NVIDIA GeForce 820M vs NVIDIA GeForce GTX 860M Comparison

The NVIDIA GeForce GTX 860M and NVIDIA GeForce 820M are both members of the GeForce 800M generation, but they occupy vastly different performance tiers. The GTX 860M is positioned as a high-performance mobile GPU, while the 820M is an entry-level part designed for basic tasks. Benchmark data confirms a decisive gulf between the two, with the GTX 860M delivering nearly four times the raw compute performance in the one shared test.

Head-to-Head Benchmarks

The only common benchmark between these two GPUs is Geekbench OpenCL, and the results are unambiguous. The GTX 860M scores 10,461 points, while the 820M scores just 2,784 points. This represents a delta of 275.8% in favor of the GTX 860M, meaning the higher-end part is approximately 3.76 times faster in this compute workload. This is not a marginal difference; it is a generational chasm within the same product family.

Looking at the broader benchmark suite for the GTX 860M, its performance profile is consistent across APIs. It scores 4,900 in Geekbench Metal, 9,648 in Geekbench Vulkan, and 3,081 in Passmark G3D. Its Passmark GPU Compute score is 1,251, and its average benchmark score across all tests is 2,967. The 820M’s average score is 2,784, which is lower than even the GTX 860M’s weakest OpenCL result.

When examining the GTX 860M’s standing relative to its nearest rivals, the data shows it sits in a tight cluster. It trails the NVIDIA GeForce 820A by 0.5% (2,983 vs. 2,953 for the 860M), but leads the NVIDIA GeForce GTX 750 Ti by 0.5% (2,953) and the NVIDIA Quadro P600 by 1.5% (2,923). Interestingly, it also edges out the NVIDIA GeForce RTX 4060 Ti 8 GB by 1.9% (2,913) in average score, though that comparison is heavily influenced by the RTX 4060 Ti’s much higher scores in modern DirectX 12 workloads versus the older tests where the 860M competes.

The 820M’s nearest rivals paint a different picture of its competitive position. It trails the NVIDIA GeForce RTX 3060 Ti GDDR6X by 0.4% (2,795 vs. 2,784) and the AMD Radeon 550 by 0.6% (2,801). It leads the Intel UHD Graphics 610 by 1.2% (2,752) and the NVIDIA GeForce RTX 5070 SUPER by 3.5% (2,690). These deltas are all within a narrow band, indicating that the 820M’s single OpenCL score places it in a similar performance class to those very different products, despite their wildly different architectures and release dates.

The percentile rankings reinforce the gap. The GTX 860M sits at the 19th percentile of all GPUs, while the 820M sits at the 18th percentile. While these are close, the absolute score difference of 183 points (2,967 vs. 2,784) represents a meaningful step up in capability, especially considering the 860M’s multiple benchmark scores versus the 820M’s single data point.

FAQ

Q: How much faster is the GTX 860M than the 820M in OpenCL compute performance?

A: The GTX 860M scores 10,461 in Geekbench OpenCL, while the 820M scores 2,784. This gives the GTX 860M a 275.8% advantage in that specific test.

Q: Do both GPUs support the same modern graphics APIs?

A: Yes. Both the GTX 860M and the 820M report DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175 support, indicating identical API feature levels despite their performance differences.

Q: Which GPU has a higher average benchmark score?

A: The GTX 860M has an average benchmark score of 2,967, whereas the 820M’s average is 2,784. The GTX 860M is therefore the higher-performing part on average.

Q: How does the GTX 860M compare to the GTX 750 Ti?

A: The GTX 860M has an average score of 2,967, which is 0.5% higher than the GTX 750 Ti’s average score of 2,953. The two are effectively tied in average performance.

Q: Is the 820M’s performance significantly different from integrated graphics?

A: The 820M’s average score of 2,784 is 1.2% higher than the Intel UHD Graphics 610’s score of 2,752. The difference is small, suggesting the 820M offers only a modest improvement over that integrated solution.

Q: Which GPU has a higher percentile ranking?

A: The GTX 860M ranks at the 19th percentile of all GPUs, while the 820M ranks at the 18th percentile. The GTX 860M is marginally better positioned in the overall distribution.

Architecture Differences

Both GPUs are built on NVIDIA’s Kepler architecture and use a 28 nm process node from TSMC, but the silicon underneath is fundamentally different. The GTX 860M uses the GK104 chip, which contains 3,540 million transistors on a die size of 294 mm². In contrast, the 820M uses the GK107 chip, which is much smaller at 118 mm² and packs only 1,270 million transistors. The transistor density also differs slightly, with the GK104 achieving 12.0 million transistors per mm² versus 10.8 million per mm² for the GK107.

The execution resources are heavily skewed toward the GTX 860M. It has 1,152 shading units, 96 texture mapping units (TMUs), and 16 ROPs. The 820M has only 384 shading units and 32 TMUs, though it also has 16 ROPs. This translates into a massive difference in raw throughput: the GTX 860M delivers 2.108 TFLOPS of FP32 compute, while the 820M manages just 622.1 GFLOPS. Pixel fill rates are similarly divergent, with the GTX 860M hitting 21.96 GPixel/s versus 6.480 GPixel/s for the 820M, and texture rates of 87.84 GTexel/s versus 25.92 GTexel/s.

Memory architecture is another point of separation. Both use a 128-bit bus, but the GTX 860M pairs it with 2 GB of GDDR5 memory running at 1250 MHz (5 Gbps effective), yielding 80.00 GB/s of bandwidth. The 820M uses 2 GB of DDR3 at 901 MHz (1802 Mbps effective), which provides only 28.83 GB/s. This bandwidth disparity is critical for memory-bound workloads and contributes significantly to the performance gap.

Clock behavior also differs. The GTX 860M has a base clock of 797 MHz and a boost clock of 915 MHz, while the 820M does not list base or boost clocks. The power envelope is lower for the 820M at 45 W TDP, versus 75 W for the GTX 860M. The GTX 860M is designed as an IGP (integrated graphics processor) for laptops, while the 820M is an MXM module. Both use a PCIe 3.0 x16 bus interface and have portable-device-dependent display outputs. Neither GPU includes ray tracing or tensor cores, and both are marked as end-of-life products. The GTX 860M was released on March 9, 2014, while the 820M came later on March 20, 2015.

The Verdict

The data is clear: the GTX 860M is the overwhelmingly superior GPU. Its 275.8% lead in OpenCL performance is the headline number, but the architectural advantages in shading units, texture units, memory bandwidth, and compute throughput all point in the same direction. Anyone needing to run compute-heavy tasks, modern games, or GPU-accelerated applications should choose the GTX 860M without hesitation.

The 820M is not without merit, but its merits are narrow. Its 45 W TDP is much lower than the 75 W of the GTX 860M, making it a more power-efficient choice for basic tasks. Its performance is competitive with integrated graphics like the Intel UHD Graphics 610, and its average score of 2,784 places it within 0.6% of the AMD Radeon 550. For users whose workloads are limited to simple 2D tasks, video playback, or very light 3D acceleration, the 820M is adequate.

However, the 820M’s single benchmark result and 18th percentile ranking suggest it is at the very bottom of the performance hierarchy. The GTX 860M, while also low-ranked at the 19th percentile, offers a substantially better experience for any graphics-intensive application. The choice is straightforward: the GTX 860M for performance, the 820M only when power consumption and basic functionality are the primary concerns.

Specification Differences

| Specification | NVIDIA GeForce GTX 860M | NVIDIA GeForce 820M |

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

| Chip | GK104 | GK107 |

| Process Node | 28 nm | 28 nm |

| Transistors | 3,540 million | 1,270 million |

| Die Size | 294 mm² | 118 mm² |

| Transistor Density | 12.0M / mm² | 10.8M / mm² |

| Base Clock | 797 MHz | Not specified |

| Boost Clock | 915 MHz | Not specified |

| Memory Clock | 1250 MHz (5 Gbps effective) | 901 MHz (1802 Mbps effective) |

| Memory Size | 2 GB | 2 GB |

| Memory Type | GDDR5 | DDR3 |

| Memory Bus Width | 128 bit | 128 bit |

| Memory Bandwidth | 80.00 GB/s | 28.83 GB/s |

| Shading Units | 1152 | 384 |

| TMUs | 96 | 32 |

| ROPs | 16 | 16 |

| Pixel Rate | 21.96 GPixel/s | 6.480 GPixel/s |

| Texture Rate | 87.84 GTexel/s | 25.92 GTexel/s |

| FP32 Performance | 2.108 TFLOPS | 622.1 GFLOPS |

| TDP | 75 W | 45 W |

| Slot Width | IGP | MXM Module |

| Power Connectors | None | Not specified |

| Release Date | March 9, 2014 | March 20, 2015 |