NVIDIA GeForce MX350 vs NVIDIA Tesla K20c Comparison

# Head-to-Head Benchmarks

The only shared benchmark between these two GPUs is Geekbench OpenCL, and the result is decisive. The NVIDIA Tesla K20c scores 11,479, while the GeForce MX350 manages 8,689 — a 32.1% advantage for the Tesla. That is not a marginal gap; it is a generational chasm in raw compute throughput, reflecting the Tesla's 2,496 shading units against the MX350's 640.

In context, the Tesla K20c's score places it within 0.4% of the AMD Radeon Pro 5500M (11,528) and 1.3% behind the AMD Radeon RX 7800 XT (11,627). It also edges out the NVIDIA GeForce GTX 780M by 1.9% (11,261). The MX350, meanwhile, sits just 0.7% above the AMD Radeon Pro 450 (10,804) and 1.1% ahead of the NVIDIA Quadro K2200 (10,761), but trails the NVIDIA GeForce GTX 1650 SUPER by 1.5% (11,047) and the AMD Radeon RX 550 by 1.7% (11,075).

The Tesla also holds a second benchmark result in its favor: the Geekbench Vulkan test. Here, the MX350 scores 13,077, but the Tesla has no Vulkan result listed, so no direct comparison is possible. Still, in the one head-to-head metric available, the Tesla wins outright. The wins tally stands at 1 for the Tesla, 0 for the MX350.

The delta becomes more striking when you consider the MX350's average benchmark score across all its tests: 10,883, which is actually higher than its OpenCL result alone. That average still trails the Tesla's single OpenCL score of 11,479 by roughly 5.5%. The MX350's percentile ranking among all GPUs is 49, while the Tesla sits at 51 — a narrow overall standing gap, but the OpenCL test reveals where the real performance difference lies.

# Architecture Differences

The Tesla K20c is built on the Kepler architecture using the GK110 chip, fabricated on a 28 nm process at TSMC. That chip packs 7,080 million transistors into a 561 mm² die, yielding a transistor density of 12.6 million per square millimeter. The MX350, by contrast, uses the Pascal architecture with the GP107S chip, built on a 14 nm process at Samsung. It houses 3,300 million transistors on a 132 mm² die, achieving a density of 25.0 million per square millimeter — nearly double the Tesla's density despite having fewer than half the transistors.

Memory configurations diverge sharply. The Tesla carries 5 GB of GDDR5 on a 320-bit bus, delivering 208.0 GB/s of bandwidth. The MX350 has 2 GB of GDDR5 on a 64-bit bus, with just 56.06 GB/s — the Tesla offers 3.7 times the memory bandwidth. Clock speeds tell a different story: the MX350 runs at a base of 1354 MHz and boosts to 1468 MHz, while the Tesla's base and boost clocks are not listed. The Tesla's memory runs at 1300 MHz (5.2 Gbps effective), while the MX350's memory operates at 1752 MHz (7 Gbps effective).

Compute resources are lopsided. The Tesla fields 2,496 shading units, 208 texture mapping units, and 40 ROPs. The MX350 has 640 shading units, 32 TMUs, and 16 ROPs. Pixel rate favors the Tesla at 36.71 GPixel/s versus 23.49 GPixel/s, and texture rate is 146.8 GTexel/s against 46.98 GTexel/s. FP32 throughput is 3.524 TFLOPS for the Tesla and 1.879 TFLOPS for the MX350 — the Tesla is 87.5% faster in single-precision compute. The MX350 does list an FP16 rate of 29.36 GFLOPS (1:64), but the Tesla has no FP16 figure, suggesting it lacks dedicated half-precision throughput.

Power and physical specs could not be more different. The Tesla is a 225 W dual-slot card requiring 1x 6-pin plus 1x 8-pin power connectors and a 550 W suggested PSU. The MX350 is a 20 W part with no power connectors and no suggested PSU — a 11.25x reduction in TDP. The Tesla measures 267 mm (10.5 inches) in length; the MX350 has no listed dimensions because it is a mobile GPU. Interface and outputs also differ: the Tesla uses PCIe 2.0 x16 with no display outputs, while the MX350 uses PCIe 3.0 x4 and has portable-device-dependent outputs.

API support shows a small edge for the MX350 in DirectX and Vulkan. The Tesla supports DirectX 12 (11_0) and Vulkan 1.2.175, while the MX350 supports DirectX 12 (12_1) and Vulkan 1.4. Both offer OpenGL 4.6. Neither GPU has ray tracing cores or tensor cores. The Tesla launched on November 11, 2012, with a launch MSRP of 3,199 USD; the MX350 arrived on February 9, 2020, with no listed MSRP. Both are end-of-life products.

# FAQ

Q: Which GPU is faster in OpenCL?

A: The Tesla K20c scores 11,479 in Geekbench OpenCL, which is 32.1% higher than the MX350's 8,689. The Tesla wins the only head-to-head benchmark available.

Q: How does each GPU compare to its nearest rivals?

A: The Tesla sits 0.4% below the AMD Radeon Pro 5500M (11,528) and 1.7% below the NVIDIA GeForce GTX 1660 (11,680), while leading the GTX 780M by 1.9%. The MX350 is 0.7% above the AMD Radeon Pro 450 (10,804) and 1.1% above the NVIDIA Quadro K2200 (10,761), but 1.7% behind the AMD Radeon RX 550 (11,075).

Q: Do the GPUs support the same modern APIs?

A: Both support OpenGL 4.6 and DirectX 12, but the MX350 supports DirectX 12 (12_1) and Vulkan 1.4, while the Tesla is limited to DirectX 12 (11_0) and Vulkan 1.2.175. Neither has ray tracing or tensor cores.

Q: Which GPU has more memory bandwidth?

A: The Tesla K20c offers 208.0 GB/s via a 320-bit bus with 5 GB GDDR5. The MX350 provides 56.06 GB/s over a 64-bit bus with 2 GB GDDR5. The Tesla has roughly 3.7 times the bandwidth.

Q: What is the power consumption difference?

A: The Tesla is rated at 225 W and needs dual-slot cooling with 1x 6-pin and 1x 8-pin connectors. The MX350 draws 20 W, requires no power connectors, and has no suggested PSU rating.

Q: Which GPU is denser in terms of transistor packing?

A: The MX350's 14 nm Samsung process achieves 25.0 million transistors per mm² (3,300M on 132 mm²), versus 12.6M per mm² for the Tesla (7,080M on 561 mm² at 28 nm TSMC).

# The Verdict

Choose the Tesla K20c if you need raw compute throughput. Its 3.524 TFLOPS FP32 performance, 208.0 GB/s memory bandwidth, and 32.1% OpenCL lead over the MX350 make it the clear choice for compute-heavy workloads that can tolerate a 225 W power draw and dual-slot footprint. Its 2,496 shading units and 208 TMUs dwarf the MX350's 640 and 32, respectively. The Tesla also carries 5 GB of VRAM, which is 2.5 times the MX350's 2 GB, and its 51st percentile ranking among all GPUs edges out the MX350's 49th.

Choose the GeForce MX350 if power efficiency and portability matter more than peak performance. At 20 W, it consumes only 8.9% of the Tesla's power budget, and it fits in thin laptops with no external power connectors. Its 14 nm process delivers higher transistor density (25.0M per mm² vs 12.6M), and it supports newer API features: DirectX 12 (12_1) and Vulkan 1.4, versus the Tesla's DirectX 12 (11_0) and Vulkan 1.2.175. The MX350 also has a boost clock of 1468 MHz, whereas the Tesla lists no boost clock at all.

The data does not support a single winner — these are different tools for different jobs. The Tesla's 11,479 OpenCL score is 32.1% higher, but that advantage comes at a 11.25x power cost. The MX350's 10,883 average benchmark score is respectable for a 20 W mobile part, and its 13,077 Vulkan score suggests better driver optimization for modern APIs, even though the Tesla has no Vulkan result to compare. If you are rendering or computing in a desktop workstation, the Tesla wins outright. If you are gaming or creating content on battery power, the MX350 is the only sensible pick.

# Specification Differences

| Specification | NVIDIA Tesla K20c | NVIDIA GeForce MX350 |

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

| Chip | GK110 | GP107S |

| Architecture | Kepler | Pascal |

| Process Node | 28 nm | 14 nm |

| Foundry | TSMC | Samsung |

| Transistors | 7,080 million | 3,300 million |

| Die Size | 561 mm² | 132 mm² |

| Transistor Density | 12.6M / mm² | 25.0M / mm² |

| Base Clock | N/A | 1354 MHz |

| Boost Clock | N/A | 1468 MHz |

| Memory Clock | 1300 MHz (5.2 Gbps effective) | 1752 MHz (7 Gbps effective) |

| Memory Size | 5 GB | 2 GB |

| Memory Bus Width | 320 bit | 64 bit |

| Memory Bandwidth | 208.0 GB/s | 56.06 GB/s |

| Shading Units | 2496 | 640 |

| TMUs | 208 | 32 |

| ROPs | 40 | 16 |

| Pixel Rate | 36.71 GPixel/s | 23.49 GPixel/s |

| Texture Rate | 146.8 GTexel/s | 46.98 GTexel/s |

| FP32 | 3.524 TFLOPS | 1.879 TFLOPS |

| FP16 | N/A | 29.36 GFLOPS (1:64) |

| TDP | 225 W | 20 W |

| Slot Width | Dual-slot | N/A |

| Power Connectors | 1x 6-pin + 1x 8-pin | None |

| Suggested PSU | 550 W | N/A |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x4 |

| Display Outputs | No outputs | Portable Device Dependent |

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

| Vulkan | 1.2.175 | 1.4 |

| Length | 267 mm (10.5 inches) | N/A |

| Release Date | 2012-11-11 | 2020-02-09 |

| Launch MSRP | 3,199 USD | N/A |