NVIDIA GeForce GTX 590 vs NVIDIA Tesla M2090 Comparison

The NVIDIA Tesla M2090 and NVIDIA GeForce GTX 590 are both built on the same GF110 chip, yet they serve entirely different purposes. The data shows two cards with nearly identical compute resources but divergent priorities: one optimized for sustained computation, the other for consumer graphics. This analysis breaks down the benchmark results, architectural differences, and practical implications of each.

Head-to-Head Benchmarks

The only shared benchmark in the data is Geekbench OpenCL, and the margin is razor-thin. The Tesla M2090 scores 13075, while the GTX 590 scores 12830. That is a delta of 1.9% in favor of the Tesla. In practical terms, this is almost a statistical tie. Both cards land within a whisker of each other, and the percentile rankings confirm it: the M2090 sits at the 53rd percentile of all GPUs, the GTX 590 at the 52nd.

To contextualize, the M2090's nearest rivals include the GeForce GTX 1660 SUPER (12986, delta 0.7%), the GeForce GTX 950 (13189, delta -0.9%), and the Radeon RX 580 (12928, delta 1.1%). The GTX 590's rivals are similarly clustered: the Radeon Pro 455 (12831, delta 0%), the FirePro W5100 (12847, delta -0.1%), and the GeForce GTX 670 (12773, delta 0.4%). Neither card dramatically outperforms its peers; they are all within a couple of percent of each other.

The real story is that the M2090 wins the only direct comparison, but the 1.9% advantage is smaller than the variance seen between its own rival cards. The GTX 590 is essentially tied with the Radeon Pro 455 at a 0% delta, making it indistinguishable from that competitor in this workload. The M2090's win is consistent but not dominant. It holds a 1% lead over the RTX 3050 Ti Mobile and a 1.1% lead over the RX 580. The GTX 590 trails the Radeon 740M by 0.3% and the FirePro W5100 by 0.1%. In raw numbers, the M2090 is 245 points ahead of the GTX 590. That is a real gap, but not a chasm.

Architecture Differences

Both cards share the same fundamental DNA. They use the GF110 chip, the Fermi 2.0 architecture, a 40 nm process node from TSMC, 3,000 million transistors, and a 520 mm² die size. The shading units, texture mapping units, and raster output units are identical: 512, 64, and 48 respectively. Neither has ray tracing or tensor cores. Both support DirectX 12 (11_0) and OpenGL 4.6, with no Vulkan support listed. The bus interface is PCIe 2.0 x16 on both.

The differences emerge in implementation. The Tesla M2090 runs its memory at 924 MHz, yielding 3.7 Gbps effective, while the GTX 590 runs at 854 MHz, or 3.4 Gbps effective. This clocks the M2090's memory bandwidth to 177.4 GB/s versus 164.0 GB/s for the GTX 590. The memory capacity is a massive divergence: the M2090 has 6 GB of GDDR5, the GTX 590 has only 1536 MB. Both use a 384-bit bus, but the M2090's larger frame buffer and higher memory clock give it a clear edge in data-intensive workloads.

Compute rates follow the memory trend. The M2090 achieves 1,332.2 GFLOPS FP32, 20.83 GPixel/s pixel rate, and 41.66 GTexel/s texture rate. The GTX 590 is lower across the board: 1,244.2 GFLOPS, 19.46 GPixel/s, and 38.91 GTexel/s. The GTX 590 is not a slower chip per se; it uses the same GF110. But its clocks are tuned down, likely for thermal and power reasons. The GTX 590 draws 365 W TDP versus 250 W for the M2090, and requires a 750 W power supply versus 600 W. The GTX 590 uses 2x 8-pin power connectors, the M2090 uses 1x 6-pin + 1x 8-pin.

Physically, the GTX 590 is larger: 279 mm (11 inches) long, 111 mm (4.4 inches) tall, and 40 mm (1.6 inches) wide. The M2090 is 248 mm (9.8 inches) long, with height and width not specified. Both are dual-slot cards. The GTX 590 has display outputs (3x DVI and 1x mini-DisplayPort), while the M2090 has no outputs at all. That is a critical distinction: the M2090 is a compute card, not a graphics card.

FAQ

Q: Which card has higher raw compute performance?

A: The Tesla M2090 leads in FP32 with 1,332.2 GFLOPS versus 1,244.2 GFLOPS for the GTX 590. It also has higher pixel and texture rates: 20.83 GPixel/s versus 19.46 GPixel/s, and 41.66 GTexel/s versus 38.91 GTexel/s.

Q: Why does the GTX 590 have a higher TDP despite lower clocks?

A: The data does not specify a direct cause, but the GTX 590 is a dual-GPU consumer card with a 365 W TDP and 2x 8-pin connectors. The M2090, a single compute card, has a 250 W TDP and uses 1x 6-pin + 1x 8-pin. The GTX 590 also requires a 750 W power supply versus 600 W.

Q: Can the GTX 590 be used for display output?

A: Yes. The GTX 590 has 3x DVI and 1x mini-DisplayPort outputs. The Tesla M2090 has no display outputs, meaning it is intended for compute-only workloads.

Q: How much memory does each card have?

A: The Tesla M2090 has 6 GB of GDDR5, while the GTX 590 has 1536 MB (1.5 GB) of GDDR5. Both use a 384-bit memory bus, but the M2090's bandwidth is 177.4 GB/s versus 164.0 GB/s.

Q: What is the benchmark score difference?

A: In Geekbench OpenCL, the M2090 scores 13075 and the GTX 590 scores 12830. The M2090 wins by 1.9%. This places the M2090 at the 53rd percentile of all GPUs and the GTX 590 at the 52nd.

Q: Are these cards from the same generation?

A: Both use the GF110 chip and Fermi 2.0 architecture, but they belong to different product lines. The M2090 is from the Tesla Fermi (x20xx) generation, released 2011-07-24. The GTX 590 is from the GeForce 500 generation, released 2011-03-23.

Specification Differences

| Specification | NVIDIA Tesla M2090 | NVIDIA GeForce GTX 590 |

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

| Generation | Tesla Fermi (x20xx) | GeForce 500 |

| Memory Clock | 924 MHz (3.7 Gbps effective) | 854 MHz (3.4 Gbps effective) |

| Memory Size | 6 GB | 1536 MB |

| Memory Bandwidth | 177.4 GB/s | 164.0 GB/s |

| Pixel Rate | 20.83 GPixel/s | 19.46 GPixel/s |

| Texture Rate | 41.66 GTexel/s | 38.91 GTexel/s |

| FP32 | 1,332.2 GFLOPS | 1,244.2 GFLOPS |

| TDP | 250 W | 365 W |

| Power Connectors | 1x 6-pin + 1x 8-pin | 2x 8-pin |

| Suggested PSU | 600 W | 750 W |

| Display Outputs | No outputs | 3x DVI, 1x mini-DisplayPort |

| Dimensions | 248 mm (9.8 inches) length | 279 mm (11 inches) length, 111 mm (4.4 inches) height, 40 mm (1.6 inches) width |

| Release Date | 2011-07-24 | 2011-03-23 |

| Predecessor | Tesla | GeForce 400 |

| Successor | Tesla Kepler | GeForce 600 |

| Launch MSRP | Not listed | 699 USD |

Where Each One Wins

The Tesla M2090 wins in every compute metric that matters for heavy workloads. It has more memory (6 GB versus 1536 MB), higher bandwidth (177.4 GB/s versus 164.0 GB/s), higher FP32 (1,332.2 GFLOPS versus 1,244.2 GFLOPS), and higher pixel and texture rates. It also has a lower TDP at 250 W versus 365 W, which is counterintuitive given its higher performance. The M2090 is the card for scientific computing, data processing, or any task that needs large datasets in memory. Its 6 GB frame buffer is four times larger, allowing it to hold far more data locally. The lack of display outputs is irrelevant in a server or compute node. The M2090 also wins the only head-to-head benchmark, albeit by 1.9%.

The GTX 590 wins in the consumer graphics domain by default. It has display outputs and a 699 USD launch MSRP, making it a viable product for a desktop workstation or gaming rig at the time. It is smaller in memory but still has 512 shading units and 64 TMUs. Its lower FP32 and bandwidth are a trade-off for a more balanced consumer feature set. The GTX 590 is the choice if the workload involves rendering to a screen, driving multiple monitors (3x DVI, 1x mini-DisplayPort), or playing games from 2011. It also has a longer card length, which may matter for case compatibility, but the data does not indicate a performance advantage from that.

The data does not show a scenario where the GTX 590 beats the M2090 in raw compute. The M2090 is ahead in every measurable performance metric. The GTX 590's only advantages are its display outputs, its launch MSRP (which is a price point, not a performance metric), and its earlier release date. For a user who needs compute, the M2090 is strictly better per the benchmark and spec data. For a user who needs graphics output, the GTX 590 is the only option of the two. The choice is not about performance parity; it is about intended use case. The M2090 is a professional compute accelerator. The GTX 590 is a dual-GPU consumer card that happens to use the same chip.