NVIDIA GeForce GTX 1080 vs NVIDIA Tesla M2090 Comparison

# NVIDIA Tesla M2090 vs NVIDIA GeForce GTX 1080

The NVIDIA Tesla M2090 and GeForce GTX 1080 are generations apart, and the benchmark data reflects that chasm clearly. In the single shared test — Geekbench OpenCL — the GTX 1080 scores 51,204 against the Tesla M2090’s 13,075, a 74.5% advantage for the Pascal card. The Tesla M2090’s average benchmark score of 13,075 places it at the 53rd percentile of all GPUs, while the GTX 1080’s average of 11,960 sits at the 51st percentile — a quirk explained by the GTX 1080’s multiple benchmark entries, including lower DirectX and 2D scores that drag its average down. For raw compute throughput, however, there is no contest.

Head-to-Head Benchmarks

The only directly comparable result is Geekbench OpenCL, and it is decisive. The GTX 1080 delivers 51,204 points, which is 4.1 times the Tesla M2090’s 13,075. That translates to a 74.5% delta in favor of the newer card. For context, the Tesla M2090’s nearest rivals in the database are the GeForce GTX 1660 SUPER (12,986, just 0.7% lower), the GeForce GTX 950 (13,189, 0.9% higher), the RTX 3050 Ti Mobile (12,940, 1% lower), and the AMD Radeon RX 580 (12,928, 1.1% lower). The M2090 is essentially a contemporary of those cards in OpenCL compute, despite being from 2011.

The GTX 1080’s nearest rivals tell a different story. Its average score of 11,960 is just 0.3% above the GeForce GTX 960A (11,998) and 1% above the AMD Radeon RX 6500 XT (11,842). It is 2.4% ahead of the GeForce GTX 1660 (11,680) and 2.9% ahead of the AMD Radeon RX 7800 XT (11,627). These deltas are small, meaning the GTX 1080’s average is pulled down by its varied test suite, even though its OpenCL result dwarfs the Tesla.

Other GTX 1080 results show its strengths and weaknesses. It scores 15,586 in Passmark G3D, 7,614 in Passmark GPU Compute, 30,398 in Geekbench Vulkan, and 23,824 in Geekbench Metal. Its DirectX 12 Passmark score is a modest 55, while DirectX 11 reaches 124 and DirectX 9 hits 211. The 2D score of 888 is respectable but not class-leading. The M2090 has no comparable results in these tests, so the only apples-to-apples comparison is OpenCL — and that single data point is enough to establish the hierarchy.

Architecture Differences

The Tesla M2090 is built on the Fermi 2.0 architecture with the GF110 chip, fabricated on TSMC’s 40 nm process. It packs 3,000 million transistors onto a 520 mm² die, yielding a transistor density of 5.8 million per square millimeter. The GTX 1080 uses the Pascal architecture with the GP104 chip on a 16 nm process, integrating 7,200 million transistors into a 314 mm² die — a density of 22.9 million per square millimeter. That is a 4x jump in density and a 2.4x increase in total transistors, all while shrinking the physical die by more than a third.

Shading units scale dramatically: the M2090 has 512, while the GTX 1080 has 2,560 — a 5x difference. Texture mapping units go from 64 to 160, and ROPs from 48 to 64. Clock speeds are not directly comparable because the M2090 lists no base or boost clock, only a memory clock of 924 MHz (3.7 Gbps effective). The GTX 1080 runs at a 1,607 MHz base and 1,733 MHz boost, with memory at 1,251 MHz (10 Gbps effective). Memory capacity rises from 6 GB GDDR5 to 8 GB GDDR5X, while the bus width narrows from 384-bit to 256-bit. Bandwidth still climbs from 177.4 GB/s to 320.3 GB/s thanks to faster memory.

Pixel rate jumps from 20.83 GPixel/s on the M2090 to 110.9 GPixel/s on the GTX 1080. Texture rate goes from 41.66 GTexel/s to 277.3 GTexel/s. FP32 compute is listed at 1,332.2 GFLOPS for the Tesla and 8.873 TFLOPS for the GTX 1080 — a 6.7x gap. The GTX 1080 also lists FP16 performance of 138.6 GFLOPS (1:64), while the M2090 has no FP16 entry. Power draw is lower on the newer card: 180 W vs 250 W, with a suggested PSU of 450 W vs 600 W. The GTX 1080 uses a single 8-pin connector; the M2090 needs 1x 6-pin plus 1x 8-pin.

Interface and outputs differ too. The Tesla uses PCIe 2.0 x16 and has no display outputs — it is a compute card. The GTX 1080 uses PCIe 3.0 x16 and offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a. API support favors the GTX 1080: it lists DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while the M2090 lists DirectX 12 (11_0), OpenGL 4.6, and no Vulkan. Physical dimensions also diverge: the M2090 is 248 mm long, the GTX 1080 is 267 mm long, 112 mm tall, and 40 mm wide.

The Verdict

The data points to one conclusion for most buyers: the GTX 1080 is the superior card in every measurable way. Its OpenCL score is 74.5% higher, its architecture is four generations newer, and it consumes 70 W less power while delivering more than 6x the FP32 throughput. For anyone needing display outputs, the M2090 is disqualified immediately — it has none. For compute workloads, the GTX 1080’s 8 GB GDDR5X and 320.3 GB/s bandwidth give it a practical edge over the M2090’s 6 GB GDDR5 and 177.4 GB/s.

The Tesla M2090 does have one niche: it sits at the 53rd percentile of all GPUs, slightly above the GTX 1080’s 51st percentile. That is a statistical artifact of the GTX 1080’s broader test suite, but it does mean the M2090’s single OpenCL score is competitive with much newer budget cards like the GTX 1660 SUPER and RX 580. If a workload is specifically tuned for Fermi’s compute capabilities and does not require modern APIs, the M2090 could still hold its own. But that is a narrow use case. The GTX 1080 was released in May 2016 with a launch MSRP of 599 USD, while the M2090 arrived in July 2011. The five-year gap shows in every specification.

Specification Differences

• Process node: M2090 is 40 nm; GTX 1080 is 16 nm.
• Transistors: M2090 has 3,000 million; GTX 1080 has 7,200 million.
• Die size: M2090 is 520 mm²; GTX 1080 is 314 mm².
• Transistor density: M2090 is 5.8M / mm²; GTX 1080 is 22.9M / mm².
• Shading units: M2090 has 512; GTX 1080 has 2,560.
• TMUs: M2090 has 64; GTX 1080 has 160.
• ROPs: M2090 has 48; GTX 1080 has 64.
• Memory size: M2090 has 6 GB; GTX 1080 has 8 GB.
• Memory type: M2090 is GDDR5; GTX 1080 is GDDR5X.
• Memory bus: M2090 is 384-bit; GTX 1080 is 256-bit.
• Memory bandwidth: M2090 is 177.4 GB/s; GTX 1080 is 320.3 GB/s.
• Pixel rate: M2090 is 20.83 GPixel/s; GTX 1080 is 110.9 GPixel/s.
• Texture rate: M2090 is 41.66 GTexel/s; GTX 1080 is 277.3 GTexel/s.
• FP32: M2090 is 1,332.2 GFLOPS; GTX 1080 is 8.873 TFLOPS.
• TDP: M2090 is 250 W; GTX 1080 is 180 W.
• Power connectors: M2090 is 1x 6-pin + 1x 8-pin; GTX 1080 is 1x 8-pin.
• Suggested PSU: M2090 is 600 W; GTX 1080 is 450 W.
• Bus interface: M2090 is PCIe 2.0 x16; GTX 1080 is PCIe 3.0 x16.
• Display outputs: M2090 has none; GTX 1080 has 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.4a.
• DirectX support: M2090 is 12 (11_0); GTX 1080 is 12 (12_1).
• Vulkan: M2090 has none; GTX 1080 has 1.4.
• Length: M2090 is 248 mm; GTX 1080 is 267 mm.
• Release date: M2090 is July 2011; GTX 1080 is May 2016.

FAQ

Q: Which card has a higher OpenCL score?

A: The GTX 1080 scores 51,204 in Geekbench OpenCL, which is 74.5% higher than the M2090’s 13,075.

Q: Does the Tesla M2090 support modern graphics APIs?

A: It supports DirectX 12 (11_0) and OpenGL 4.6, but it has no Vulkan support. The GTX 1080 adds Vulkan 1.4 and DirectX 12 (12_1).

Q: Can the M2090 be used for display output?

A: No. The M2090 has no display outputs, while the GTX 1080 includes DVI, HDMI 2.0, and three DisplayPort 1.4a connectors.

Q: How do their power requirements compare?

A: The M2090 draws 250 W and recommends a 600 W PSU, while the GTX 1080 draws 180 W and recommends a 450 W PSU. The M2090 also needs both a 6-pin and an 8-pin connector, whereas the GTX 1080 uses only one 8-pin.

Q: Which card has more memory bandwidth?

A: The GTX 1080 has 320.3 GB/s, nearly double the M2090’s 177.4 GB/s, despite using a narrower 256-bit bus versus the M2090’s 384-bit bus.

Q: How do their transistor counts compare?

A: The GTX 1080 has 7,200 million transistors on a 16 nm process, while the M2090 has 3,000 million on 40 nm. The GTX 1080’s die is smaller (314 mm² vs 520 mm²) but far denser.

Where Each One Wins

The GTX 1080 wins in every scenario that involves modern workloads, display output, or energy efficiency. Its 8 GB GDDR5X memory and 320.3 GB/s bandwidth handle large textures and datasets far better than the M2090’s 6 GB GDDR5. The 2,560 shading units and 8.873 TFLOPS FP32 performance make it suitable for compute tasks, gaming, and content creation alike. Its Vulkan 1.4 and DirectX 12 (12_1) support ensure compatibility with current software. The lower 180 W TDP and single 8-pin connector mean it fits into more builds without PSU upgrades.

The M2090’s only statistical win is its 53rd percentile ranking versus the GTX 1080’s 51st, which is misleading given the GTX 1080’s broader test results. In practice, the M2090 is best suited for legacy Fermi-specific compute workloads that do not require modern APIs or display output. Its 177.4 GB/s bandwidth and 1,332.2 GFLOPS are adequate for older scientific or enterprise applications, and its 6 GB frame buffer is still useful for certain data-parallel tasks. But for anyone building a system today, the GTX 1080 is the clear choice — the benchmark data shows a 74.5% lead in the one test they share, and every architectural advantage favors the Pascal card.