AMD Radeon 890M vs NVIDIA GeForce GTX TITAN BLACK Comparison

The NVIDIA GeForce GTX TITAN BLACK and the AMD Radeon 890M represent two extremes of GPU design separated by a decade of progress. The TITAN BLACK is a 2014-era flagship discrete card built on a massive die, while the 890M is a 2024 integrated graphics processor (IGP) designed for thin-and-light mobile devices. Benchmark results show a surprisingly close overall contest, with the AMD part winning 5 of 9 head-to-head tests, but the nature of those wins reveals starkly different strengths.

Head-to-Head Benchmarks

The most decisive victories belong to the AMD Radeon 890M in modern API compute workloads. In Geekbench OpenCL, the 890M scores 37,254 against the TITAN BLACK's 27,218, a 26.9% advantage. The gap widens further in Geekbench Vulkan, where the 890M posts 40,808 versus 29,561, a 27.6% lead. These are not marginal wins; they represent a fundamental shift in compute efficiency that favors the newer architecture.

The TITAN BLACK, however, dominates in legacy DirectX scenarios. In Passmark DirectX 10, it scores 47 versus 33, a 42.4% margin. In DirectX 9, the lead is 37.1%, with scores of 133 versus 97. The older card also takes Passmark G3D by 13.7% (9,184 vs 8,076), which is the aggregate 3D gaming performance metric. The two cards tie exactly in Passmark DirectX 11 at 73 points each, a rare dead heat.

The AMD 890M wins the remaining tests. In Passmark DirectX 12, it leads 37 to 34, an 8.1% edge. The most lopsided result of the entire comparison is Passmark G2D, where the 890M scores 979 versus 595, a 39.2% advantage in 2D graphics throughput. In Passmark GPU Compute, the 890M takes a 10.4% win with 4,157 versus 3,726.

Overall averages reflect this split. The TITAN BLACK has an average benchmark score of 9,474, placing it in the 46th percentile of all GPUs. The 890M averages 9,210, sitting at the 45th percentile. Despite the TITAN BLACK's slightly higher average, the 890M wins the majority of individual tests, indicating that the modern part is more balanced across different workload types.

Architecture Differences

The two GPUs could not be more different in construction. The TITAN BLACK uses the GK110B chip on the Kepler architecture, built on a 28 nm process at TSMC. It contains 7,080 million transistors on a 561 mm² die, yielding a transistor density of 12.6 million per square millimeter. The Radeon 890M uses the Strix Point chip on RDNA 3.5, also fabricated by TSMC but on a 4 nm node. It packs 34,000 million transistors onto a 233 mm² die, achieving a density of 145.9 million per square millimeter—more than 11 times denser.

Memory configuration is another stark divider. The TITAN BLACK carries 6 GB of GDDR5 on a 384-bit bus, delivering 336.0 GB/s of bandwidth. The 890M uses shared system memory, with its size, type, bus width, and bandwidth all listed as "System Shared" or "System Dependent." This makes the 890M's memory performance entirely dependent on the host laptop's RAM configuration, whereas the TITAN BLACK has dedicated, fixed-bandwidth VRAM.

Shader resources differ significantly. The TITAN BLACK has 2,880 shading units, 240 texture mapping units (TMUs), and 48 raster output units (ROPs). The 890M has 1,024 shaders, 64 TMUs, and 32 ROPs. Despite fewer units, the 890M achieves a higher pixel rate of 92.80 GPixel/s versus 58.80 GPixel/s, thanks to its much higher boost clock of 2,900 MHz versus 980 MHz. Texture rate favors the TITAN BLACK at 235.2 GTexel/s versus 185.6 GTexel/s. In raw FP32 compute, the 890M edges ahead at 5.939 TFLOPS versus 5.645 TFLOPS, and it also offers FP16 at 5.939 TFLOPS (1:1), which the TITAN BLACK lacks entirely.

The 890M includes 16 ray tracing cores, a feature completely absent from the Kepler architecture. API support also diverges: the 890M supports DirectX 12 Ultimate (12_2), while the TITAN BLACK is limited to DirectX 12 (11_1). Both support OpenGL 4.6, but Vulkan support differs—the 890M supports Vulkan 1.4, the TITAN BLACK only 1.2.175.

Where Each One Wins

The TITAN BLACK is the clear choice for older DirectX 9 and DirectX 10 titles. Its 42.4% and 37.1% margins in those respective tests indicate strong legacy driver optimization and raw rasterization muscle that modern IGPs have deprioritized. It also wins aggregate 3D performance in Passmark G3D, making it the better pick for overall gaming throughput in traditional benchmarks. Its dedicated 6 GB GDDR5 frame buffer with 336.0 GB/s bandwidth is a major asset for high-resolution textures, though the data does not specifically measure that scenario.

The Radeon 890M wins where modern APIs and compute matter. Its 27.6% Vulkan lead and 26.9% OpenCL lead show it is substantially faster for contemporary compute workloads like physics simulation, AI inference, and content creation tasks that leverage these APIs. The 39.2% G2D advantage makes it vastly better for desktop compositing, video playback, and general 2D interface work. Its DirectX 12 win, while modest at 8.1%, is important because DirectX 12 is the baseline for current and upcoming games. The GPU compute win of 10.4% further cements its position for non-gaming parallel workloads.

FAQ

Q: Which GPU is faster overall?

A: The TITAN BLACK has a higher average benchmark score of 9,474 versus 9,210 for the 890M, a 2.9% difference. However, the 890M wins 5 of 9 individual head-to-head tests, so the "faster" label depends entirely on workload type.

Q: How do they compare in DirectX 12 performance?

A: The Radeon 890M wins Passmark DirectX 12 with a score of 37 versus 34 for the TITAN BLACK, an 8.1% advantage. The 890M also supports DirectX 12 Ultimate (12_2), while the TITAN BLACK is limited to DirectX 12 (11_1).

Q: Is the TITAN BLACK better for older games?

A: Yes, benchmark results clearly show this. The TITAN BLACK leads by 42.4% in DirectX 10 and 37.1% in DirectX 9, making it substantially stronger for legacy titles that rely on those APIs.

Q: Does the Radeon 890M support ray tracing?

A: Yes, it has 16 ray tracing cores. The TITAN BLACK has no ray tracing hardware whatsoever, as the Kepler architecture predates that technology.

Q: How much memory bandwidth does each GPU have?

A: The TITAN BLACK has dedicated 6 GB GDDR5 memory with 336.0 GB/s bandwidth on a 384-bit bus. The 890M uses system shared memory, with bandwidth listed as "System Dependent," meaning it varies based on the host system's RAM.

Q: What is the TITAN BLACK's launch price?

A: The launch MSRP is 999 USD. The 890M has no launch MSRP listed.

The Verdict

The data paints a clear picture for two distinct buyers. Choose the NVIDIA GeForce GTX TITAN BLACK if you primarily run older DirectX 9/10 games, need guaranteed dedicated VRAM bandwidth of 336.0 GB/s, or require the higher texture rate of 235.2 GTexel/s. Its 13.7% lead in Passmark G3D shows it remains the stronger traditional 3D rasterizer, and its 42.4% DirectX 10 win makes it the obvious pick for that specific use case. The 250 W TDP and dual-slot cooler are acceptable for a desktop system with a 600 W suggested PSU.

Choose the AMD Radeon 890M if you prioritize modern API support, compute performance, or efficiency. Its 27.6% Vulkan and 26.9% OpenCL leads make it decisively better for current-generation compute tasks. The 39.2% G2D win makes it superior for everyday desktop use. Its 15 W TDP, lack of power connectors, and IGP form factor mean it can power a thin laptop with no dedicated cooling. The 890M also offers ray tracing cores and DirectX 12 Ultimate support, future-proofing it for upcoming software standards. Its 8.1% DirectX 12 win, while small, indicates better performance for the API that matters most going forward.

The TITAN BLACK sits in the 46th percentile of all GPUs, just ahead of the 890M's 45th, but the modern part achieves near-parity while consuming 94% less power (15 W vs 250 W) and occupying no slot space. For a new build in 2024 or later, the 890M's architecture, API support, and compute efficiency make it the more practical choice. The TITAN BLACK remains relevant only for a narrow niche of legacy gaming and specific rasterization-heavy workloads.

Specification Differences

| Specification | NVIDIA GeForce GTX TITAN BLACK | AMD Radeon 890M |

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

| Architecture | Kepler | RDNA 3.5 |

| Process Node | 28 nm | 4 nm |

| Transistors | 7,080 million | 34,000 million |

| Die Size | 561 mm² | 233 mm² |

| Transistor Density | 12.6M / mm² | 145.9M / mm² |

| Base Clock | 889 MHz | 400 MHz |

| Boost Clock | 980 MHz | 2900 MHz |

| Memory Size | 6 GB | System Shared |

| Memory Type | GDDR5 | System Shared |

| Memory Bus Width | 384 bit | System Shared |

| Memory Bandwidth | 336.0 GB/s | System Dependent |

| Shading Units | 2880 | 1024 |

| TMUs | 240 | 64 |

| ROPs | 48 | 32 |

| RT Cores | None | 16 |

| Pixel Rate | 58.80 GPixel/s | 92.80 GPixel/s |

| Texture Rate | 235.2 GTexel/s | 185.6 GTexel/s |

| FP32 | 5.645 TFLOPS | 5.939 TFLOPS |

| FP16 | Not listed | 5.939 TFLOPS (1:1) |

| TDP | 250 W | 15 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 1x 6-pin + 1x 8-pin | None |

| Suggested PSU | 600 W | Not listed |

| Bus Interface | PCIe 3.0 x16 | PCIe 4.0 x8 |

| DirectX | 12 (11_1) | 12 Ultimate (12_2) |

| Vulkan | 1.2.175 | 1.4 |

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

| Release Date | 2014-02-17 | 2024-07-14 |

| Launch MSRP | 999 USD | Not listed |