NVIDIA GeForce GTX 950 OEM

The GeForce GTX 950 OEM is a Maxwell 2.0 part from NVIDIA, produced on TSMC's 28 nm process, with 2,940 million transistors on a 228 mm² die. The data record contains no benchmark scores and no nearest-rival comparisons, so its positional signal comes from the 50th percentile ranking against all GPUs in the database, placing it in the middle of the overall distribution.

Benchmark Performance

The benchmark record for this SKU is empty: the average benchmark score is 0, and the nearestRivals list has no entries. As a result, exact percentage deltas versus competing GPUs cannot be calculated from the data provided. What the record does include are the card's raw throughput specifications, which serve as the only available indicators of computational capacity. Peak FP32 performance is 2.464 TFLOPS, texture rate is 76.99 GTexel/s, and pixel rate is 38.50 GPixel/s. These figures describe the capability of the shader array, texture units, and ROPs, but they are not frame-rate measurements.

The percentileVsAllGpus value of 50 suggests a midpoint placement relative to the full GPU distribution. Without actual scores, however, that percentile cannot be translated into a specific margin over any named card. The absence of score data also means no settings-specific or resolution-specific performance claims can be substantiated. The 50th percentile should therefore be read as a broad positional marker, not as a quantified performance advantage. The raw throughput numbers indicate a modest shader and fill-rate design, but competitive ranking remains undefined.

Ray Tracing and Feature Set

The card has no RT cores and no tensor cores in the fact record. It is based on Maxwell 2.0 architecture, so the feature set is defined by conventional fixed-function and shader hardware: 1024 shading units, 64 texture mapping units, and 32 ROPs. API support includes DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The API list does not include any ray-tracing extension in the data. Because no dedicated RT or tensor hardware is present, workloads that depend on those units cannot be offloaded to this GPU.

The display output field is listed as "No outputs", indicating an OEM-oriented board without integrated display connectivity. The bus interface is PCIe 3.0 x16. The feature set is therefore centered on standard rasterization workloads, with no support for ray tracing acceleration hardware and no tensor-core capabilities. The absence of display outputs also separates this card from consumer cards that can drive monitors directly.

How It Compares

The nearestRivals data is empty, so there are no rival names, scores, or deltaPct values to report. In place of a per-rival comparison, the only positional frame from the data is the 50th percentile across all GPUs. The predecessor is GeForce 700 and the successor is GeForce 10, which places this part in the GeForce 900 generation. Without rival scores, no exact competitive advantage can be quantified.

The card's 2.464 TFLOPS FP32 throughput and 76.99 GTexel/s texture rate are the primary specification anchors, but any direct comparison to a named competitor would require matching benchmark data. The record leaves the comparative story incomplete. What can be stated is that this OEM part occupies the middle of the database percentile distribution and belongs to the GeForce 900 generation between the GeForce 700 and GeForce 10 lines.

FAQ

Q: Does the GeForce GTX 950 OEM support ray tracing?

A: No. The fact record lists no RT cores, so ray tracing acceleration hardware is not present.

Q: How much memory does it have and what is the bus width?

A: It has 4 GB of GDDR5 memory on a 128-bit bus, with 80.19 GB/s of bandwidth. The memory clock is 1253 MHz, or 5 Gbps effective.

Q: What APIs are supported?

A: The supported APIs are DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Q: What power supply is recommended?

A: The suggested PSU rating is 200 W, and the card uses no power connectors.

Q: Is this GPU still in production?

A: No. Its production status is end-of-life.

Q: Does it have tensor cores?

A: No. The fact record lists no tensor cores.

Q: What display outputs does it have?

A: None. The display outputs field is "No outputs."

Who Should Consider It

Because the average benchmark score is 0 and no frame-rate data is present, recommendations must be based on the supplied specifications. With 4 GB of GDDR5 and 80.19 GB/s of bandwidth, the card can address moderate memory workloads. The 1024 shading units and 2.464 TFLOPS FP32 throughput indicate a shader array suited to modest rendering tasks. The 50th percentile placement suggests it sits in the middle of the GPU field, but no settings-specific results are listed to define a precise performance tier.

The card's API support for DirectX 12 (12_1) and Vulkan 1.4 makes it usable in modern API environments, but the lack of display outputs means it cannot directly drive a monitor from its own outputs. This part is therefore more appropriate in systems where another path handles display output, or where no display output is required. The 4 GB memory capacity sets a hard limit on allocation size, though the data does not define a resolution ceiling. Users with modest texture and shading demands are the relevant audience; those requiring ray tracing or tensor acceleration will not find the needed hardware here.

Memory Subsystem

The memory subsystem consists of 4 GB of GDDR5 on a 128-bit bus, producing 80.19 GB/s of bandwidth at 1253 MHz, with 5 Gbps effective data rate. The 128-bit bus width constrains peak transfer compared to wider interfaces, making bandwidth a potential limiting resource for large frame buffers. The 4 GB capacity provides a substantial allocation space, but the 80.19 GB/s rate dictates how quickly data can be moved into and out of the GPU.

In workloads where textures or frame buffers exceed the comfortable throughput of the 128-bit bus, performance will be gated by memory transfer rather than shader count. The 32 ROPs deliver 38.50 GPixel/s, which is relevant for fill-rate-oriented tasks. The memory subsystem should be viewed as moderate: sufficient capacity for many scenes, with a bandwidth figure that is modest in absolute terms. The combination of 4 GB capacity and 80.19 GB/s bandwidth defines the memory ceiling of this OEM part.

Power and Cooling

The data sheet does not list a TDP. The only power figures available are the suggested PSU rating of 200 W and the fact that the card requires no power connectors. The slot width is dual-slot. Because no auxiliary power connector is listed, the data specifies only the 200 W PSU recommendation and no external connector requirement.

The card's production facts also inform the power and cooling picture: the GPU is built on a 28 nm process at TSMC, with a die size of 228 mm² and 2,940 million transistors. With no TDP value, thermal output cannot be quantified, but dual-slot cooling indicates a cooler occupying two expansion slots. This part is end-of-life, and the power delivery data is limited to the 200 W system recommendation. Without a TDP figure, further thermal analysis cannot be derived from the record.