NVIDIA GeForce 820M vs NVIDIA GeForce GTX 750 Ti Comparison

# NVIDIA GeForce GTX 750 Ti vs NVIDIA GeForce 820M

The NVIDIA GeForce GTX 750 Ti and NVIDIA GeForce 820M sit on opposite ends of the mobile-desktop spectrum, yet both target users who need basic 3D acceleration without breaking a power budget. The GTX 750 Ti is a desktop card with a 60W TDP, while the 820M is a 45W mobile module designed for laptops. Benchmark data shows a massive gulf between them: the GTX 750 Ti averages 2953 across all tests, while the 820M averages just 2784. That 169-point gap translates to a 288.9% difference in the one benchmark where both were tested—Geekbench OpenCL. The 750 Ti also holds a 19th percentile ranking among all GPUs versus the 820M's 18th, meaning both are near the bottom of the performance ladder, but the 750 Ti is still meaningfully ahead.

Where Each One Wins

The GTX 750 Ti wins every head-to-head benchmark in the data set. The sole shared test, Geekbench OpenCL, shows the 750 Ti scoring 10827 against the 820M's 2784, a 288.9% advantage. That is not a close contest; it is a generational and architectural stomping. The 750 Ti also has a full suite of additional benchmark results—Passmark G3D at 3903, Passmark GPU Compute at 1744, and Geekbench Vulkan at 10065—while the 820M has no recorded scores in those tests. If you are running OpenCL workloads, the 750 Ti is the only rational choice between these two.

The 820M's only "win" is its 45W TDP versus 60W for the 750 Ti, making it the lower-power part. That matters in a laptop where thermal headroom is scarce. But lower power consumption does not translate to any benchmark victory. The 820M is also an MXM module, meaning it is designed for portable devices, whereas the 750 Ti is a single-slot desktop card. So the 820M wins on form factor and power draw, but loses on every measurable performance metric.

Architecture Differences

The 750 Ti uses the GM107 chip built on the Maxwell architecture, while the 820M uses the GK107 chip from the older Kepler architecture. Both are fabricated on TSMC's 28nm process, but they differ substantially in transistor count and die size. The GM107 packs 1,870 million transistors into a 148mm² die, giving a transistor density of 12.6M per mm². The GK107 has 1,270 million transistors on a 118mm² die, for a density of 10.8M per mm². Maxwell is the newer design, and the density advantage reflects that.

The 750 Ti features 640 shading units, 40 texture mapping units, and 16 ROPs. The 820M is cut down to 384 shading units, 32 TMUs, and 16 ROPs. That is a 40% reduction in shader count and a 20% reduction in TMUs, which directly explains the performance gap. Clock speeds reinforce the difference: the 750 Ti runs at 1020 MHz base and 1085 MHz boost, while the 820M has no base or boost clock listed in the data. Memory also diverges sharply. The 750 Ti uses 2GB of GDDR5 at 1350 MHz (5.4 Gbps effective) with 86.40 GB/s of bandwidth. The 820M uses 2GB of DDR3 at 901 MHz (1802 Mbps effective) with just 28.83 GB/s of bandwidth. That is a threefold bandwidth advantage for the 750 Ti.

Pixel and texture rates tell the same story. The 750 Ti outputs 17.36 GPixel/s and 43.40 GTexel/s, while the 820M manages only 6.480 GPixel/s and 25.92 GTexel/s. FP32 compute is 1,388.8 GFLOPS for the 750 Ti versus 622.1 GFLOPS for the 820M. Both support DirectX 12 (11_0) and OpenGL 4.6, but the 750 Ti has Vulkan 1.4 support while the 820M is limited to Vulkan 1.2.175. The 750 Ti also has a PCIe 3.0 x16 bus interface, same as the 820M, but the desktop card offers display outputs (2x DVI, 1x mini-HDMI 1.4a) whereas the 820M's outputs are portable-device dependent.

The Verdict

Pick the GTX 750 Ti if you need desktop performance and have a 250W PSU available. The data shows it is nearly 4x faster in OpenCL, has 3x the memory bandwidth, and supports a newer Vulkan version. Its 60W TDP is modest for a desktop card, and it requires no power connectors, making it an easy drop-in for an existing system. The 750 Ti's 19th percentile ranking is low in absolute terms, but it is still one full percentile above the 820M, and its average benchmark score of 2953 is 169 points higher.

Pick the 820M only if you are constrained to a laptop or an MXM module and need the absolute lowest power draw. At 45W, it consumes 25% less power than the 750 Ti. But you pay for that efficiency with dramatically lower performance: 622.1 GFLOPS of FP32 compute, 28.83 GB/s of bandwidth, and a single benchmark score of 2784 in OpenCL. The 820M has no launch MSRP listed, so its value proposition is unclear; the 750 Ti launched at 149 USD, but you should not base a purchasing decision on that alone given the performance delta.

For any compute workload, gaming at low settings, or GPU-accelerated tasks, the 750 Ti is the clear winner. The 820M is a fallback for portable systems where a discrete GPU is better than integrated graphics—but marginally so, given its 18th percentile ranking.

FAQ

Q: Which GPU is faster in OpenCL?

A: The GTX 750 Ti scores 10827 in Geekbench OpenCL, while the 820M scores 2784. That is a 288.9% difference in favor of the 750 Ti.

Q: Do both cards support the same DirectX version?

A: Yes, both support DirectX 12 (11_0) and OpenGL 4.6. However, the GTX 750 Ti supports Vulkan 1.4, while the 820M is limited to Vulkan 1.2.175.

Q: How much memory bandwidth does each card have?

A: The GTX 750 Ti has 86.40 GB/s from 2GB of GDDR5 on a 128-bit bus. The 820M has 28.83 GB/s from 2GB of DDR3 on the same 128-bit bus.

Q: Which card has more shading units?

A: The GTX 750 Ti has 640 shading units, while the 820M has 384. The 750 Ti also has 40 TMUs versus 32 on the 820M, but both have 16 ROPs.

Q: What is the power consumption difference?

A: The GTX 750 Ti has a 60W TDP and requires no power connectors, with a suggested PSU of 250W. The 820M has a 45W TDP and is an MXM module designed for portable devices.

Q: Are there any benchmarks where the 820M wins?

A: No. The only head-to-head benchmark, Geekbench OpenCL, is won by the GTX 750 Ti. The 820M has no recorded wins in any test.

Head-to-Head Benchmarks

The single shared benchmark is Geekbench OpenCL, and it is a rout. The GTX 750 Ti posts 10827, while the 820M manages 2784. That is a 288.9% delta, meaning the 750 Ti is nearly four times faster. In practical terms, OpenCL workloads like video encoding, physics simulation, or compute shaders will finish in roughly a quarter of the time on the 750 Ti.

The 750 Ti's additional benchmark results put this gap in context. Its Passmark G3D score of 3903 and Passmark GPU Compute score of 1744 show it is a capable entry-level desktop GPU. Its Geekbench Vulkan score of 10065 is nearly identical to its OpenCL score, indicating consistent compute performance across APIs. The 820M has no comparable data, so we cannot assess its Vulkan or DirectX performance directly. But given its OpenCL deficit, it is reasonable to infer similar gaps elsewhere.

The nearest rival data adds another layer. The 750 Ti's closest competitor is the GTX 860M, which averages 2967—just 0.5% higher. The 820A averages 2983, a 1% delta. The Quadro P600 is 1% slower, and the RTX 4060 Ti 8GB is 1.4% slower in this specific comparison set. That last entry is odd but instructive: the 750 Ti's average score of 2953 is within 1.4% of an RTX 4060 Ti's average of 2913 in this dataset, though that does not mean they perform alike in real-world gaming—it reflects the benchmark averaging methodology.

The 820M's nearest rivals include the RTX 3060 Ti GDDR6X at 2795 (0.4% higher), the AMD Radeon 550 at 2801 (0.6% higher), the Intel UHD Graphics 610 at 2752 (1.2% lower), and the RTX 5070 SUPER at 2690 (3.5% lower). This places the 820M in extremely weak company—it is trading blows with integrated graphics and entry-level desktop cards. The 750 Ti, while still low-end, is a full tier above.

Specification Differences

The two cards differ in nearly every meaningful specification. The 750 Ti uses the GM107 chip (Maxwell architecture, 1,870 million transistors, 148mm² die), while the 820M uses the GK107 chip (Kepler architecture, 1,270 million transistors, 118mm² die). Both are 28nm TSMC parts, but the Maxwell design is denser at 12.6M transistors per mm² versus 10.8M.

Clock speeds: the 750 Ti runs at 1020 MHz base and 1085 MHz boost; the 820M has no listed base or boost clocks. Memory clocks also differ: 1350 MHz (5.4 Gbps effective) for the 750 Ti versus 901 MHz (1802 Mbps effective) for the 820M. Memory type is GDDR5 on the 750 Ti versus DDR3 on the 820M, with the same 2GB capacity and 128-bit bus width. Bandwidth is 86.40 GB/s versus 28.83 GB/s.

Compute resources: 640 shading units, 40 TMUs, and 16 ROPs on the 750 Ti; 384 shading units, 32 TMUs, and 16 ROPs on the 820M. Pixel rate is 17.36 GPixel/s versus 6.480 GPixel/s. Texture rate is 43.40 GTexel/s versus 25.92 GTexel/s. FP32 is 1,388.8 GFLOPS versus 622.1 GFLOPS. TDP is 60W versus 45W. The 750 Ti is single-slot with no power connectors and a 250W suggested PSU; the 820M is an MXM module with no power connector data. Display outputs are 2x DVI and 1x mini-HDMI on the 750 Ti versus portable-device-dependent on the 820M. Vulkan support is 1.4 on the 750 Ti versus 1.2.175 on the 820M. The 750 Ti has a 145mm length (5.7 inches); the 820M has no dimensions listed. Release dates are February 2014 for the 750 Ti and March 2015 for the 820M.