NVIDIA GeForce GT 740 vs NVIDIA GeForce GTX 860M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GT 740 posts an average benchmark score of 3431, while the NVIDIA GeForce GTX 860M averages 2967. The GT 740 sits in the 21st percentile of all GPUs, whereas the GTX 860M is in the 19th percentile.

Q: How do the two cards compare in raw compute performance?

A: The GTX 860M delivers 2.108 TFLOPS of FP32 throughput, nearly three times the 762.6 GFLOPS of the GT 740. Its texture rate of 87.84 GTexel/s also dwarfs the GT 740's 31.78 GTexel/s.

Q: What memory configurations do the two cards use?

A: The GT 740 has 1024 MB of GDDR5 on a 128-bit bus, yielding a bandwidth of 80.19 GB/s. The GTX 860M doubles capacity to 2 GB of GDDR5 on the same 128-bit bus, at a nearly identical 80.00 GB/s.

Q: Which GPU offers better API support?

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

Q: How does the GTX 860M perform in the Geekbench Vulkan test compared to the GT 740?

A: The GTX 860M scores 9648 in Geekbench Vulkan, which is 57.7% higher than the GT 740's 4083. This is the smallest relative gap among the three head-to-head benchmarks.

Q: What is the transistor count difference between the two chips?

A: The GTX 860M's GK104 chip contains 3,540 million transistors on a 294 mm² die, whereas the GT 740's GK107 has 1,270 million transistors on a 118 mm² die. The GTX 860M also has a slightly higher transistor density at 12.0M / mm² versus 10.8M / mm².

Where Each One Wins

The benchmark data paints a clear picture: the GTX 860M wins every single head-to-head test. In Geekbench Metal, it scores 4900 against the GT 740's 2363 — a 51.8% advantage. The OpenCL gap is even wider, with 10461 versus 3847, a 63.2% delta. The Vulkan test shows a 57.7% lead, with 9648 versus 4083. The GT 740 does not win a single benchmark in the comparison set.

However, the GT 740's higher average score across all benchmarks (3431 vs 2967) suggests it may perform better in workloads not captured by the head-to-head tests. Its nearest rivals include the Intel HD Graphics P4600 (avg 3389, deltaPct 1.2) and Intel HD Graphics 530 (avg 3332, deltaPct 3), indicating it slots into the integrated-graphics performance tier in broader testing. The GTX 860M, by contrast, sits near the GeForce GTX 750 Ti (avg 2953, deltaPct 0.5) and the Quadro P600 (avg 2923, deltaPct 1.5) — a different performance neighborhood entirely.

For pure compute-heavy tasks like OpenCL workloads, the GTX 860M is the clear choice. For scenarios where the average of many benchmark types matters, the GT 740's higher percentile ranking suggests it may be more consistent across diverse applications. The GTX 860M's PassMark scores show strength in DirectX 9 (55) and G2D (226), but comparatively weak DirectX 10 (15), 11 (23), and 12 (13) results — a peculiar pattern that hints at driver or architecture quirks in legacy API paths.

Architecture Differences

Both GPUs use NVIDIA's Kepler architecture, but they are fundamentally different chips. The GT 740 is built on the GK107 die, while the GTX 860M uses the larger GK104. The transistor counts reflect this: 1,270 million versus 3,540 million. The die sizes differ substantially as well — 118 mm² for the GT 740 against 294 mm² for the GTX 860M — which explains the GTX 860M's higher transistor density of 12.0M / mm² versus 10.8M / mm².

The shading engine is where the gap becomes enormous. The GTX 860M packs 1152 shading units and 96 texture mapping units, versus just 384 shading units and 32 TMUs on the GT 740. Both have 16 ROPs. This threefold difference in shading units and texture units directly explains the GTX 860M's massive lead in FP32 throughput (2.108 TFLOPS vs 762.6 GFLOPS) and texture fill rate (87.84 vs 31.78 GTexel/s).

The pixel rate tells a similar story: 21.96 GPixel/s for the GTX 860M versus 7.944 GPixel/s for the GT 740. Both are manufactured on TSMC's 28 nm process, and both come from the same foundry. The GTX 860M's higher TDP of 75 W versus 64 W is modest given the nearly threefold increase in compute resources, suggesting the GK104 chip is more power-efficient per transistor.

The GTX 860M has no power connectors — it is an IGP (integrated graphics processor) for laptops, with display outputs described as "Portable Device Dependent." The GT 740 is a single-slot desktop card requiring a 1x 6-pin power connector and a suggested 250 W PSU. The GT 740 measures 145 mm (5.7 inches) in length. Both use PCIe 3.0 x16 interfaces.

Specification Differences

The two cards diverge on nearly every specification beyond their shared Kepler architecture and 28 nm process. The GT 740 uses the GK107 chip; the GTX 860M uses GK104. The GT 740 belongs to the GeForce 700 generation, while the GTX 860M is from the GeForce 800M series. Their predecessors and successors differ accordingly: the GT 740 follows GeForce 600 and leads to GeForce 900, whereas the GTX 860M follows GeForce 700M and leads to GeForce 900M.

Clock speeds differ significantly. The GT 740 does not list a base or boost clock in the data, while the GTX 860M runs at 797 MHz base and 915 MHz boost. Memory clocks are nearly identical — 1253 MHz for the GT 740 and 1250 MHz for the GTX 860M, both yielding 5 Gbps effective.

Memory capacity is the clearest differentiator: 1024 MB on the GT 740 versus 2 GB on the GTX 860M. Bus width is the same at 128 bit, and bandwidth is virtually identical (80.19 vs 80.00 GB/s). The GT 740 has 384 shading units, 32 TMUs, and 16 ROPs; the GTX 860M has 1152 shading units, 96 TMUs, and 16 ROPs.

Power and physical design differ substantially. The GT 740 draws 64 W and is a single-slot card with 2x DVI and 1x mini-HDMI 1.4a outputs. The GTX 860M draws 75 W, has no power connectors, no slot width (IGP), and its display outputs are portable-device dependent. The GT 740 has a suggested PSU of 250 W; the GTX 860M lists none. The GT 740's launch MSRP was 89 USD.

Head-to-Head Benchmarks

The three head-to-head tests all favor the GTX 860M, but the margins vary meaningfully. The smallest gap is in Geekbench Metal, where the GTX 860M scores 4900 against 2363 for the GT 740 — a delta of -51.8% from the GT 740's perspective. This means the GT 740 achieves less than half the GTX 860M's Metal performance.

The OpenCL test shows the largest disparity: 10461 versus 3847, a -63.2% delta. The GTX 860M's score is roughly 2.7 times higher. This test likely stresses the raw compute throughput, where the GTX 860M's 1152 shading units and 2.108 TFLOPS dominate the GT 740's 384 units and 762.6 GFLOPS.

The Vulkan benchmark falls in between: 9648 versus 4083, a -57.7% delta. The GTX 860M leads by a factor of about 2.4. Interestingly, the GT 740's Vulkan score of 4083 is higher than its OpenCL score of 3847, while the GTX 860M's OpenCL score of 10461 exceeds its Vulkan score of 9648. This inverted pattern suggests the two cards have different driver optimization paths for these APIs.

The GTX 860M's PassMark results add context beyond the Geekbench trio. Its DirectX 9 score of 55 is strong, but DirectX 10 (15), 11 (23), and 12 (13) are all notably lower. The G2D score of 226 and G3D score of 3081 indicate solid 2D performance and moderate 3D capability. The GPU compute score of 1251 is respectable for a laptop part.

The GT 740's nearest rival data shows it is competitive with the Quadro 2000M (avg 3434, deltaPct -0.1) and GeForce 920MX (avg 3528, deltaPct -2.7). The GTX 860M's nearest rivals include the GTX 750 Ti (avg 2953, deltaPct 0.5) and GeForce 820A (avg 2983, deltaPct -0.5). These rival groupings suggest the GT 740 and GTX 860M occupy different tiers in the broader GPU landscape, despite the GT 740's higher average score.

The Verdict

The data directs different users to different cards. If the workload involves OpenCL compute, Metal graphics, or Vulkan rendering — the three head-to-head tests — the GTX 860M is the unambiguous winner. Its 63.2% lead in OpenCL and 51.8% lead in Metal are decisive. The GTX 860M's 2 GB memory capacity also gives it a practical edge for modern textures, while the GT 740's 1024 MB may prove limiting.

However, the GT 740's higher average benchmark score (3431 vs 2967) and better percentile ranking (21st vs 19th) tell a different story. In a broader benchmark suite that includes diverse workloads beyond the three head-to-head tests, the GT 740 appears more consistent. Its nearest rivals are integrated graphics parts like the Intel HD Graphics 530 and P4600, suggesting it is a capable entry-level desktop solution. The GTX 860M's nearest rivals are more varied, including the GTX 750 Ti and even the RTX 4060 Ti 8 GB (avg 2913, deltaPct 1.9) — an odd grouping that may reflect specific workload characteristics.

Desktop users with a 250 W PSU and a 6-pin connector can install the GT 740 as a single-slot card with DVI and mini-HDMI outputs. Laptop users get the GTX 860M as a mobile IGP with no power connectors and portable-dependent display outputs. The GT 740's 64 W TDP is lower than the GTX 860M's 75 W, but both are modest power consumers.

The verdict depends on context. For portable gaming and compute-heavy tasks, the GTX 860M's threefold advantage in shading units and its 2 GB frame buffer make it the stronger choice — the head-to-head data is unambiguous. For a desktop system where average performance across many benchmark types matters more than peak compute, the GT 740's higher average score and percentile position suggest it holds up well in general-purpose use. The GT 740 also has the advantage of a defined launch MSRP of 89 USD, though pricing considerations are beyond the scope of this analysis. In short: the GTX 860M wins where it matters most — raw performance — while the GT 740 offers a more balanced profile across diverse workloads.