NVIDIA GeForce GT 1010 vs NVIDIA GeForce GTX 950M Comparison

The NVIDIA GeForce GTX 950M and the NVIDIA GeForce GT 1010 sit in adjacent rows of the database yet belong to entirely different eras and design philosophies. One is a Maxwell-based mobile part from the GeForce 900M generation, manufactured by TSMC on a 28 nm process and carrying 640 shading units fed by 4 GB of DDR3 memory. The other is a Pascal-based desktop card from the GeForce 10 generation, built by Samsung on a 14 nm process with a far smaller die, fewer resources, and faster GDDR5 attached to a narrow 64-bit bus. The recorded benchmark data gives the GTX 950M a clear lead in compute, but the underlying hardware tells a more nuanced story about efficiency, memory bandwidth, and feature support.

Head-to-Head Benchmarks

The database contains one direct comparison test, and it is decisive. In Geekbench OpenCL, the GTX 950M scores 9745 against 6698 for the GT 1010, a 45.5 percent advantage for the older Maxwell part. That margin is substantial: it means the GTX 950M delivers nearly half again the GPU compute throughput of the GT 1010 in general-purpose workloads, despite being the older design by roughly six years of release dates.

The supporting averages reinforce the picture. The GTX 950M's average benchmark score across the database is 8135, placing it in the 42nd percentile of all recorded GPUs. The GT 1010 averages 6698 and sits in the 38th percentile. The gap between the two percentile positions is modest, four points, but the average score difference of 1437 points is not.

Context from each card's nearest rivals sharpens the comparison. The GTX 950M's average of 8135 sits within a fraction of a percent of the AMD Radeon R9 M360 at 8129 (a 0.1 percent delta), the NVIDIA GeForce GTX 980 at 8167 (-0.4 percent), the NVIDIA GeForce 945M at 8099 (0.4 percent), and the NVIDIA GRID K2 at 8080 (0.7 percent). The GT 1010's average of 6698 is likewise bracketed by the AMD Radeon R7 M370 at 6764 (-1 percent), the AMD Radeon R7 M460 at 6612 (1.3 percent), the AMD Radeon HD 7730M at 6581 (1.8 percent), and the AMD FirePro M5100 at 6830 (-1.9 percent). Both cards compete in tightly clustered performance neighborhoods, but the GTX 950M's neighborhood sits noticeably higher.

The GTX 950M also holds a second recorded result the GT 1010 lacks: a Geekbench Vulkan score of 6525. There is no corresponding GT 1010 Vulkan entry in the database, so no direct comparison can be drawn there, but the presence of the result indicates the Maxwell part was exercised across more API pathways in the recorded data.

The win tally in the head-to-head set is one win for the GTX 950M and zero for the GT 1010.

Architecture Differences

The architectural split is generational. The GTX 950M uses the GM107 chip on the Maxwell architecture, produced at TSMC on a 28 nm process. It packs 1,870 million transistors onto a 148 mm² die, yielding a transistor density of 12.6 million transistors per square millimeter. The GT 1010 uses the GP108 chip on the Pascal architecture, produced at Samsung on a 14 nm process. Its die carries 1,800 million transistors in just 74 mm², exactly half the GTX 950M's die area, for a density of 24.3 million transistors per square millimeter, nearly double that of the Maxwell part.

The resource allocation diverges sharply. The GTX 950M fields 640 shading units, 40 texture mapping units, and 16 render output units. The GT 1010 fields 256 shading units, 16 TMUs, and 8 ROPs, exactly 40 percent of the Maxwell card's count across all three resource categories. The theoretical rates follow: the GTX 950M posts 17.98 GPixel/s of pixel fill and 44.96 GTexel/s of texture fill, versus 11.74 GPixel/s and 23.49 GTexel/s for the GT 1010. FP32 compute stands at 1,438.7 GFLOPS for the GTX 950M against 751.6 GFLOPS for the GT 1010, a raw throughput advantage of roughly 1.9x for the Maxwell part, which aligns closely with the measured OpenCL result.

Clocks run in the GT 1010's favor but cannot overcome the resource deficit. The Pascal card boosts to 1468 MHz from a 1228 MHz base, while the GTX 950M boosts to 1124 MHz from a base of 993 MHz. The GTX 950M compensates with far wider parallelism.

Memory design points the other way. The GTX 950M pairs 4 GB of DDR3 with a 128-bit bus running at an effective 1800 Mbps, for 28.80 GB/s of bandwidth. The GT 1010 pairs 2 GB of GDDR5 with a 64-bit bus at an effective 6 Gbps, for 48.06 GB/s of bandwidth. The Pascal card delivers substantially more bandwidth from a much smaller memory footprint, while the Maxwell card offers double the capacity, which matters for texture-heavy or memory-hungry workloads.

Power draw is a defining contrast. The GTX 950M is rated at 75 W and relies on power delivery from its host platform, with no external power connectors listed. The GT 1010 draws just 30 W, also with no power connectors, and carries a suggested 200 W system power supply. It is a single-slot card measuring 147 mm in length, whereas the GTX 950M is classified as an IGP-style integrated solution with portable-device-dependent display outputs. Bus interfaces differ too: PCIe 3.0 x8 for the GTX 950M versus PCIe 3.0 x4 for the GT 1010.

Feature support slightly favors the newer card. Both report DirectX 12, OpenGL 4.6, and Vulkan 1.4, but the GT 1010's DirectX feature level is 12_1 versus 11_0 for the GTX 950M. Neither card has RT cores or tensor cores. Both are end-of-life products: the GTX 950M was released in March 2015 and the GT 1010 in January 2021.

Where Each One Wins

The GTX 950M wins on raw compute and graphics throughput. Its 45.5 percent Geekbench OpenCL victory, its higher FP32 rating, doubled pixel and texture fill rates, and quadrupled shading resources make it the stronger option for GPU compute workloads, rendering, and any task where parallel throughput dominates. Its 4 GB of memory also gives it headroom the GT 1010's 2 GB cannot match when working sets grow large, and its wider 128-bit bus suits bandwidth-hungry shading loads even at lower effective memory speeds.

The GT 1010 wins on efficiency and bandwidth-per-resource. Its 30 W TDP against the GTX 950M's 75 W means it delivers usable Pascal-era performance at well under half the power draw, and its 48.06 GB/s of GDDR5 bandwidth actually exceeds the Maxwell card's 28.80 GB/s despite the narrower bus. For display duties, light compute, and space-constrained builds, its single-slot 147 mm form factor, DVI and mini-HDMI 2.0 outputs, and higher DirectX feature level of 12_1 give it practical advantages the older mobile part cannot offer. The GTX 950M's portable-device-dependent outputs underscore that it was never intended as a standalone desktop card.

Against their respective rival sets, both cards sit near the middle of tightly packed fields, but the GTX 950M's field is the stronger one, trading places with cards like the GTX 980 database entry and the R9 M360 at sub-percent deltas.

FAQ

Q: Which GPU is faster in benchmarks?

A: The GTX 950M. It won the Geekbench OpenCL head-to-head 9745 to 6698, a 45.5 percent margin, and holds a higher database average of 8135 versus 6698.

Q: How much more compute throughput does the GTX 950M have?

A: Its FP32 rating is 1,438.7 GFLOPS versus 751.6 GFLOPS for the GT 1010, close to a 1.9x advantage, consistent with the measured OpenCL result.

Q: Which card has more memory?

A: The GTX 950M, with 4 GB of DDR3. The GT 1010 has 2 GB of GDDR5.

Q: Which card has more memory bandwidth?

A: The GT 1010, with 48.06 GB/s from its GDDR5 running at 6 Gbps effective, compared with 28.80 GB/s for the GTX 950M's DDR3 at 1800 Mbps effective.

Q: How do their power requirements compare?

A: The GT 1010 is rated at 30 W with no power connectors and a suggested 200 W power supply. The GTX 950M is rated at 75 W and also lists no power connectors.

Q: Are these cards still in production?

A: No. Both are end-of-life. The GTX 950M launched in March 2015 and the GT 1010 in January 2021.

The Verdict

The data points to the GTX 950M for anyone whose priority is performance. It wins the only recorded head-to-head test by 45.5 percent, holds a higher percentile position at 42 versus 38, and doubles the GT 1010's shading units, TMUs, ROPs, and compute output. Workloads that lean on GPU compute or fill rate favor it decisively, and its 4 GB memory buffer adds capacity the GT 1010 lacks.

The GT 1010 is the pick when efficiency and practicality matter more than throughput. It draws 30 W instead of 75 W, offers more memory bandwidth despite its smaller footprint, ships as a short single-slot desktop card with DVI and mini-HDMI 2.0 outputs, and supports the newer DirectX 12_1 feature level. For a compact build or a display-and-light-compute role, those attributes count for more than the benchmark gap.

Both are end-of-life parts occupying the lower-middle of the database's historical rankings, separated by four percentile points and one clear benchmark result in the Maxwell card's favor.