Intel Arc A380 vs NVIDIA GeForce GTX 950A Comparison

Head-to-Head Benchmarks

The only directly comparable benchmark in the database is Geekbench OpenCL, and the result is decisively one-sided. The Intel Arc A380 scores 38,224, while the NVIDIA GeForce GTX 950A scores 10,273. That is a 73.1% margin in favor of the Intel part, meaning the Arc A380 delivers nearly four times the compute throughput in this particular workload. The delta is stark enough that no amount of driver tuning or overclocking on the older Maxwell card could close the gap.

Looking at the rivals surrounding each card in the database, the GTX 950A sits within a tight cluster. Its closest competitor, the AMD Radeon RX 6500M, averages 10,362, which is 0.9% higher than the GTX 950A. The NVIDIA Tesla C2075 is 1.2% ahead at 10,400. On the flip side, the GTX 950A beats the AMD Radeon R9 M375 by 2% (10,070) and trails the AMD Radeon RX 550X by 2% (10,481). These are all sub-3% margins, indicating the GTX 950A is firmly entrenched in a mid-pack performance tier from roughly seven years ago.

The Arc A380, by contrast, sits in a different competitive neighborhood entirely. Its nearest rival, the AMD FirePro W5170M, averages 8,595, which is 0.4% behind the Intel card. The AMD Radeon HD 8870M is 1.1% behind at 8,462, and the NVIDIA GeForce MX330 is 1.2% behind at 8,458. The AMD Radeon 880M is 1.4% behind at 8,436. Interestingly, the Arc A380's average benchmark score across all recorded tests is 8,558, which is lower than its Geekbench OpenCL score. This suggests that while the OpenCL result is strong, other workloads (like the PassMark DirectX tests) pull the overall average down.

The database records zero wins for the GTX 950A in head-to-head comparisons and one win for the Arc A380. That single benchmark tells the story: in raw compute, the Intel card is in a different league. But the broader benchmark suite for the Arc A380 reveals a more nuanced picture. In PassMark DirectX 9, it scores 73, but in DirectX 10, 11, and 12, scores drop to 37, 38, and 35 respectively. The DirectX 9 result is roughly double the newer API scores, which is a curious pattern for a card that supports DirectX 12 Ultimate. The PassMark G3D score of 6,252 and GPU compute score of 2,762 further indicate that the card's strength is not uniformly distributed across all rendering paths.

Architecture Differences

The two GPUs come from different eras and fundamentally different design philosophies. The GTX 950A uses the GM107 chip on NVIDIA's Maxwell architecture, built on a 28 nm process at TSMC. The die is 148 mm² and packs 1,870 million transistors, yielding a transistor density of 12.6 million per mm². The Arc A380 uses the DG2-128 chip on Intel's Xe-HPG architecture, built on a 6 nm process, also at TSMC. The die is slightly larger at 157 mm² but contains 7,200 million transistors, resulting in a density of 45.9 million per mm². That is a 3.6x improvement in transistor density, reflecting the massive process node advantage.

Memory configurations diverge sharply. The GTX 950A has 2 GB of DDR3 on a 128-bit bus, producing 32.03 GB/s of bandwidth. Memory clock is 1,001 MHz with 2 Gbps effective data rate. The Arc A380 has 6 GB of GDDR6 on a 96-bit bus, delivering 186.0 GB/s of bandwidth. Memory clock is 1,937 MHz with 15.5 Gbps effective. Despite the narrower bus, the Arc A380 has nearly 6x the memory bandwidth due to the far faster GDDR6 signaling. The GTX 950A's 2 GB frame buffer is also a limiting factor for modern textures and resolutions, whereas the Arc A380's 6 GB is more comfortable for current game assets.

Core counts follow the architectural generational leap. The GTX 950A has 640 shading units, 40 TMUs, and 16 ROPs. The Arc A380 has 1,024 shading units, 64 TMUs, and 32 ROPs. The Intel card also includes 8 ray tracing cores, a feature entirely absent from the Maxwell design. Pixel rate for the Arc A380 is 65.60 GPixel/s versus 17.98 GPixel/s for the GTX 950A. Texture rate is 131.2 GTexel/s versus 44.96 GTexel/s. FP32 compute is 4.198 TFLOPS versus 1,438.7 GFLOPS. The Arc A380 also has FP16 throughput of 8.397 TFLOPS (2:1 ratio), while the GTX 950A has no recorded FP16 capability.

Clock speeds differ due to the process advantage. The GTX 950A runs at 993 MHz base and 1,124 MHz boost. The Arc A380 runs at 2,000 MHz base and 2,050 MHz boost. Both cards have the same 75 W TDP, but the Arc A380 requires a 1x 8-pin power connector while the GTX 950A has none. The GTX 950A uses an MXM-B (3.0) interface and is a MXM module, making it a portable-device-oriented part. The Arc A380 is a dual-slot card using PCIe 4.0 x8, with dimensions of 222 mm length, 114 mm height, and 42 mm width. Display outputs also differ: the GTX 950A has portable-device-dependent outputs, while the Arc A380 offers 1x HDMI 2.1 and 3x DisplayPort 2.0.

API support reflects the generational gap. The GTX 950A supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The Arc A380 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The DirectX feature level difference means the Intel card can handle mesh shaders, variable rate shading, and other modern rendering techniques that the Maxwell part cannot.

The Verdict

The data is unambiguous for compute-heavy workloads. The Intel Arc A380 beats the NVIDIA GeForce GTX 950A by 73.1% in Geekbench OpenCL. The Arc A380 also has 6x the memory bandwidth, 3.6x the transistor density, 8 ray tracing cores, and support for DirectX 12 Ultimate. The GTX 950A is end-of-life, released March 12, 2015, while the Arc A380 is also end-of-life but released June 13, 2022, and had a launch MSRP of 149 USD.

For users running modern games with ray tracing or DirectX 12 Ultimate features, the Arc A380 is the only viable choice from these two. The GTX 950A lacks the hardware capabilities entirely. For legacy software that relies on older DirectX paths, the GTX 950A's PassMark DirectX 9 score of 73 is not directly comparable to the Arc A380's 73, but the database shows the Arc A380 scores 73 in DirectX 9 as well. However, the Arc A380's DirectX 10, 11, and 12 scores (37, 38, 35) are notably lower than its DirectX 9 score, suggesting some driver overhead or architectural inefficiency in those paths.

The percentile rankings place the GTX 950A at the 48th percentile of all GPUs, while the Arc A380 sits at the 44th percentile. This is counterintuitive given the OpenCL results, but it reflects the fact that the Arc A380's average benchmark score (8,558) is dragged down by its low PassMark scores. The GTX 950A's average score is 10,273, with only a single benchmark recorded. The Arc A380's broader benchmark suite shows inconsistent performance, with its best results in compute (Geekbench OpenCL 38,224, Geekbench Vulkan 36,736) and its worst in legacy DirectX paths.

For anyone needing raw compute in applications that use OpenCL or Vulkan, the Arc A380 is the clear winner. For anyone stuck with a MXM-based laptop and limited to the GTX 950A's form factor, the comparison is moot because the Arc A380 is a dual-slot PCIe card that cannot fit. The GTX 950A's advantage is its 75 W TDP with no external power connector, making it a drop-in upgrade for portable devices with MXM slots. The Arc A380 requires a 250 W suggested PSU and an 8-pin connector.

FAQ

Q: Which GPU has higher memory bandwidth?

A: The Intel Arc A380 has 186.0 GB/s bandwidth, while the NVIDIA GeForce GTX 950A has 32.03 GB/s. The Arc A380's GDDR6 memory on a 96-bit bus far outpaces the GTX 950A's DDR3 on a 128-bit bus.

Q: Does the GTX 950A support ray tracing?

A: No. The GTX 950A has no ray tracing cores. The Intel Arc A380 includes 8 ray tracing cores.

Q: What is the difference in FP32 compute performance?

A: The Arc A380 delivers 4.198 TFLOPS of FP32 compute, while the GTX 950A delivers 1,438.7 GFLOPS. The Intel card is roughly 2.9x faster in raw FP32 throughput.

Q: Are both cards still in production?

A: No. Both are marked as end-of-life. The GTX 950A was released on March 12, 2015, and the Arc A380 on June 13, 2022.

Q: Which card has more shading units?

A: The Arc A380 has 1,024 shading units, compared to the GTX 950A's 640 shading units. The Intel card also has 64 TMUs and 32 ROPs versus 40 TMUs and 16 ROPs.

Q: What is the form factor difference?

A: The GTX 950A is an MXM module using the MXM-B (3.0) interface, while the Arc A380 is a dual-slot PCIe 4.0 x8 card measuring 222 mm by 114 mm by 42 mm.

Where Each One Wins

The Intel Arc A380 wins in every modern compute and rendering scenario. The Geekbench OpenCL score of 38,224 versus 10,273 is a 73.1% advantage. The Vulkan score of 36,736 further confirms its strength in modern APIs. Its 8 ray tracing cores and DirectX 12 Ultimate support make it the only option for games using hardware-accelerated ray tracing, mesh shaders, or variable rate shading. The 6 GB GDDR6 frame buffer with 186.0 GB/s bandwidth allows it to handle higher resolution textures and larger scene complexity without hitting memory walls. The 2,000 MHz base clock and 2,050 MHz boost clock, combined with 4.198 TFLOPS FP32, make it suitable for GPU compute tasks like OpenCL-based rendering or Vulkan compute shaders.

The NVIDIA GeForce GTX 950A wins in scenarios where form factor and power delivery are constraints. It is an MXM module with no external power connector, drawing 75 W. This makes it a direct replacement for older MXM-based laptops that lack the physical space or power supply for a dual-slot PCIe card. Its 28 nm process and Maxwell architecture are older, but the card remains functional for legacy DirectX 9/10/11 titles that do not demand modern features. The 2 GB DDR3 memory is small but adequate for older games at lower resolutions. Its single benchmark score of 10,273 in OpenCL places it at the 48th percentile of all GPUs, which is actually 4 percentile points higher than the Arc A380's 44th percentile, indicating that relative to the entire market, the GTX 950A is not as far behind as the head-to-head delta suggests.

For users with a desktop PCIe slot and a 250 W PSU, the Arc A380 is the obvious pick. For users with an MXM slot and a strict 75 W power budget, the GTX 950A is the only option that fits. The Arc A380's PassMark DirectX 9 score of 73 matches the GTX 950A's only PassMark result (which is not recorded for the GTX 950A in this database), but the Arc A380's DirectX 10/11/12 scores are all below 40, suggesting that its legacy API performance is not a strong point. The GTX 950A has no recorded PassMark scores, so direct comparison in those workloads is not possible.

In summary, the Arc A380 is the superior GPU in nearly every measurable way, with the sole exception of its lower percentile ranking and its requirement for an external power connector. The GTX 950A is a product of its time, and the database shows it performing at a level that is competitive with other mid-range GPUs from its era, but it cannot compete with the architectural advancements of the Xe-HPG generation.