NVIDIA GeForce GTX 1070 Ti vs NVIDIA Tesla M2090 Comparison

# NVIDIA Tesla M2090 vs NVIDIA GeForce GTX 1070 Ti

The NVIDIA Tesla M2090 and NVIDIA GeForce GTX 1070 Ti represent two vastly different eras of GPU design, separated by six years of architectural evolution. The Tesla M2090, a Fermi 2.0 compute card from 2011, was built for professional workloads with no display outputs, while the GTX 1070 Ti, a Pascal gaming card from 2017, delivers modern API support and a benchmark score nearly four times higher in OpenCL. The data reveals a generational chasm: the GTX 1070 Ti outscores the M2090 by 36144 points in Geekbench OpenCL, a delta of -73.4% from the newer card's perspective. Yet the M2090 still holds a 53rd percentile ranking among all GPUs, showing that even a decade-old compute card can remain relevant in specific niches.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 1070 Ti has an average benchmark score of 14277, while the NVIDIA Tesla M2090 scores 13075. This places the GTX 1070 Ti in the 56th percentile of all GPUs, compared to the M2090's 53rd percentile.

Q: How do the two cards compare in raw OpenCL performance?

A: In the Geekbench OpenCL test, the GTX 1070 Ti scores 49219, which is 36144 points higher than the M2090's 13075. This represents a 276% performance advantage for the newer card, with the deltaPct showing -73.4% when viewed from the GTX 1070 Ti's perspective.

Q: What are the memory specifications of each card?

A: The Tesla M2090 features 6 GB of GDDR5 memory on a 384-bit bus, delivering 177.4 GB/s bandwidth. The GTX 1070 Ti has 8 GB of GDDR5 on a 256-bit bus, achieving 256.3 GB/s bandwidth despite the narrower interface.

Q: Which card supports modern graphics APIs?

A: The GTX 1070 Ti supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The Tesla M2090 only reaches DirectX 12 (11_0) and OpenGL 4.6, with no Vulkan support listed.

Q: How do the transistor counts and die sizes compare?

A: The GTX 1070 Ti packs 7200 million transistors on a 314 mm² die using TSMC's 16 nm process, achieving a density of 22.9M transistors per mm². The M2090 uses 3000 million transistors on a much larger 520 mm² die at 40 nm, with only 5.8M transistors per mm².

Q: What is the power consumption difference?

A: The Tesla M2090 has a TDP of 250 W and requires a 600 W suggested PSU with 1x 6-pin and 1x 8-pin connectors. The GTX 1070 Ti draws 180 W, needs only a 450 W PSU, and uses a single 8-pin connector.

Architecture Differences

The architectural gap between these two NVIDIA cards is enormous, reflecting six years of process and design advancement. The Tesla M2090 is built on Fermi 2.0 architecture using TSMC's 40 nm process node, featuring a GF110 chip with 3000 million transistors spread across a massive 520 mm² die. This yields a transistor density of just 5.8M per mm², a figure that seems primitive by modern standards. In contrast, the GTX 1070 Ti employs the Pascal architecture on a 16 nm process with the GP104 chip, fitting 7200 million transistors into a much smaller 314 mm² die — achieving 22.9M transistors per mm², nearly four times the density.

The compute resources differ dramatically. The M2090 has 512 shading units, 64 texture mapping units, and 48 ROPs. The GTX 1070 Ti scales this up to 2432 shading units, 152 TMUs, and 64 ROPs — 4.75 times more shaders and 2.375 times more texture units. Clock speeds tell a similar story: the M2090's memory runs at 924 MHz (3.7 Gbps effective), while the GTX 1070 Ti's memory operates at 2002 MHz (8 Gbps effective). The GTX 1070 Ti also has explicit base and boost clocks of 1607 MHz and 1683 MHz, while the M2090 lists no base or boost clock.

Memory architecture diverges in an interesting way. The M2090 uses a wider 384-bit bus with 6 GB of GDDR5, while the GTX 1070 Ti narrows to 256-bit but compensates with faster memory and 8 GB capacity. The result: the GTX 1070 Ti achieves 256.3 GB/s bandwidth versus 177.4 GB/s for the M2090 — a 44% advantage despite using fewer memory channels. The M2090's compute capabilities peak at 1,332.2 GFLOPS FP32, while the GTX 1070 Ti reaches 8.186 TFLOPS, over six times higher. The GTX 1070 Ti also lists FP16 performance at 127.9 GFLOPS at a 1:64 ratio, while the M2090 has no FP16 specification.

The Verdict

The data overwhelmingly favors the GTX 1070 Ti for nearly any use case. In the only head-to-head benchmark available, Geekbench OpenCL, the GTX 1070 Ti delivers 49219 points versus the M2090's 13075 — a 276% advantage. The newer card also offers modern API support including Vulkan 1.4 and DirectX 12 (12_1), while the M2090 is stuck at DirectX 12 (11_0) with no Vulkan. For gaming, the GTX 1070 Ti has display outputs (1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a), while the M2090 has none — it is a compute-only card. The GTX 1070 Ti also wins on efficiency: 180 W TDP versus 250 W, and a lower suggested PSU of 450 W versus 600 W.

The Tesla M2090's only theoretical advantages are its wider 384-bit memory bus and its 48 ROPs, though these do not translate into benchmark wins. Its 53rd percentile ranking shows it still outperforms many GPUs, but its nearest rivals — the GTX 1660 SUPER (12986, 0.7% delta) and GTX 950 (13189, -0.9% delta) — are all newer consumer cards. The GTX 1070 Ti's nearest rivals include the GTX TITAN (14373, -0.7% delta) and AMD Radeon RX Vega 11 (14385, -0.7% delta), showing it competes at a much higher tier. The verdict is clear: the GTX 1070 Ti is the superior choice for virtually any workload, with the M2090 remaining only as a historical compute artifact.

Specification Differences

| Specification | NVIDIA Tesla M2090 | NVIDIA GeForce GTX 1070 Ti |

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

| Architecture | Fermi 2.0 | Pascal |

| Process Node | 40 nm | 16 nm |

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

| Die Size | 520 mm² | 314 mm² |

| Transistor Density | 5.8M / mm² | 22.9M / mm² |

| Memory | 6 GB GDDR5 | 8 GB GDDR5 |

| Memory Bus | 384 bit | 256 bit |

| Memory Bandwidth | 177.4 GB/s | 256.3 GB/s |

| Memory Clock | 924 MHz (3.7 Gbps) | 2002 MHz (8 Gbps) |

| Shading Units | 512 | 2432 |

| TMUs | 64 | 152 |

| ROPs | 48 | 64 |

| Base Clock | N/A | 1607 MHz |

| Boost Clock | N/A | 1683 MHz |

| FP32 Performance | 1,332.2 GFLOPS | 8.186 TFLOPS |

| Pixel Rate | 20.83 GPixel/s | 107.7 GPixel/s |

| Texture Rate | 41.66 GTexel/s | 255.8 GTexel/s |

| TDP | 250 W | 180 W |

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

| Suggested PSU | 600 W | 450 W |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Display Outputs | No outputs | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a |

| DirectX | 12 (11_0) | 12 (12_1) |

| Vulkan | N/A | 1.4 |

| Release Date | 2011-07-24 | 2017-11-01 |

| Length | 248 mm (9.8 inches) | 267 mm (10.5 inches) |

Head-to-Head Benchmarks

The only direct benchmark comparison available is Geekbench OpenCL, and it tells a decisive story. The GTX 1070 Ti scores 49219, while the Tesla M2090 manages just 13075. The deltaPct of -73.4% indicates the M2090 trails by nearly three-quarters of the newer card's score. This single result encapsulates the generational leap: the GTX 1070 Ti's score is 3.76 times higher than the M2090's.

Beyond the head-to-head, the GTX 1070 Ti has a rich benchmark profile that the M2090 lacks entirely. The GTX 1070 Ti scores 61006 in Geekbench Vulkan, 22180 in Geekbench Metal, and 14674 in Passmark G3D. Its Passmark GPU compute score is 7165, while its DirectX 9 performance reaches 207. The M2090 has no equivalent data for these tests, meaning the comparison is limited to a single OpenCL data point.

The percentile rankings provide additional context. The M2090 sits at the 53rd percentile of all GPUs, while the GTX 1070 Ti ranks at the 56th percentile. This narrow gap in percentile, despite the massive OpenCL score difference, suggests that the M2090's compute-oriented design still holds up well against the broader GPU landscape. However, the average benchmark scores show a clearer picture: the GTX 1070 Ti averages 14277 across all its benchmarks, while the M2090 averages 13075 from its single test.

Where Each One Wins

The GTX 1070 Ti wins in every measurable category from the data. In raw compute, it dominates with 8.186 TFLOPS FP32 versus 1,332.2 GFLOPS. Its pixel rate of 107.7 GPixel/s is over five times the M2090's 20.83 GPixel/s, and its texture rate of 255.8 GTexel/s dwarfs the 41.66 GTexel/s of the older card. The GTX 1070 Ti also wins on efficiency, delivering vastly higher performance at 180 W compared to the M2090's 250 W.

For gaming and general consumer use, the GTX 1070 Ti is the only viable option — the M2090 has no display outputs, making it impossible to connect a monitor. The GTX 1070 Ti's modern API support (Vulkan 1.4, DirectX 12_1) ensures compatibility with contemporary software, while the M2090's DirectX 12 (11_0) and lack of Vulkan limit its software ecosystem. The GTX 1070 Ti's 8 GB memory capacity and higher bandwidth (256.3 GB/s) also make it more suitable for modern workloads.

The Tesla M2090's only potential advantages are its wider 384-bit memory bus and its end-of-life status as a dedicated compute card. Its 6 GB GDDR5 memory might theoretically benefit from the wider bus in specific memory-latency-sensitive workloads, but this does not translate into any benchmark victory. The M2090's 53rd percentile ranking, while respectable, places it alongside the GTX 950 and GTX 1660 SUPER — all consumer cards from later generations. The data shows no workload where the M2090 outperforms the GTX 1070 Ti; the newer card wins the sole head-to-head benchmark and offers superior specifications across every architectural metric. For anyone choosing between these two, the GTX 1070 Ti is the definitive pick, with the M2090 serving only as a historical reference point for Fermi-era compute performance.