AMD Radeon 8040S vs NVIDIA GeForce GTX 860M Comparison

The AMD Radeon 8040S is the definitive winner in this comparison, taking all seven head-to-head benchmark tests against the NVIDIA GeForce GTX 860M. The data shows a generational leap, with the Radeon 8040S delivering anywhere from 59% to 78.5% higher scores across DirectX, compute, and 2D workloads. While the GTX 860M was a capable mobile part in its era, the Radeon 8040S simply operates in a different performance class, making this a one-sided affair that favors the newer AMD silicon in every measurable way.

Head-to-Head Benchmarks

The Radeon 8040S dominates the DirectX suite. In the passmark_directx_11 test, it scores 80 against the GTX 860M's 23, a delta of -71.2% (meaning the NVIDIA part trails by that margin). The pattern repeats in passmark_directx_12, where the AMD card posts 47 versus 13, a -72.3% gap, and in passmark_directx_10, where the score is 48 versus 15, a -68.7% difference. Even in the legacy passmark_directx_9 test, the Radeon 8040S nearly triples the GTX 860M's output, scoring 134 against 55, a -59% delta. These aren't marginal wins; they are consistent, overwhelming advantages across every API generation.

The compute and 3D results are even more lopsided. In passmark_g3d, the Radeon 8040S achieves 10578, while the GTX 860M manages only 3081, a -70.9% delta. That is over three times the raw graphics performance. The passmark_gpu_compute test tells a similar story: 5138 for AMD versus 1251 for NVIDIA, a -75.7% gap, meaning the Radeon 8040S delivers more than four times the compute throughput. The 2D performance gap is equally stark, with the Radeon 8040S scoring 1052 in passmark_g2d against the GTX 860M's 226, a -78.5% delta — the largest relative difference in the entire benchmark set.

These numbers translate into a clear hierarchy. The GTX 860M's best result relative to its rival is in DirectX 9, where it still loses by 59%. Its worst is in 2D, where it loses by 78.5%. The Radeon 8040S wins every single test, and the margin is never less than roughly two-thirds of the NVIDIA card's score. For context, the GTX 860M's average benchmark score is 2967, placing it at the 19th percentile of all GPUs, while the Radeon 8040S has an average score of 2440, placing it at the 17th percentile — but that average is skewed by the fact that the AMD card has fewer benchmark entries and no Geekbench scores to boost its mean. The head-to-head data is unambiguous: the Radeon 8040S is the faster GPU in every workload tested.

Architecture Differences

The two GPUs come from entirely different design eras. The NVIDIA GeForce GTX 860M uses the GK104 chip based on the Kepler architecture, built on a 28 nm process at TSMC. It packs 3,540 million transistors into a 294 mm² die, yielding a transistor density of 12.0M per mm². The AMD Radeon 8040S, by contrast, uses the Strix Halo chip based on RDNA 3.5, fabricated on a much more advanced 4 nm process, also at TSMC. Its die size is 308 mm², slightly larger than the NVIDIA chip, but with an unspecified transistor count that prevents a direct density comparison. The process node difference alone — 28 nm versus 4 nm — explains much of the performance gulf.

Core configurations diverge significantly. The GTX 860M has 1152 shading units, 96 texture mapping units, and 16 raster output pipelines. The Radeon 8040S has 1024 shading units, 64 TMUs, and 32 ROPs. Despite having fewer shaders and TMUs, the AMD card achieves far higher throughput because of its clock speeds and architecture efficiency. The Radeon 8040S also features 16 dedicated ray tracing cores, a capability the Kepler-based GTX 860M entirely lacks, as it predates hardware ray tracing. Neither GPU has tensor cores, but the AMD part supports DirectX 12 Ultimate (12_2), while the GTX 860M is limited to DirectX 12 (11_0) — a full feature level behind.

Memory architecture is another fundamental split. The GTX 860M uses 2 GB of dedicated GDDR5 memory on a 128-bit bus, delivering 80.00 GB/s of bandwidth. The Radeon 8040S uses System Shared memory, meaning it has no dedicated VRAM and instead relies on the host system's RAM, with bus width and bandwidth listed as system dependent. This makes the AMD card more flexible in theory but ties its memory performance to the laptop's configuration. The NVIDIA part’s fixed 80.00 GB/s is a hard ceiling, while the AMD card's bandwidth could vary widely depending on the system it ships in.

Clock speeds and power draw also tell a story of architectural progress. The GTX 860M runs at a 797 MHz base and 915 MHz boost, while the Radeon 8040S operates at 1295 MHz base and 2800 MHz boost — nearly three times the boost clock. Despite the massive clock advantage, the AMD card has a lower TDP of 55 W versus the GTX 860M's 75 W. The Radeon 8040S also supports newer I/O standards, with PCIe 5.0 x16 compared to the GTX 860M's PCIe 3.0 x16, and it offers Vulkan 1.4 support versus Vulkan 1.2.175 on the NVIDIA part.

FAQ

Q: Which GPU wins in DirectX 12 performance?

A: The AMD Radeon 8040S wins decisively, scoring 47 in passmark_directx_12 compared to the NVIDIA GeForce GTX 860M's 13, a -72.3% delta. The AMD card also supports DirectX 12 Ultimate (12_2), while the NVIDIA part is limited to DirectX 12 (11_0).

Q: How do the two compare in raw compute workloads?

A: The Radeon 8040S delivers 5138 in passmark_gpu_compute versus 1251 for the GTX 860M, a -75.7% gap. This means the AMD card achieves roughly four times the compute throughput of the NVIDIA part.

Q: Does the GTX 860M have any ray tracing capability?

A: No. The GTX 860M's Kepler architecture has no ray tracing cores, while the Radeon 8040S includes 16 dedicated RT cores as part of its RDNA 3.5 architecture.

Q: What are the clock speed differences between the two?

A: The GTX 860M runs at 797 MHz base and 915 MHz boost. The Radeon 8040S runs at 1295 MHz base and 2800 MHz boost, giving the AMD card a significantly higher operating frequency across the board.

Q: How does memory configuration affect these GPUs?

A: The GTX 860M has 2 GB of dedicated GDDR5 memory on a 128-bit bus with 80.00 GB/s bandwidth. The Radeon 8040S uses System Shared memory, meaning its memory size, type, and bus width are all system dependent, with bandwidth listed as system dependent as well.

Q: Which GPU is more power efficient?

A: The Radeon 8040S has a lower TDP of 55 W compared to the GTX 860M's 75 W, despite delivering far higher performance. This reflects the efficiency gains from the 4 nm process node versus 28 nm.

Specification Differences

| Specification | NVIDIA GeForce GTX 860M | AMD Radeon 8040S |

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

| Architecture | Kepler | RDNA 3.5 |

| Process Node | 28 nm | 4 nm |

| Transistors | 3,540 million | unknown |

| Die Size | 294 mm² | 308 mm² |

| Transistor Density | 12.0M / mm² | null |

| Base Clock | 797 MHz | 1295 MHz |

| Boost Clock | 915 MHz | 2800 MHz |

| Memory Size | 2 GB | System Shared |

| Memory Type | GDDR5 | System Shared |

| Memory Bus Width | 128 bit | System Shared |

| Memory Bandwidth | 80.00 GB/s | System Dependent |

| Memory Clock | 1250 MHz (5 Gbps effective) | System Shared |

| Shading Units | 1152 | 1024 |

| TMUs | 96 | 64 |

| ROPs | 16 | 32 |

| RT Cores | null | 16 |

| Pixel Rate | 21.96 GPixel/s | 89.60 GPixel/s |

| Texture Rate | 87.84 GTexel/s | 179.2 GTexel/s |

| FP32 | 2.108 TFLOPS | 5.734 TFLOPS |

| FP16 | null | 5.734 TFLOPS (1:1) |

| TDP | 75 W | 55 W |

| Bus Interface | PCIe 3.0 x16 | PCIe 5.0 x16 |

| DirectX | 12 (11_0) | 12 Ultimate (12_2) |

| Vulkan | 1.2.175 | 1.4 |

| Release Date | 2014-03-09 | 2025-01-05 |

| Production Status | End-of-life | Active |

Where Each One Wins

The AMD Radeon 8040S wins everywhere that matters. It takes all seven benchmark tests, including every DirectX version, 2D, 3D, and compute. The largest margin is in passmark_g2d, where the AMD card leads by 78.5%, and the smallest is in passmark_directx_9, where it still leads by 59%. The Radeon 8040S also wins on architecture, offering ray tracing support, a modern 4 nm process, higher pixel and texture rates (89.60 GPixel/s and 179.2 GTexel/s versus 21.96 GPixel/s and 87.84 GTexel/s), and more than double the FP32 throughput at 5.734 TFLOPS versus 2.108 TFLOPS.

The NVIDIA GeForce GTX 860M has no benchmark wins to claim. Its only advantages lie in specific specification categories: it has more shading units (1152 versus 1024) and more TMUs (96 versus 64), plus a fixed 80.00 GB/s memory bandwidth that does not depend on system configuration. It also has a smaller die size at 294 mm² versus 308 mm², though that is hardly a performance benefit. The GTX 860M's dedicated 2 GB GDDR5 memory could be seen as a plus in scenarios where the host system's shared memory is slow or limited, but the benchmark data shows that even with this dedicated memory, it cannot compete with the Radeon 8040S's performance.

The Verdict

The choice is clear from the data: the AMD Radeon 8040S is the superior GPU for virtually any workload. It wins every benchmark by a wide margin, offers modern features like ray tracing and DirectX 12 Ultimate, runs cooler at 55 W versus 75 W, and provides more than double the raw FP32 compute power. The GTX 860M is an end-of-life product from 2014 with no path to competitiveness against the 2025-era Radeon 8040S.

Buyers or system integrators considering the GTX 860M should look to the Radeon 8040S for any modern gaming, compute, or general-purpose graphics task. The only scenario where the GTX 860M might be preferable is in a system that absolutely requires dedicated VRAM with guaranteed 80.00 GB/s bandwidth, rather than relying on system shared memory that could bottleneck depending on the laptop's RAM configuration. But even then, the performance gap is so large that the GTX 860M's fixed memory advantage does not translate into a single benchmark win. For anyone choosing between these two, the Radeon 8040S is the only rational pick based on the evidence.