GPU Comparison

NVIDIA GeForce 920M and NVIDIA GeForce RTX 5080 SUPER represent two extremes of the GPU spectrum, separated by a decade of architectural evolution and a massive gulf in compute capability. The 920M is a 2015 entry-level mobile part built on the Kepler 2.0 architecture with 384 shading units, while the RTX 5080 SUPER is a 2025 flagship desktop-class GPU built on the Blackwell 2.0 architecture packing 10,752 shading units. Direct head-to-head benchmark data does not exist between these two products, as they were never tested against each other in the same suite. However, the available individual benchmark results and architectural specifications reveal a performance chasm so wide that the comparison is almost purely academic.

Head-to-Head Benchmarks

The benchmark data available for these two GPUs comes from entirely different test suites, making direct numerical comparison impossible. The 920M was tested in Geekbench OpenCL and Vulkan, scoring 3725 and 2849 respectively, yielding an average benchmark score of 3287. The RTX 5080 SUPER was tested in 3DMark Steel Nomad DX12, scoring 3075, which serves as its average benchmark score. These are not comparable metrics, Geekbench measures compute workloads while 3DMark Steel Nomad targets modern gaming DX12 performance, so the raw numbers cannot be directly pitted against each other.

What the data does show is the positioning of each card relative to its own era and peer group. The 920M sits at the 20th percentile of all GPUs, meaning it outperforms only about one-fifth of the graphics cards in the database. Its nearest rivals in average score are the Intel HD Graphics 530 at 3332 (1.4% higher), the NVIDIA Quadro P1000 at 3163 (which the 920M trails by 2.8% in the rival comparison), and the NVIDIA GeForce GT 730M at 3316 (0.9% higher). The 920M also edges out the NVIDIA GeForce GT 640 by 2.4% and the Intel HD Graphics P4600 by 3%, showing it clusters tightly with low-end integrated and entry-level discrete GPUs from its generation.

The RTX 5080 SUPER presents a more peculiar picture. Despite being a flagship Blackwell 2.0 part with 45,600 million transistors, it sits at only the 19th percentile of all GPUs in the database. Its nearest rivals include the NVIDIA GeForce 820A at 2983 (which the RTX 5080 SUPER leads by 3.1%), the NVIDIA GeForce GTX 860M at 2967 (leading by 3.6%), and the Intel Arc Pro B60 at 3182 (trailing by 3.4%). The RTX 5080 SUPER also comes within 2.8% of the Quadro P1000, another low-end part. This anomalous percentile ranking likely reflects the benchmark database's composition, the 3DMark Steel Nomad DX12 test is a demanding next-generation workload, and many older or less capable GPUs may not have been tested in it, skewing the percentile distribution.

The performance delta between the two cards becomes apparent through architectural math rather than direct benchmarks. The 920M delivers 732.7 GFLOPS of FP32 compute, while the RTX 5080 SUPER delivers 56.28 TFLOPS, a difference that can be expressed as the RTX 5080 SUPER being roughly 76 times more powerful in raw shader throughput. Texture rate tells a similar story: 30.53 GTexel/s versus 879.3 GTexel/s. Pixel rate jumps from 7.632 GPixel/s to 293.1 GPixel/s. These are not rival comparisons but direct specification deltas, and they paint a picture of generational leap that defies simple percentage framing.

The Verdict

The data unambiguously shows that the RTX 5080 SUPER is the superior product by every measurable compute and graphics metric. Its 56.28 TFLOPS FP32 performance, 1.02 TB/s memory bandwidth, and 24 GB of GDDR7 memory place it in a completely different performance class than the 920M's 732.7 GFLOPS, 14.40 GB/s bandwidth, and 2 GB of DDR3. The RTX 5080 SUPER also supports modern feature sets including 84 ray tracing cores and 336 tensor cores, while the 920M has no such hardware. The 920M's only advantages are its 33 W TDP versus 415 W, its integrated form factor versus dual-slot, and its lack of external power connectors versus a 1x 16-pin requirement, qualities that matter only for power-constrained or legacy systems.

Who should pick the 920M? The data suggests only users with a 2015-era laptop that has no other graphics solution and needs basic display output. Its Geekbench OpenCL score of 3725 and Vulkan score of 2849 indicate it can handle lightweight compute tasks, and its 20th percentile ranking places it above only the most basic integrated graphics. It supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, so it can run older games at low settings, but it will struggle with anything modern.

Who should pick the RTX 5080 SUPER? The data supports its use for high-end gaming and professional workloads given its 3DMark Steel Nomad DX12 score of 3075, its 84 RT cores for real-time ray tracing, and its 24 GB frame buffer for large assets. Its 19th percentile ranking is misleading given its specifications; the 3DMark Steel Nomad test is a stress test that even many contemporary GPUs fail. The launch MSRP is 999 USD, making it a premium product, but the raw compute capabilities justify the positioning for users who need maximum performance.

Where Each One Wins

The 920M wins in power efficiency and form factor flexibility. At 33 W TDP, it requires no external power connectors and fits as an integrated graphics processor, making it suitable for thin-and-light laptops from its era. Its 87 mm² die size and 1,020 million transistors on a 28 nm TSMC process allow for minimal thermal footprint. For users with legacy 900M-series laptops, the 920M provides basic display output and light 2D acceleration without taxing the system's cooling solution.

The RTX 5080 SUPER wins in every performance category that matters for modern applications. Its 1.02 TB/s memory bandwidth is essential for high-resolution textures and large datasets. Its 112 ROPs and 336 TMUs handle rasterization and texturing at resolutions the 920M cannot even display. The 56.28 TFLOPS FP16 performance, identical to its FP32 output at a 1:1 ratio, accelerates AI inference and machine learning workloads. Its 84 RT cores bring hardware-accelerated ray tracing, a feature entirely absent from the Kepler 2.0 architecture. The PCIe 5.0 x16 interface offers 16 lanes of fifth-generation bandwidth versus the 920M's PCIe 3.0 x8, reducing data transfer bottlenecks for CPU-GPU communication.

FAQ

Q: Which GPU has higher raw FP32 compute performance?

A: The RTX 5080 SUPER delivers 56.28 TFLOPS compared to the 920M's 732.7 GFLOPS, making it approximately 76 times faster in single-precision floating-point operations.

Q: What are the memory specifications of each card?

A: The 920M has 2 GB of DDR3 memory on a 64-bit bus with 14.40 GB/s bandwidth. The RTX 5080 SUPER has 24 GB of GDDR7 memory on a 256-bit bus with 1.02 TB/s bandwidth.

Q: Do both GPUs support ray tracing?

A: No. The RTX 5080 SUPER has 84 dedicated RT cores for hardware-accelerated ray tracing. The 920M has no RT cores, as the Kepler 2.0 architecture predates ray tracing hardware.

Q: How do their transistor counts and die sizes compare?

A: The 920M uses 1,020 million transistors on an 87 mm² die. The RTX 5080 SUPER uses 45,600 million transistors on a 378 mm² die, leading to a transistor density of 120.6M per mm² versus 11.7M per mm² for the 920M.

Q: What is the power consumption difference?

A: The 920M has a 33 W TDP and requires no power connectors. The RTX 5080 SUPER has a 415 W TDP and requires a single 16-pin power connector.

Q: Which API versions does each GPU support?

A: The 920M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The RTX 5080 SUPER supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Architecture Differences

The 920M is built on the Kepler 2.0 architecture using the GK208B chip, fabricated on a 28 nm TSMC process. This architecture focuses on power efficiency for mobile devices, with 384 shading units, 32 TMUs, and 8 ROPs. It has no RT cores or tensor cores, as those technologies did not exist in consumer GPUs in 2015. The 920M supports DirectX 12 (11_0), meaning it can run DX12 titles but only at the 11_0 feature level, missing many modern rendering features. Its memory subsystem uses DDR3, which is slow by modern standards but was appropriate for its entry-level positioning.

The RTX 5080 SUPER is built on the Blackwell 2.0 architecture using the GB203 chip, fabricated on a 5 nm TSMC process. This represents a dramatic shift in both manufacturing and design philosophy. It packs 10,752 shading units, 336 TMUs, and 112 ROPs, along with 84 RT cores and 336 tensor cores for ray tracing and AI acceleration. The Blackwell 2.0 architecture supports DirectX 12 Ultimate (12_2), which enables features like mesh shaders, variable rate shading, and hardware-accelerated ray tracing. Its memory subsystem uses GDDR7 with a 256-bit bus, and its PCIe 5.0 x16 interface doubles the bandwidth available to the 920M's PCIe 3.0 x8. The transistor density jump from 11.7M to 120.6M per mm² reflects the process node shrink and architectural advances over the decade separating the two products.

Specification Differences

The two GPUs differ in nearly every specification field. The 920M has a base and boost clock of 954 MHz, while the RTX 5080 SUPER has a base clock of 2295 MHz and a boost clock of 2617 MHz. Memory clocks differ substantially: 900 MHz (1800 Mbps effective) for the 920M versus 2000 MHz (32 Gbps effective) for the RTX 5080 SUPER. The 920M has 2 GB of DDR3 memory, while the RTX 5080 SUPER has 24 GB of GDDR7. The shading unit count jumps from 384 to 10,752, TMUs from 32 to 336, and ROPs from 8 to 112. The RTX 5080 SUPER adds 84 RT cores and 336 tensor cores, neither of which exist on the 920M.

The pixel rate goes from 7.632 GPixel/s to 293.1 GPixel/s, and texture rate from 30.53 GTexel/s to 879.3 GTexel/s. The 920M has a 64-bit memory bus, while the RTX 5080 SUPER has a 256-bit bus. Bandwidth increases from 14.40 GB/s to 1.02 TB/s. The 920M is an integrated graphics processor with no slot width or power connectors, while the RTX 5080 SUPER is a dual-slot card measuring 304 mm by 137 mm by 40 mm, requiring a 16-pin connector. The 920M uses PCIe 3.0 x8, while the RTX 5080 SUPER uses PCIe 5.0 x16. Display outputs differ as well: the 920M is portable-device dependent, while the RTX 5080 SUPER offers 1x HDMI 2.1b and 3x DisplayPort 2.1b. Production status also differs, with the 920M marked as end-of-life and the RTX 5080 SUPER active, and the RTX 5080 SUPER has a launch MSRP of 999 USD while the 920M has no listed MSRP.