NVIDIA GeForce GTX 950A vs NVIDIA TITAN V CEO Edition Comparison

Where Each One Wins

The recorded data splits these two NVIDIA parts into entirely different performance classes, so the use-case separation is stark. The GeForce GTX 950A is a modest, low-power mobile-class GPU built for portability and light duties. Its single benchmark entry, a Geekbench OpenCL score of 10,273, places it at the 48th percentile of all GPUs in the database, which indicates a mid-pack result for everyday compute workloads. Its nearest rivals in the database include the AMD Radeon RX 6500M at 10,362 (0.9% higher) and the AMD Radeon R9 M375 at 10,070 (2% lower), so the GTX 950A sits in a tight cluster of similar-performance parts. This is not a card for heavy rendering or high-end gaming; it is a capable entry-level solution for systems where space and power draw matter more than raw throughput.

The NVIDIA TITAN V CEO Edition is the polar opposite. It is a flagship-class compute monster with a 5120 shading units, 320 texture mapping units, and 128 ROPs, backed by 32 GB of HBM2 memory on a 4096-bit bus. Its Passmark G3D score of 16,988 is massive relative to the GTX 950A's profile, and its Passmark G2D score of 1,086 reflects strong 2D desktop performance. The TITAN V CEO Edition also carries 640 tensor cores, making it explicitly oriented toward deep learning and scientific compute, not just rasterization. The database shows its average benchmark score at 9,037, which is lower than its G3D score because the average includes only the two Passmark tests; the G2D result drags the average down. Still, the recorded data indicates that this GPU wins decisively in every compute-heavy category where it has a benchmark entry.

The GeForce GTX 950A wins in the narrow category of power efficiency and physical footprint. It draws only 75 W, uses an MXM Module slot width, requires no power connectors, and is described as "Portable Device Dependent" for display outputs. The TITAN V CEO Edition, by contrast, draws 250 W, needs a 600 W suggested PSU, uses a dual-slot cooler with one 6-pin and one 8-pin power connector, and measures 267 mm in length, 112 mm in height, and 40 mm in width. For a portable workstation or a compact embedded system, the GTX 950A is the only viable choice from the data. For anything involving large datasets, high-resolution compute, or AI workloads, the TITAN V CEO Edition is the clear winner.

Architecture Differences

The two GPUs come from different manufacturing generations and entirely different architectural families. The GTX 950A is built on the Maxwell architecture using the GM107 chip, fabricated on a 28 nm process at TSMC. It packs 1,870 million transistors into a 148 mm² die, yielding a transistor density of 12.6 million transistors per square millimeter. The TITAN V CEO Edition uses the Volta architecture with the GV100 chip, fabricated on a 12 nm process, also at TSMC. It contains 21,100 million transistors on an 815 mm² die, for a density of 25.9 million transistors per square millimeter. That is more than double the transistor density of the GTX 950A, and over eleven times the raw transistor count.

Memory subsystems are fundamentally different as well. The GTX 950A uses 2 GB of DDR3 memory on a 128-bit bus, with a bandwidth of 32.03 GB/s and a memory clock of 1001 MHz (2 Gbps effective). The TITAN V CEO Edition uses 32 GB of HBM2 memory on a 4096-bit bus, with a bandwidth of 868.4 GB/s and a memory clock of 848 MHz (1696 Mbps effective). The bandwidth gap is enormous: 868.4 GB/s versus 32.03 GB/s represents a 27x difference in raw memory throughput. That single factor drives most of the performance separation in memory-bound workloads.

Compute resources differ by an order of magnitude. The GTX 950A has 640 shading units, 40 TMUs, and 16 ROPs, with a pixel rate of 17.98 GPixel/s and a texture rate of 44.96 GTexel/s. Its FP32 throughput is 1,438.7 GFLOPS, and it has no tensor cores and no FP16 entry in the database. The TITAN V CEO Edition has 5120 shading units, 320 TMUs, 128 ROPs, and 640 tensor cores. Its pixel rate is 186.2 GPixel/s, its texture rate is 465.6 GTexel/s, and its FP32 throughput is 14.90 TFLOPS. It also has a recorded FP16 rate of 29.80 TFLOPS at a 2:1 ratio, which the GTX 950A lacks entirely.

Clock speeds and API support also differ. The GTX 950A runs at a base clock of 993 MHz and a boost of 1124 MHz, while the TITAN V CEO Edition runs at 1200 MHz base and 1455 MHz boost. Both support OpenGL 4.6 and Vulkan 1.4, but the TITAN V CEO Edition supports DirectX 12 (12_1) while the GTX 950A only supports DirectX 12 (11_0). The TITAN V CEO Edition also has a PCIe 3.0 x16 bus interface, while the GTX 950A uses an MXM-B (3.0) interface.

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark entries comparing these two GPUs, and the wins counter shows zero wins for each. However, the individual benchmark records still allow a meaningful comparison. The GTX 950A has one benchmark: Geekbench OpenCL at 10,273. The TITAN V CEO Edition has two: Passmark G2D at 1,086 and Passmark G3D at 16,988. These are different tests, so direct numerical comparison is not possible. What is possible is contextualizing each score against the database's nearest rivals.

The GTX 950A's Geekbench OpenCL score of 10,273 sits within 2% of several peers. The AMD Radeon RX 6500M scores 10,362, which is 0.9% higher. The NVIDIA Tesla C2075 scores 10,400, which is 1.2% higher. The AMD Radeon R9 M375 scores 10,070, which is 2% lower. The AMD Radeon RX 550X scores 10,481, which is 2% higher. This tells us the GTX 950A is firmly in the lower-midrange of the database's recorded GPUs, with its 48th percentile ranking confirming a median position.

The TITAN V CEO Edition's Passmark G3D score of 16,988 is its headline result. Its average benchmark score of 9,037 is pulled down by the G2D score of 1,086, but the G3D score alone places it far above the GTX 950A's class. The nearest rivals for the TITAN V CEO Edition are all much older or much weaker cards: the NVIDIA GeForce GTX 660 at 9,022 (0.2% higher average), the NVIDIA GeForce GTX 560 at 9,058 (0.2% lower), the AMD Radeon 550X at 8,918 (1.3% lower), and the AMD Radeon 890M at 9,210 (1.9% higher). These rivals are clustered tightly around the TITAN V CEO Edition's average score, which reflects the fact that its average is dominated by the low G2D score. But for raw 3D performance, the G3D score of 16,988 is what matters, and it dwarfs anything in the GTX 950A's vicinity.

The biggest win for the TITAN V CEO Edition is in memory bandwidth, pixel rate, and texture rate. The 868.4 GB/s bandwidth is 27x the GTX 950A's 32.03 GB/s. The 186.2 GPixel/s pixel rate is over 10x the GTX 950A's 17.98 GPixel/s. The 465.6 GTexel/s texture rate is over 10x the GTX 950A's 44.96 GTexel/s. The FP32 throughput of 14.90 TFLOPS is also over 10x the GTX 950A's 1,438.7 GFLOPS. These are not incremental improvements; they are class-defining gaps.

The GTX 950A's only clear advantage in the recorded data is power consumption: 75 W versus 250 W. That is a 3.3x difference in favor of the GTX 950A, and it comes with no power connectors required, versus the TITAN V CEO Edition's need for one 6-pin and one 8-pin connector.

The Verdict

The data points to a simple conclusion: the NVIDIA TITAN V CEO Edition is the superior GPU in every compute metric recorded in the database, and the GeForce GTX 950A is only preferable where power draw, physical size, and system integration constraints are paramount. The TITAN V CEO Edition offers 5120 shading units versus 640, 32 GB of HBM2 versus 2 GB of DDR3, 868.4 GB/s bandwidth versus 32.03 GB/s, and 640 tensor cores versus none. Its FP32 throughput of 14.90 TFLOPS is more than ten times the GTX 950A's 1,438.7 GFLOPS. For any workload that stresses the GPU, the TITAN V CEO Edition wins by an overwhelming margin.

The GeForce GTX 950A is a 75 W MXM module with no power connectors, designed for portable devices. It has a 48th percentile ranking and its nearest rivals are all in the 10,000 to 10,500 Geekbench OpenCL range. It is not a competitive performer for modern heavy workloads, but it is a low-power, low-profile option that fits into systems where a dual-slot 267 mm card with a 600 W PSU requirement is simply not an option.

Pick the TITAN V CEO Edition if your application demands high compute throughput, large memory capacity, or tensor core acceleration. The recorded data shows no scenario in which the GTX 950A outperforms it in raw performance. Pick the GeForce GTX 950A only if your system cannot physically or electrically accommodate the TITAN V CEO Edition. The GTX 950A is end-of-life, as is the TITAN V CEO Edition, but the GTX 950A's performance class is firmly in the entry-level segment.

FAQ

Q: Which GPU has more shading units?

A: The NVIDIA TITAN V CEO Edition has 5120 shading units, while the GeForce GTX 950A has 640.

Q: How much memory bandwidth does each card provide?

A: The TITAN V CEO Edition provides 868.4 GB/s over a 4096-bit HBM2 interface, while the GTX 950A provides 32.03 GB/s over a 128-bit DDR3 interface.

Q: Does the GeForce GTX 950A support tensor cores?

A: No. The GTX 950A has no tensor cores recorded in the database. The TITAN V CEO Edition has 640 tensor cores.

Q: What is the TITAN V CEO Edition's Passmark G3D score?

A: The TITAN V CEO Edition scores 16,988 in Passmark G3D and 1,086 in Passmark G2D.

Q: What is the GeForce GTX 950A's Geekbench OpenCL score?

A: The GTX 950A scores 10,273 in Geekbench OpenCL, placing it at the 48th percentile of all GPUs in the database.

Q: Which card has a higher pixel rate?

A: The TITAN V CEO Edition has a pixel rate of 186.2 GPixel/s, compared to the GTX 950A's 17.98 GPixel/s.

Specification Differences

| Specification | NVIDIA GeForce GTX 950A | NVIDIA TITAN V CEO Edition |

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

| Architecture | Maxwell | Volta |

| Chip | GM107 | GV100 |

| Process Node | 28 nm | 12 nm |

| Transistors | 1,870 million | 21,100 million |

| Die Size | 148 mm² | 815 mm² |

| Transistor Density | 12.6M / mm² | 25.9M / mm² |

| Base Clock | 993 MHz | 1200 MHz |

| Boost Clock | 1124 MHz | 1455 MHz |

| Memory Clock | 1001 MHz, 2 Gbps effective | 848 MHz, 1696 Mbps effective |

| Memory Size | 2 GB | 32 GB |

| Memory Type | DDR3 | HBM2 |

| Memory Bus Width | 128 bit | 4096 bit |

| Memory Bandwidth | 32.03 GB/s | 868.4 GB/s |

| Shading Units | 640 | 5120 |

| TMUs | 40 | 320 |

| ROPs | 16 | 128 |

| Tensor Cores | None | 640 |

| Pixel Rate | 17.98 GPixel/s | 186.2 GPixel/s |

| Texture Rate | 44.96 GTexel/s | 465.6 GTexel/s |

| FP32 | 1,438.7 GFLOPS | 14.90 TFLOPS |

| FP16 | Not recorded | 29.80 TFLOPS (2:1) |

| TDP | 75 W | 250 W |

| Slot Width | MXM Module | Dual-slot |

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

| Suggested PSU | Not recorded | 600 W |

| Bus Interface | MXM-B (3.0) | PCIe 3.0 x16 |

| Display Outputs | Portable Device Dependent | 1x HDMI 2.0, 3x DisplayPort 1.4a |

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

| Dimensions | Not recorded | 267 mm x 112 mm x 40 mm |

| Release Date | 2015-03-12 | 2018-06-20 |

| Production Status | End-of-life | End-of-life |