NVIDIA GeForce GTX 1070 Ti vs NVIDIA GeForce GTX 590 Comparison

# The Verdict

The recorded data places the NVIDIA GeForce GTX 1070 Ti and the NVIDIA GeForce GTX 590 in very different segments of the GPU landscape. The GTX 1070 Ti, a Pascal generation part, is a factory-tuned variant that sits close to the top of the GeForce 10 stack. The GTX 590, a Fermi 2.0 dual-GPU card, is a much older, end-of-life product. In the database averages, the GTX 1070 Ti scores 14,277 across all recorded benchmarks, while the GTX 590 averages 12,830. That gives the 1070 Ti an 11.3% lead in aggregate performance.

The GTX 590 has a few legacy strengths. It is a dual-GPU card, and in raw compute workloads that scale with a second die, it can pull ahead. But its age shows up in driver support, API levels, and efficiency metrics. The GTX 1070 Ti is the better choice for current software, lower power draw, and any workload that depends on modern geometry or compute features. For legacy SLI-era software that explicitly targets Fermi dual-GPU setups, the GTX 590 still has a niche, but the data shows its lead is narrow and workload-specific.

FAQ

Q: Which card wins in raw compute, the GeForce GTX 1070 Ti or the GTX 590?

A: The GTX 1070 Ti wins the compute-heavy workloads. In the Geekbench OpenCL test, the GTX 1070 Ti scores 49,219 against 12,830 for the GTX 590, which is a 283.6% advantage. The Pascal architecture's much higher clock rates, at 1607 MHz base and 1683 MHz boost, drive this margin.

Q: How does the GTX 590 compare in memory bandwidth?

A: The GTX 590 has a 384-bit memory bus and 1536 MB of GDDR5, with a bandwidth of 164.0 GB/s. The GTX 1070 Ti has a slimmer 256-bit bus, 8 GB of GDDR5, and a much higher 256.3 GB/s bandwidth. In bandwidth-bound work, the newer card is 56.3% ahead in the recorded data.

Q: Is the GTX 590 a dual-GPU card, and does that matter?

A: Yes, it is a dual-GPU card with two GF110 dies. The database shows the GTX 590 with 512 shading units and 48 ROPs per GPU, doubling to 1024 shading units and 96 ROPs combined. The GTX 1070 Ti has 2432 shading units and 64 ROPs. Software and driver support for multi-GPU in modern APIs is poor, so the 590's second die is often idle.

Q: Which card has the higher memory clock?

A: The GTX 590's memory clock is 854 MHz, with an effective 3.4 Gbps. The GTX 1070 Ti's memory clock is 2002 MHz, or 8 Gbps effective. The newer card has a 25% higher effective memory clock, which helps its bandwidth advantage.

Q: How do the transistors and process node compare?

A: The GTX 1070 Ti has 7,200 million transistors on a 314 mm² die at TSMC's 16 nm node, for a density of 22.9M per mm². The GTX 590 uses 3,000 million transistors on a 520 mm² die at 40 nm, with a density of 5.8M per mm². The modern card is roughly 4x denser.

Q: What about API support?

A: The GTX 1070 Ti supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 590 supports DirectX 12 (11_0), OpenGL 4.6, and has no Vulkan support listed. The newer card has broader and more current API coverage.

Architecture Differences

The two cards come from different architecture families separated by nearly five years of chip design. The GTX 1070 Ti is built on GP104, a Pascal-generation chip fabricated at TSMC on a 16 nm process. The die is 314 mm², and it packs 7,200 million transistors, a density of 22.9M per mm². The GTX 590 is built on the GF110, a Fermi 2.0 chip from TSMC on a 40 nm node. Its die is 520 mm², holds 3,000 million transistors, and has a density of 5.8M per mm².

The GTX 1070 Ti has 2432 shaders, 152 texture mapping units, 64 ROPs, and no ray tracing or tensor cores. Its pixel rate is 107.7 GPixel/s and texture rate is 255.8 GTexel/s. The GTX 590 offers 512 shaders per GPU (1024 total), 64 TMUs, and 48 ROPs per GPU. Its pixel rate is 19.46 GPixel/s and texture rate is 38.91 GTexel/s. The newer card has roughly 12x the pixel fill rate and 6.6x the texture rate.

The GTX 1070 Ti uses a 267 mm card with a dual-slot cooler and one 8-pin power connector. The GTX 590 is a dual-slot card, 279 mm long, with two 8-pin connectors. The 1070 Ti's total board power is 180 W with a suggested 450 W power supply; the 590's board power is 365 W and recommends a 750 W power supply.

Specification Differences

| Field | NVIDIA GeForce GTX 1070 Ti | NVIDIA GeForce GTX 590 |

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

| Architecture | Pascal | Fermi 2.0 |

| Process node | 16 nm | 40 nm |

| Transistors | 7,200 million | 3,000 million |

| Die size | 314 mm² | 520 mm² |

| Transistor density | 22.9M / mm² | 5.8M / mm² |

| Base clock | 1607 MHz | Not listed |

| Boost clock | 1683 MHz | Not listed |

| Memory clock | 2002 MHz (8 Gbps effective) | 1070 MHz (3.4 Gbps effective) |

| Memory size | 8 GB | 1536 MB |

| Memory type | GDDR5 | GDDR5 |

| Bus width | 256 bit | 384 bit |

| Memory bandwidth | 256.3 GB/s | 164.0 GB/s |

| Shading units | 2432 | 512 (1024 total) |

| TMUs | 152 | 64 (128 total) |

| ROPs | 64 | 48 (96 total) |

| Pixel rate | 107.7 GPixel/s | 19.46 GPixel/s |

| Texture rate | 255.8 GTexel/s | 38.91 GTexel/s |

| FP32 | 8.186 TFLOPS | 1,244.2 GFLOPS |

| FP16 | 127.9 GFLOPS (1:64) | Not listed |

| TDP | 180 W | 365 W |

| Power connectors | 1x 8-pin | 2x 8-pin |

| Bus interface | PCIe 3.0 x16 | PCIe 2.0 x16 |

| Display outputs | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a | 3x DVI, 1x mini-DisplayPort |

| Release date | 2017-11-01 | 2011-03-23 |

| Production status | End-of-life | End-of-life |

| Predecessor | GeForce 900 | GeForce 400 |

| Successor | GeForce 20 | GeForce 600 |

Head-to-Head Benchmarks

The recorded benchmark data shows one clear head-to-head result. In the Geekbench OpenCL test, the GTX 1070 Ti scored 49,219 to 12,830 for the GTX 590. The result is a 283.6% win for the newer card. This is the largest and most decisive margin in the data set, and it reflects the 1070 Ti's massive lead in raw compute, memory bandwidth, and driver optimization.

The GTX 590 does not win any of the recorded head-to-head tests. Its average benchmark score of 12,830 places it near the bottom of the modern GPU range, while the 1070 Ti's average of 14,277 places it at the 56th percentile of all GPUs in the database. The 590's closest rivals in the database are the AMD Radeon Pro 455, with an average score of 12831, and the AMD FirePro W5100 with 12847, showing that the 590 slots into the performance range of much smaller, lower-power, single-GPU professional cards. By contrast, the 1070 Ti's nearest rivals are the Intel Iris Xe MAX at 14315, the AMD Radeon Vega 11 at 14352, the NVIDIA GTX TITAN at 14373, and the AMD Radeon RX Vega 11 at 14385, which is a much higher performance class.

Looking at the specification deltas, the 1070 Ti dominates almost every metric that matters for modern software: it has 3.8x the memory bandwidth, 5.5x the pixel fill rate, 6.6x the texture rate, 4x the shading units, and a 3.3x higher FP32 throughput (8.186 TFLOPS vs 1.244 TFLOPS). Its power draw is half of the 590's (180 W vs 365 W), its bus interface is a full generation newer (PCIe 3.0 vs 2.0), and it supports Vulkan, which the older card lacks entirely. The 590's only physical advantages are a wider 384-bit bus, more memory (1536 MB vs 8 GB, though the 1070 Ti has a higher effective clock), and more display outputs (4 vs 3). For modern content, the 1070 Ti is the faster, more efficient, and more feature-complete product. The 590 is a legacy product best suited for period-correct builds or software that was specifically multi-GPU aware from the Fermi era.