M375X, while the GTX 465 ranks marginally above three of its four. The percentile fields frame this the same way: 46th versus 43rd.

The practical reading: in OpenCL compute the newer mobile chip is decisively ahead of the older desktop card, but in overall database standing the two are separated by only a few percentile points, and each sits in a tight cluster of similarly performing hardware.

Specification Differences

The cards differ on nearly every specification that matters. Process node: 40 nm versus 28 nm. Transistor count: 3,100 million versus 3,540 million. Die size: 529 mm² versus 294 mm². Transistor density: 5.9M per mm² versus 12.0M per mm².

Memory favors the GTX 675MX in capacity and speed: 2 GB of GDDR5 versus 1024 MB, both on a 256-bit bus, with bandwidth of 115.2 GB/s versus 102.7 GB/s. Memory clock is 900 MHz (3.6 Gbps effective) against 802 MHz (3.2 Gbps effective).

Shader resources: 960 shading units and 80 TMUs versus 352 and 44, with ROPs equal at 32. FP32 compute: 1,255.7 GFLOPS versus 855.4 GFLOPS. Texture rate: 52.32 GTexel/s versus 26.75 GTexel/s. Pixel rate: 13.08 GPixel/s versus 13.38 GPixel/s.

Power and form factor are the biggest divides. The GTX 465 is a 200 W dual-slot desktop card requiring 2x 6-pin PCIe power connectors and a suggested 550 W power supply, on a PCIe 2.0 x16 interface. The GTX 675MX is a 100 W MXM module with no power connectors, using an MXM-B (3.0) interface, intended for laptops. The desktop card measures 241 mm long; the module's dimensions are not recorded.

API support: identical DirectX 12 (11_0) and OpenGL 4.6 on both, Vulkan 1.2.175 only on the GTX 675MX. Release dates: 2010 for the GTX 465 (with a launch MSRP of 279 USD) and 2012 for the GTX 675MX, which has no recorded launch MSRP. Both are End-of-life, with the GTX 465 sitting between the GeForce 200 and GeForce 500 generations, and the GTX 675MX between the GeForce 500M and GeForce 700M.

FAQ

Q: Which card wins the shared benchmark? A: The GTX 675MX wins geekbench_opencl with 10,723 points to the GTX 465's 9,294, a 13.3 percent advantage.

Q: How much more compute does the GTX 675MX have? A: Its FP32 throughput is 1,255.7 GFLOPS versus 855.4 GFLOPS for the GTX 465, reflecting 960 shading units against 352.

Q: Which has more memory? A: The GTX 675MX, with 2 GB of GDDR5 and 115.2 GB/s of bandwidth, versus 1024 MB and 102.7 GB/s for the GTX 465.

Q: Does either card support Vulkan? A: Only the GTX 675MX, at version 1.2.175. The GTX 465 has no Vulkan entry in the database. Both support DirectX 12 (11_0) and OpenGL 4.6.

Q: How do their power requirements compare? A: The GTX 465 is rated at 200 W, needs 2x 6-pin connectors, and carries a suggested 550 W power supply. The GTX 675MX is a 100 W MXM module with no power connectors.

Q: Where does each rank against all GPUs? A: The GTX 465 sits at the 46th percentile and the GTX 675MX at the 43rd, a three-point gap in the older card's favor.

The Verdict

In raw compute, the data favors the GTX 675MX: a 13.3 percent OpenCL win, roughly 47 percent higher FP32 throughput, nearly double the texture rate, twice the memory capacity, and higher bandwidth, all at half the power draw. Anyone choosing between them purely on recorded performance should take the Kepler module.

The counterweights are context, not benchmarks. The GTX 465 is a desktop card with fixed outputs (2x DVI, 1x mini-HDMI 1.3a), a PCIe 2.0 x16 interface, and a slightly higher database standing at the 46th percentile versus 43rd, plus a marginal edge in pixel fill rate (13.38 versus 13.08 GPixel/s). The GTX 675MX exists only as an MXM-B (3.0) module inside portable devices with device-dependent display options. So the realistic split is simple: for compute throughput and efficiency per watt, the GTX 675MX; for a self-contained desktop platform with defined display connectivity, the GTX 465. Both are End-of-life parts, and neither wins anything the database does not explicitly record.