NVIDIA GeForce MX350 vs NVIDIA Tesla C2070 Comparison

Head-to-Head Benchmarks

The only direct benchmark comparison available in the database is the Geekbench OpenCL test. In this test, the NVIDIA Tesla C2070 records a score of 9716, while the NVIDIA GeForce MX350 scores 8689. This gives the Tesla C2070 a 10.6% lead over the MX350, the sole head-to-head win recorded between these two parts.

Looking at the broader context, the MX350's average benchmark score of 10883 places it at the 49th percentile among all GPUs. Its nearest rivals show how tightly clustered this performance tier is: the AMD Radeon Pro 450 averages 10804, just 0.7% behind, and the NVIDIA Quadro K2200 averages 10761, a 1.1% gap. On the other side, the NVIDIA GeForce GTX 1650 SUPER scores 11047, putting the MX350 1.5% behind it, while the AMD Radeon RX 550 scores 11075, a 1.7% difference. These deltas are small enough that driver optimization or workload type could easily shift the ranking between these cards.

The Tesla C2070, by contrast, sits at the 47th percentile with an average score of 9716. Its nearest rivals follow a similar pattern: the NVIDIA Tesla M10 averages 9724 (0.1% ahead of the C2070), the NVIDIA Quadro P4000 averages 9665 (0.5% behind), and the AMD Radeon Pro WX 2100 averages 9653 (0.7% behind). The NVIDIA GeForce GTX 1070, a much newer consumer part, averages 9780, which is only 0.7% ahead of this 2011 compute card. The C2070's single benchmark score does not capture its full capabilities, but the data suggests it still competes in a similar performance band to much newer hardware.

The MX350's other recorded benchmark, Geekbench Vulkan, shows a score of 13077, a figure that is significantly higher than its OpenCL score. This indicates the MX350 has solid driver support for Vulkan workloads, which matters for modern gaming and compute applications. The Tesla C2070 has no recorded Vulkan benchmark in the database, and its API support list does not include Vulkan at all, so direct comparison on that API is not possible.

Where Each One Wins

The Tesla C2070 wins the only direct benchmark comparison, taking the Geekbench OpenCL test by 10.6%. This result aligns with its compute-oriented design: the C2070 was built for general-purpose GPU computing, and its 448 shading units, 56 texture mapping units, and 48 render output units give it a wider memory bus and higher memory bandwidth (143.4 GB/s) than the MX350. In OpenCL workloads that scale with memory throughput, the C2070's advantage is clear.

The MX350, however, wins in other dimensions that the average benchmark score reflects. Its average score of 10883 is 12.0% higher than the Tesla C2070's 9716 average, meaning that across a broader set of tasks, the MX350's newer architecture and higher clock speeds (1354 MHz base, 1468 MHz boost) deliver better overall throughput. The MX350 also has a recorded Vulkan score of 13077, which the C2070 cannot match because it lacks Vulkan support entirely. For any workload using Vulkan, the MX350 is the only option between the two.

In terms of efficiency, the MX350's 20 W TDP versus the C2070's 238 W TDP is a decisive practical difference. The MX350 achieves its performance at roughly one-twelfth the power draw, making it suitable for thin-and-light laptops and systems without dedicated power connectors. The C2070 requires a 1x 6-pin plus 1x 8-pin power connector setup and a 550 W suggested power supply, which limits it to desktop workstations with substantial power delivery.

Architecture Differences

The two GPUs come from entirely different architectural generations. The MX350 uses the GP107S chip built on the Pascal architecture, manufactured on a 14 nm process by Samsung. The Tesla C2070 uses the GF100 chip on the Fermi architecture, manufactured on a 40 nm process by TSMC. This process node difference is substantial: 14 nm versus 40 nm, which explains the massive transistor density gap. The MX350 packs 3,300 million transistors into a 132 mm² die, yielding a density of 25.0M transistors per mm². The C2070 has 3,100 million transistors spread across a 529 mm² die, giving a density of only 5.9M per mm².

The architecture generations also differ in API support. The MX350 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The C2070 supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support listed. The DirectX feature level difference (12_1 versus 11_0) means the MX350 supports newer rendering features that the C2070 cannot handle. The MX350 also has a dedicated FP16 throughput figure of 29.36 GFLOPS (1:64), while the C2070 has no FP16 data recorded, reflecting the Fermi architecture's lack of dedicated half-precision support.

Memory technology differs as well. The MX350 uses 2 GB of GDDR5 memory on a 64-bit bus, while the C2070 uses 6 GB of GDDR5 on a 384-bit bus. The C2070's memory clock is 747 MHz (3 Gbps effective), while the MX350's is 1752 MHz (7 Gbps effective). The C2070's much wider bus gives it more than double the bandwidth: 143.4 GB/s versus 56.06 GB/s for the MX350. This memory bandwidth advantage is why the C2070 wins the OpenCL test despite its older architecture and lower clock speeds.

Specification Differences

The two cards differ in nearly every measurable specification. The MX350 has 640 shading units, 32 TMUs, and 16 ROPs. The C2070 has 448 shading units, 56 TMUs, and 48 ROPs. Despite having more shading units, the MX350's pixel rate of 23.49 GPixel/s is higher than the C2070's 16.07 GPixel/s, and its texture rate of 46.98 GTexel/s exceeds the C2070's 32.14 GTexel/s. The MX350's FP32 throughput is 1.879 TFLOPS, while the C2070's is 1,027.7 GFLOPS, which converts to roughly 1.03 TFLOPS. So the MX350 has about 83% higher FP32 compute throughput.

Clock speeds also favor the MX350. It runs at 1354 MHz base and 1468 MHz boost, while the C2070 has no base or boost clock listed in the database, only its memory clock of 747 MHz. The MX350's memory clock of 1752 MHz (7 Gbps effective) is more than double the C2070's memory clock, though the C2070's wider bus compensates with higher overall bandwidth.

Physical and power specifications diverge sharply. The MX350 has no power connectors, a 20 W TDP, and a PCIe 3.0 x4 bus interface. The C2070 is dual-slot, 248 mm long (9.8 inches), requires 1x 6-pin plus 1x 8-pin power connectors, has a 238 W TDP, a 550 W suggested power supply, and uses PCIe 2.0 x16. The MX350's display outputs are portable device dependent, while the C2070 has a single DVI output. The MX350 was released on February 9, 2020, while the C2070 came out on July 24, 2011, a gap of nearly nine years.

FAQ

Q: Which GPU has higher raw compute performance in FP32?

A: The NVIDIA GeForce MX350 has 1.879 TFLOPS FP32 throughput, while the NVIDIA Tesla C2070 has 1,027.7 GFLOPS. The MX350 is roughly 83% faster in this metric.

Q: Which GPU has more memory bandwidth?

A: The NVIDIA Tesla C2070 has 143.4 GB/s bandwidth from its 384-bit bus, while the NVIDIA GeForce MX350 has 56.06 GB/s from a 64-bit bus. The C2070 has more than double the bandwidth.

Q: Does the Tesla C2070 support Vulkan?

A: No. The database lists DirectX 12 (11_0) and OpenGL 4.6 for the Tesla C2070, with no Vulkan support. The MX350 supports Vulkan 1.4 and has a recorded Geekbench Vulkan score of 13077.

Q: Which GPU has the higher average benchmark score?

A: The MX350 averages 10883 across its recorded benchmarks, placing at the 49th percentile. The Tesla C2070 averages 9716, placing at the 47th percentile. The MX350's average is about 12% higher.

Q: What are the power requirements for each card?

A: The MX350 has a 20 W TDP and needs no power connectors. The Tesla C2070 has a 238 W TDP, requires a 1x 6-pin plus 1x 8-pin power connector setup, and has a 550 W suggested power supply.

Q: Which GPU has more memory capacity?

A: The Tesla C2070 has 6 GB of GDDR5 memory, while the MX350 has 2 GB. Both use GDDR5, but the C2070's capacity is three times larger.

The Verdict

The data paints a clear split between these two cards. The NVIDIA GeForce MX350 is the better all-round performer for modern workloads. Its average benchmark score of 10883 beats the Tesla C2070's 9716 by 12%, it supports Vulkan 1.4 and DirectX 12 (12_1), and it does so at 20 W TDP with no external power connectors. For anyone building a low-power system, a laptop, or a small form factor machine, the MX350 is the obvious choice. Its nearest rivals sit within 1.7% on either side, so it competes directly with cards like the GeForce GTX 1650 SUPER and Radeon RX 550, and the MX350 is effectively their peer.

The NVIDIA Tesla C2070 wins the specific OpenCL test by 10.6%, and its 143.4 GB/s memory bandwidth and 6 GB capacity make it better suited for memory-bound compute tasks. Its 384-bit bus is the key advantage, and the C2070's nearest rivals (Tesla M10, Quadro P4000, Radeon Pro WX 2100) all score within 0.7% of it, meaning it remains competitive in its niche. However, the C2070's lack of Vulkan support, older DirectX feature level, and 238 W power draw make it impractical for modern gaming or efficiency-focused systems. The C2070 also requires a 550 W power supply and a dual-slot chassis, which limits where it can be installed.

The verdict depends on the use case. For general computing, gaming, or any Vulkan-based workload, the MX350 is the superior choice, and its higher FP32 throughput (1.879 TFLOPS) confirms it is the faster chip for most tasks. For raw memory bandwidth or legacy OpenCL compute that can use 6 GB of VRAM, the C2070 retains a genuine advantage, but it comes with significant power and space costs. The MX350 is the practical pick for nearly all modern scenarios, while the C2070 is a specialist tool for specific compute work where its memory subsystem matters more than its older architecture.