NVIDIA Tesla M2090
NVIDIA graphics card specifications and benchmark scores
At a Glance
NVIDIANVIDIA Tesla M2090 Specifications
GPU Core
Shader units and compute resources
The NVIDIA Tesla M2090 GPU core specifications define its raw processing power for graphics and compute workloads. Shading units (also called CUDA cores, stream processors, or execution units depending on manufacturer) handle the parallel calculations required for rendering. TMUs (Texture Mapping Units) process texture data, while ROPs (Render Output Units) handle final pixel output. Higher shader counts generally translate to better GPU benchmark performance, especially in demanding games and 3D applications.
Tesla M2090 Clock Speeds
GPU and memory frequencies
Clock speeds directly impact the Tesla M2090's performance in GPU benchmarks and real-world gaming. The base clock represents the minimum guaranteed frequency, while the boost clock indicates peak performance under optimal thermal conditions. Memory clock speed affects texture loading and frame buffer operations. The Tesla M2090 by NVIDIA dynamically adjusts frequencies based on workload, temperature, and power limits to maximize performance while maintaining stability.
NVIDIA's Tesla M2090 Memory
VRAM capacity and bandwidth
VRAM (Video RAM) is dedicated memory for storing textures, frame buffers, and shader data. The Tesla M2090's memory capacity determines how well it handles high-resolution textures and multiple displays. Memory bandwidth, measured in GB/s, affects how quickly data moves between the GPU and VRAM. Higher bandwidth improves performance in memory-intensive scenarios like 4K gaming. The memory bus width and type (GDDR6, GDDR6X, HBM) significantly influence overall GPU benchmark scores.
Tesla M2090 by NVIDIA Cache
On-chip cache hierarchy
On-chip cache provides ultra-fast data access for the Tesla M2090, reducing the need to fetch data from slower VRAM. L1 and L2 caches store frequently accessed data close to the compute units. AMD's Infinity Cache (L3) dramatically increases effective bandwidth, improving GPU benchmark performance without requiring wider memory buses. Larger cache sizes help maintain high frame rates in memory-bound scenarios and reduce power consumption by minimizing VRAM accesses.
Tesla M2090 Theoretical Performance
Compute and fill rates
Theoretical performance metrics provide a baseline for comparing the NVIDIA Tesla M2090 against other graphics cards. FP32 (single-precision) performance, measured in TFLOPS, indicates compute capability for gaming and general GPU workloads. FP64 (double-precision) matters for scientific computing. Pixel and texture fill rates determine how quickly the GPU can render complex scenes. While real-world GPU benchmark results depend on many factors, these specifications help predict relative performance levels.
Fermi 2.0 Architecture & Process
Manufacturing and design details
The NVIDIA Tesla M2090 is built on NVIDIA's Fermi 2.0 architecture, which defines how the GPU processes graphics and compute workloads. The manufacturing process node affects power efficiency, thermal characteristics, and maximum clock speeds. Smaller process nodes pack more transistors into the same die area, enabling higher performance per watt. Understanding the architecture helps predict how the Tesla M2090 will perform in GPU benchmarks compared to previous generations.
Power & Thermal
TDP and power requirements
Power specifications for the NVIDIA Tesla M2090 determine PSU requirements and thermal management needs. TDP (Thermal Design Power) indicates the heat output under typical loads, guiding cooler selection. Power connector requirements ensure adequate power delivery for stable operation during demanding GPU benchmarks. The suggested PSU wattage accounts for the entire system, not just the graphics card. Efficient power delivery enables the Tesla M2090 to maintain boost clocks without throttling.
Tesla M2090 by NVIDIA Physical & Connectivity
Dimensions and outputs
Physical dimensions of the NVIDIA Tesla M2090 are critical for case compatibility. Card length, height, and slot width determine whether it fits in your chassis. The PCIe interface version affects bandwidth for communication with the CPU. Display outputs define monitor connectivity options, with modern cards supporting multiple high-resolution displays simultaneously. Verify these specifications against your case and motherboard before purchasing to ensure a proper fit.
NVIDIA API Support
Graphics and compute APIs
API support determines which games and applications can fully utilize the NVIDIA Tesla M2090. DirectX 12 Ultimate enables advanced features like ray tracing and variable rate shading. Vulkan provides cross-platform graphics capabilities with low-level hardware access. OpenGL remains important for professional applications and older games. CUDA (NVIDIA) and OpenCL enable GPU compute for video editing, 3D rendering, and scientific applications. Higher API versions unlock newer graphical features in GPU benchmarks and games.
Tesla M2090 Product Information
Release and pricing details
The NVIDIA Tesla M2090 is manufactured by NVIDIA as part of their graphics card lineup. Release date and launch pricing provide context for comparing GPU benchmark results with competing products from the same era. Understanding the product lifecycle helps evaluate whether the Tesla M2090 by NVIDIA represents good value at current market prices. Predecessor and successor information aids in tracking generational improvements and planning future upgrades.
About NVIDIA Tesla M2090
The NVIDIA Tesla M2090 is a dual-slot compute card built on the 40 nm Fermi 2.0 architecture, using the GF110 chip with 3,000 million transistors on a 520 mm² die. It targets professional and server workloads rather than consumer gaming, and its data reflects that positioning: a single OpenCL benchmark score of 13,075 places it at the 52nd percentile of all GPUs, squarely in the middle of the pack for its era.
Power and Cooling
The Tesla M2090 carries a 250 W TDP, which is substantial for a card of its generation but manageable with the right power supply. The suggested PSU rating is 600 W, so builders should ensure their system has headroom beyond the card's own draw. Power delivery requires two connectors: one 6-pin and one 8-pin. This is a non-negotiable requirement — the card will not function without both attached, so verify your PSU's cable inventory before installation.
The card is dual-slot, meaning it occupies two expansion brackets in the chassis. Its physical length is 248 mm (9.8 inches), which fits most mid-tower cases but may be tight in smaller form-factor builds. There are no display outputs on this card — it is strictly a compute accelerator, not a graphics output device. That also means no cooling air is exhausted through a display bracket; the dual-slot cooler pushes heat out the rear of the case. The 250 W TDP generates significant heat, so proper case airflow is essential. The 40 nm process node means the die runs hot under sustained load, and the dual-slot design is the minimum viable cooling solution for this power envelope.
How It Compares
The nearest rivals by average benchmark score are a mixed group of mobile and desktop parts, and the Tesla M2090 sits almost exactly in the middle of them. Against the NVIDIA GeForce MX350, the delta is -0.1%, meaning the Tesla is essentially tied — a difference of 20 points out of roughly 13,000 is within noise. That is notable because the MX350 is a low-power mobile chip, while the Tesla is a 250 W server card; the architectural gulf is enormous, yet the OpenCL result is nearly identical.
The AMD FirePro M6100 scores 13,012, which is 0.5% behind the Tesla. This is another mobile workstation part, and again the Tesla's margin is razor-thin. The FirePro has the advantage of being a much newer design, but the Tesla's raw shader count of 512 and high memory bandwidth keep it competitive. The NVIDIA GeForce RTX 3050 Ti Mobile scores 12,940, putting it 1% behind the Tesla. This is a modern Ampere-based laptop GPU, and it trails the 2011 Tesla in this specific OpenCL workload — proof of the Tesla's compute-oriented design.
The strongest rival is the NVIDIA GeForce GTX 1660, which scores 13,265 and beats the Tesla by 1.4%. That is the only rival to exceed the Tesla's score, and it does so with a desktop Turing architecture that benefits from newer instruction sets. The Tesla's position is clear: it trades blows with entry-level and mid-range parts from much later generations, but it cannot outrun a dedicated desktop GPU from the same era's successor architecture.
Ray Tracing and Feature Set
The Tesla M2090 has no ray tracing cores and no tensor cores. It is a pure Fermi 2.0 design from 2011, predating both RTX and Tensor Core technology by several years. This means hardware-accelerated ray tracing is entirely absent; any ray tracing workload would have to run on the shader units, which is inefficient for real-time use.
API support is limited by its age. DirectX support is listed as 12 (11_0), meaning it can run DirectX 12 titles at the 11_0 feature level — effectively DirectX 11 functionality with a DX12 API wrapper. OpenGL support is 4.6, which is current and sufficient for most professional applications. There is no Vulkan support listed, which is a significant gap for modern compute and gaming workloads that rely on Vulkan's low-overhead access to the hardware.
The feature set is therefore entirely compute-focused. The 512 shading units, 64 texture mapping units, and 48 ROPs are the core of its processing power. Pixel rate is 20.83 GPixel/s and texture rate is 41.66 GTexel/s, both respectable for the era but dwarfed by modern parts. FP32 performance is 1,332.2 GFLOPS, which is the headline compute figure for this card. There is no FP16 support listed, so half-precision workloads are not accelerated. For the intended scientific and HPC use case, the FP32 throughput is what matters, and it is competitive within its generation.
FAQ
Q: Does the Tesla M2090 support ray tracing?
A: No. It has no ray tracing cores and no tensor cores. It is a Fermi 2.0 architecture GPU from 2011, and hardware ray tracing was not introduced until much later generations.
Q: What power supply do I need for this card?
A: The suggested PSU rating is 600 W. The card itself has a 250 W TDP and requires both a 6-pin and an 8-pin power connector.
Q: Can I use this card for gaming with modern titles?
A: DirectX support is 12 (11_0), which means it can run DX12 games at the 11_0 feature level — effectively DX11 functionality. There is no Vulkan support, so modern Vulkan-only games will not run. It also has no display outputs, so it cannot drive a monitor directly.
Q: How does it compare to a GeForce GTX 1660?
A: The GTX 1660 scores 13,265 in OpenCL, which is 1.4% higher than the Tesla's 13,075. The GTX 1660 is the closest rival that beats the Tesla.
Q: What is the memory configuration?
A: The Tesla M2090 has 6 GB of GDDR5 memory on a 384-bit bus, with 177.4 GB/s of bandwidth. Memory clock is 924 MHz, which translates to 3.7 Gbps effective.
Q: Is this card still in production?
A: No. Production status is end-of-life. It was released on 2011-07-24 and has been succeeded by the Tesla Kepler generation.
Benchmark Performance
The single benchmark result for the Tesla M2090 is 13,075 in Geekbench OpenCL. That score places it at the 52nd percentile of all GPUs, meaning it sits just above the median — half of all tested GPUs are slower, half are faster. For a compute card from 2011, that is a strong showing, but it is not exceptional by modern standards.
The nearest rival comparisons are tight. The Tesla is 0.1% behind the GeForce MX350 (13,095 vs 13,075), a margin of 20 points. This is effectively a statistical tie. The MX350 is a low-TDP mobile chip, so the Tesla's equal performance despite a 250 W TDP and much larger die suggests the OpenCL workload is not fully utilizing the Tesla's compute resources, or the MX350's newer architecture is more efficient per watt.
Against the AMD FirePro M6100, the Tesla leads by 0.5% (13,075 vs 13,012). The FirePro is a mobile workstation GPU, and the Tesla's 63-point advantage is small but consistent. The RTX 3050 Ti Mobile trails by 1% (12,940), a 135-point deficit. This is the most surprising result — a modern Ampere laptop GPU losing to a Fermi compute card. The GTX 1660 is the only rival ahead, beating the Tesla by 1.4% (13,265 vs 13,075), a 190-point gap.
The pattern is clear: the Tesla M2090 holds its own against entry-level and mid-range parts from 2018-2021 in this specific OpenCL test. Its FP32 throughput of 1,332.2 GFLOPS is the likely driver, as OpenCL compute workloads scale well with raw shader count. The 512 shading units are numerous for the era, and the 250 W TDP gives them room to run at high clocks. However, the lack of newer features like tensor cores and Vulkan support means this performance is not portable to modern gaming or AI workloads. The 52nd percentile ranking reflects a card that is competent but not competitive with dedicated high-end parts from any generation after its own.
Memory Subsystem
The memory configuration is a 6 GB pool of GDDR5 on a 384-bit bus, running at 924 MHz with an effective data rate of 3.7 Gbps. Total bandwidth is 177.4 GB/s. This is a wide bus design — 384 bits is the same width used by many high-end consumer cards of the era, and it provides substantial parallel access to the framebuffer.
For high-resolution workloads, 6 GB of VRAM is adequate for 1080p and 1440p compute tasks, but it is limiting for 4K textures and large datasets. The bandwidth of 177.4 GB/s is modest by modern standards — current cards exceed 500 GB/s — but it is well-matched to the FP32 compute rate. The memory clock of 924 MHz is low compared to later GDDR5 implementations, but the wide bus compensates: 384 bits at 3.7 Gbps effective yields the 177.4 GB/s figure.
The 48 ROPs are the pixel output stage, and the pixel rate of 20.83 GPixel/s is derived from the ROP count and clock. At high resolutions, the ROPs can become a bottleneck if the workload is fill-rate limited. The texture rate of 41.66 GTexel/s, driven by 64 TMUs, is more than sufficient for most compute tasks that rely on texture fetch operations. The memory subsystem is balanced for its generation — not a standout, but not a weak point. The 6 GB capacity is the most future-proof aspect, as it allows larger datasets to reside in VRAM without spilling to system memory. However, the 177.4 GB/s bandwidth will be the limiting factor for memory-bound workloads at 4K or above.
Detailed benchmark scores and charts for the NVIDIA Tesla M2090 are below.
Benchmark Scores
geekbench_openclSource
Geekbench OpenCL tests GPU compute performance using the cross-platform OpenCL API. This shows how NVIDIA Tesla M2090 handles parallel computing tasks like video encoding and scientific simulations.
Top 5 Performers
Top 5 PerformersNearby Performers
Nearby PerformersPopular NVIDIA Tesla M2090 Comparisons
See how the Tesla M2090 stacks up against similar graphics cards from the same generation and competing brands.
Compare with Other GPUs
Select another GPU to compare specifications and benchmarks side-by-side.
Browse GPUs