Intel Arc A310 vs NVIDIA GeForce GTX 650 Ti Comparison

The NVIDIA GeForce GTX 650 Ti and Intel Arc A310 represent two very different eras of GPU design, and the benchmark data reflects a decisive generational shift. In the two head-to-head tests available, the Intel Arc A310 dominates completely, winning both Geekbench OpenCL and Vulkan tests. The Geekbench OpenCL score shows the Arc A310 at 30,607 against the GTX 650 Ti’s 7,877, a delta of -74.3% from the perspective of the NVIDIA card, meaning the Intel part is roughly 3.9 times faster. The Vulkan test tells a similar story: the Arc A310 scores 28,964 versus 8,229 for the GTX 650 Ti, a -71.6% delta. These are not marginal improvements; they are overwhelming victories that suggest the older Kepler architecture is simply outclassed in compute-heavy workloads.

Head-to-Head Benchmarks

The only direct comparisons available are the two Geekbench tests, and both are lopsided. In Geekbench OpenCL, the Intel Arc A310’s score of 30,607 is nearly four times the GTX 650 Ti’s 7,877. This is a massive gap that points to fundamental differences in raw compute throughput. The Arc A310’s FP32 performance is listed at 2.688 TFLOPS, while the GTX 650 Ti manages just 1,425.4 GFLOPS (or roughly 1.43 TFLOPS). That near-doubling of floating-point capability alone explains a large portion of the OpenCL delta, though the Intel card’s newer architecture likely extracts more efficiency from its shaders as well.

The Vulkan benchmark is slightly closer in percentage terms but still a clear rout. The Arc A310 posts 28,964 against the GTX 650 Ti’s 8,229, a -71.6% delta. Vulkan is a low-overhead API that tends to favor architectures with modern scheduling and command processing. The GTX 650 Ti supports Vulkan 1.2.175, while the Arc A310 supports Vulkan 1.4, and that newer specification, combined with the Xe-HPG architecture’s design for modern APIs, appears to give Intel a decisive edge. Interestingly, the GTX 650 Ti actually scores higher in Vulkan (8,229) than in OpenCL (7,877) on its own, suggesting its legacy driver overhead is less of a bottleneck in the older API. But the Arc A310’s Vulkan score is still 2,520 points higher than its OpenCL result, indicating that Intel’s driver stack and hardware are particularly well-tuned for Vulkan’s explicit control model.

Looking at the broader benchmark context, the GTX 650 Ti’s average benchmark score is 8,053, placing it in the 42nd percentile of all GPUs. Its nearest rivals include the GeForce GTX 650 Ti Boost (8,067, -0.2% delta), the GRID K2 (8,080, -0.3%), and the GTX 880M (8,040, +0.2%). These are all close contemporaries, showing that the GTX 650 Ti sits in a tight performance cluster of early-to-mid-2010s mid-range parts. The Arc A310, by contrast, has an average score of 7,550, which is actually lower than the GTX 650 Ti’s average. This is a crucial nuance: the Arc A310’s average is dragged down by its Passmark DirectX scores, which are extraordinarily low (DirectX 10: 31, DirectX 11: 33, DirectX 12: 29). These single-digit-to-low-double-digit scores appear to be anomalies or driver issues, as the Arc A310’s Geekbench scores are in the 28,000-30,000 range, far above its Passmark G3D score of 5,433. The GTX 650 Ti has no Passmark scores in the data, so its average of 8,053 is based solely on its two Geekbench results. This means the Arc A310’s average of 7,550 is misleading when comparing across the two cards’ full benchmark suites; in the head-to-head tests that both cards have, the Intel part wins by a landslide.

Architecture Differences

The architectural gap between these two GPUs spans over a decade of silicon evolution. The GTX 650 Ti uses the GK106S chip, built on a 28 nm TSMC process, with 2,540 million transistors packed into a 221 mm² die. That yields a transistor density of 11.5 million per square millimeter. The Arc A310 uses the DG2-128 chip on a 6 nm TSMC process, with 7,200 million transistors in a 157 mm² die, achieving a density of 45.9 million per square millimeter. The Intel part has nearly three times the transistor count in a smaller physical area, which explains its ability to deliver far higher compute throughput while consuming a fraction of the power.

The memory subsystems are equally divergent. The GTX 650 Ti has 1024 MB of GDDR5 on a 128-bit bus, yielding 86.40 GB/s of bandwidth at an effective 5.4 Gbps. The Arc A310 has 4 GB of GDDR6 on a 64-bit bus, but thanks to faster 15.5 Gbps effective memory, it achieves 124.0 GB/s of bandwidth. Despite half the bus width, the Intel card has 43% more bandwidth, which is critical for modern workloads that are bandwidth-hungry. The GTX 650 Ti’s memory clock is listed as 1350 MHz (5.4 Gbps effective), while the Arc A310 runs at 1937 MHz (15.5 Gbps effective).

Shader configurations tell a story of similar core counts but different efficiency. Both cards have 768 shading units, but the GTX 650 Ti pairs them with 64 TMUs and 16 ROPs, while the Arc A310 has 32 TMUs and 16 ROPs. The GTX 650 Ti’s texture rate is 59.39 GTexel/s, slightly higher than the Arc A310’s 56.00 GTexel/s. However, the Arc A310’s pixel rate of 28.00 GPixel/s is nearly double the GTX 650 Ti’s 14.85 GPixel/s. This suggests the Intel architecture can fill pixels much faster, likely due to its higher clock speeds (1750 MHz base and boost for Intel versus no listed base or boost for NVIDIA) and more efficient ROP design.

Feature support is a major differentiator. The Arc A310 includes 6 ray tracing cores, which the GTX 650 Ti completely lacks. The Intel card also supports DirectX 12 Ultimate (12_2), while the GTX 650 Ti only manages DirectX 12 (11_0). Both support OpenGL 4.6, but the Arc A310’s Vulkan 1.4 support is newer than the GTX 650 Ti’s 1.2.175. The Arc A310 also has FP16 support at 5.376 TFLOPS (2:1 ratio), while the GTX 650 Ti has no listed FP16 capability. These feature gaps mean the Arc A310 can handle modern rendering techniques like ray tracing and variable rate shading, while the GTX 650 Ti is locked to legacy feature sets.

FAQ

Q: Why is the Intel Arc A310’s average benchmark score lower than the GTX 650 Ti’s, despite winning all head-to-head tests?

A: The Arc A310’s average of 7,550 is pulled down by its Passmark DirectX scores, which range from 29 to 69. These are anomalously low compared to its Geekbench scores (30,607 OpenCL, 28,964 Vulkan) and its Passmark G3D score of 5,433. The GTX 650 Ti has no Passmark scores in the data, so its average of 8,053 only reflects its two Geekbench results.

Q: Does the GTX 650 Ti have any advantage in raw texture processing?

A: Yes, the GTX 650 Ti has a texture rate of 59.39 GTexel/s, which is 6% higher than the Arc A310’s 56.00 GTexel/s. This comes from having 64 TMUs versus the Arc A310’s 32 TMUs, but the Intel card’s higher clocks close the gap.

Q: Can the Arc A310 do ray tracing?

A: The Arc A310 includes 6 dedicated ray tracing cores, while the GTX 650 Ti has none. This is a defining architectural difference, as the Intel card can handle hardware-accelerated ray tracing workloads that are impossible on the NVIDIA Kepler part.

Q: What memory capacity difference matters most?

A: The Arc A310 has 4 GB of GDDR6, while the GTX 650 Ti has only 1024 MB of GDDR5. The Intel card also has higher bandwidth (124.0 GB/s versus 86.40 GB/s), meaning it can handle larger textures and datasets without swapping to system memory.

Q: How do power requirements compare?

A: The GTX 650 Ti has a TDP of 110 W and requires a 300 W power supply, plus a 1x 6-pin connector. The Arc A310 has a TDP of just 30 W, requires a 200 W power supply, and needs no external power connectors. The Intel card is dramatically more efficient.

Q: Which GPU has better API compliance?

A: The Arc A310 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 650 Ti only supports DirectX 12 (11_0) and Vulkan 1.2.175. The Intel card is fully compliant with modern graphics APIs, including ray tracing requirements.

Specification Differences

The two cards differ in almost every measurable specification. The GTX 650 Ti uses a 28 nm process, while the Arc A310 uses 6 nm. Transistor count is 2,540 million versus 7,200 million. Die size is 221 mm² versus 157 mm². Memory is 1024 MB GDDR5 versus 4 GB GDDR6, with bus widths of 128-bit versus 64-bit. Bandwidth is 86.40 GB/s versus 124.0 GB/s. The clock speeds differ significantly: the GTX 650 Ti has no listed base or boost clock, while the Arc A310 runs at 1750 MHz for both. The GTX 650 Ti’s memory clocks at 1350 MHz (5.4 Gbps effective), while the Arc A310’s memory runs at 1937 MHz (15.5 Gbps effective).

Shader units are identical at 768, but TMUs differ (64 versus 32) and ROPs are the same (16). Pixel rates are 14.85 GPixel/s versus 28.00 GPixel/s. Texture rates are 59.39 GTexel/s versus 56.00 GTexel/s. FP32 performance is 1,425.4 GFLOPS versus 2.688 TFLOPS. The Arc A310 has FP16 at 5.376 TFLOPS, while the GTX 650 Ti has none. TDP is 110 W versus 30 W. The GTX 650 Ti requires a 300 W PSU and 1x 6-pin connector, while the Arc A310 needs a 200 W PSU and no connector. Bus interfaces are PCIe 3.0 x16 versus PCIe 4.0 x8. Display outputs are 2x DVI and 1x mini-HDMI 1.4a versus 4x mini-DisplayPort 2.0. The GTX 650 Ti has a launch MSRP of 149 USD, while the Arc A310 has no listed launch MSRP.

The Verdict

The data is unambiguous: the Intel Arc A310 is the superior GPU for almost any modern workload. It wins both head-to-head benchmarks by margins of over 70%, offers 4 GB of memory versus 1 GB, supports ray tracing, and consumes just 30 W compared to the GTX 650 Ti’s 110 W. The GTX 650 Ti’s only wins are in texture rate (59.39 versus 56.00 GTexel/s) and its slightly higher average benchmark score (8,053 versus 7,550), but the latter is an artifact of incomplete data. The Arc A310’s nearest rivals include the GeForce GTX 1650 (7,472, +1% delta) and AMD Radeon R7 250 (7,557, -0.1%), placing it in a similar performance class to those 2019-era parts despite its low TDP. The GTX 650 Ti’s nearest rivals are all from its own era, like the GTX 650 Ti Boost and GRID K2, confirming it is a product of 2012 technology.

For anyone choosing between these two today, the Arc A310 is the clear pick for general computing, modern gaming, or any workload that benefits from Vulkan or DirectX 12 Ultimate features. The GTX 650 Ti might be relevant only for retro compatibility or systems with extremely specific driver requirements, but the data shows it is outmatched in every meaningful compute test.

Where Each One Wins

The Arc A310 wins in every benchmark category where both cards have data. In Geekbench OpenCL, it is 288% faster. In Geekbench Vulkan, it is 252% faster. It also has better memory bandwidth (124.0 GB/s versus 86.40 GB/s), higher pixel rate (28.00 GPixel/s versus 14.85 GPixel/s), and more than double the FP32 throughput (2.688 TFLOPS versus 1,425.4 GFLOPS). The Intel card is the choice for any task involving modern APIs, ray tracing, or compute workloads like machine learning inference or video encoding.

The GTX 650 Ti’s sole numerical advantage is its texture fill rate of 59.39 GTexel/s, which beats the Arc A310’s 56.00 GTexel/s by 6%. This could theoretically benefit older DirectX 9 or OpenGL games that are heavily texture-bound, but the advantage is marginal. The GTX 650 Ti also has a slightly higher average benchmark score (8,053 versus 7,550), but that is misleading because the Arc A310’s Passmark DirectX scores are clearly anomalous. In practical terms, the GTX 650 Ti wins only in legacy compatibility scenarios—systems with PCIe 3.0 slots that predate PCIe 4.0, or users needing DVI outputs instead of DisplayPort. The GTX 650 Ti’s 145 mm length and single-slot design make it physically similar to the Arc A310 (which has no listed dimensions), but its 110 W TDP and 6-pin power requirement make it a less attractive drop-in upgrade for low-power systems. The verdict is clear: the Arc A310 wins everywhere that matters, and the GTX 650 Ti is a relic of a bygone era.