GPU Comparison

The data shows a decisive generational and architectural gap between the NVIDIA GeForce 920M and the AMD Radeon 550. Across the two shared benchmark tests, the AMD Radeon 550 wins both by massive margins, delivering roughly 2.8x to 4x the performance of the older NVIDIA part. The GeForce 920M is a legacy entry-level mobile chip from 2015, while the Radeon 550 is a 2017 desktop part built on a newer, denser process with significantly more raw compute resources. For any workload measured, the AMD Radeon 550 is categorically the faster GPU.

FAQ

Q: Which GPU is faster in Geekbench OpenCL?

A: The AMD Radeon 550 scores 10,339 in Geekbench OpenCL, while the NVIDIA GeForce 920M scores 3,725. The AMD part leads by 64%, meaning it delivers nearly triple the raw compute score of the NVIDIA chip.

Q: How big is the gap in Vulkan performance?

A: The AMD Radeon 550 scores 11,257 in Geekbench Vulkan, compared to 2,849 for the NVIDIA GeForce 920M. This represents a 74.7% advantage for the AMD card, a difference of over 4x in raw benchmark score.

Q: What is the average benchmark score for each GPU?

A: The NVIDIA GeForce 920M has an average benchmark score of 3,287, while the AMD Radeon 550 has a lower average of 2,801. Despite the Radeon 550 winning the head-to-head tests, its average is dragged down by other benchmarks in its suite, including low DirectX 10 and 9 scores.

Q: Which GPU has a higher percentile ranking among all GPUs?

A: The NVIDIA GeForce 920M ranks in the 20th percentile, while the AMD Radeon 550 ranks in the 18th percentile. This suggests the 920M slightly outperforms the 550 relative to the broader GPU landscape, despite losing the direct comparisons.

Q: How do the two compare in DirectX API support?

A: The AMD Radeon 550 supports DirectX 12 (12_0) and Vulkan 1.3, whereas the NVIDIA GeForce 920M supports DirectX 12 (11_0) and Vulkan 1.2.175. The AMD card supports a higher feature level of DirectX 12 and a newer Vulkan version.

Q: What are the memory bandwidth differences?

A: The AMD Radeon 550 uses GDDR5 memory with a bandwidth of 56.00 GB/s, while the NVIDIA GeForce 920M uses DDR3 with a bandwidth of 14.40 GB/s. The AMD GPU provides roughly 3.9x the memory bandwidth, which is a critical factor in its benchmark dominance.

Architecture Differences

The two GPUs represent distinct architectural generations and design philosophies. The NVIDIA GeForce 920M is built on the Kepler 2.0 architecture, using the GK208B chip manufactured on a 28 nm process at TSMC. It packs 1,020 million transistors into a die size of 87 mm², yielding a transistor density of 11.7M per mm². In contrast, the AMD Radeon 550 uses the GCN 4.0 architecture with the Lexa chip, fabricated on a 14 nm process at GlobalFoundries. This newer process allows AMD to fit 2,200 million transistors into a 103 mm² die, achieving a much higher transistor density of 21.4M per mm².

The compute resources differ sharply. The AMD Radeon 550 features 512 shading units, 32 texture mapping units, and 16 ROPs. The NVIDIA GeForce 920M has 384 shading units, 32 TMUs, but only 8 ROPs. This doubling of ROPs on the AMD side directly contributes to its much higher pixel fill rate. The AMD card also has a significant clock speed advantage, with a base clock of 1100 MHz and a boost clock of 1183 MHz, versus the NVIDIA chip’s flat 954 MHz base and boost.

Memory architecture is another major divide. The AMD Radeon 550 uses GDDR5 memory running at 1750 MHz with a 7 Gbps effective data rate, while the NVIDIA GeForce 920M uses DDR3 at 900 MHz with a 1800 Mbps effective rate. Both use a 64-bit bus, but the GDDR5 standard gives AMD a 56.00 GB/s bandwidth versus NVIDIA’s 14.40 GB/s. The AMD card also supports FP16 at a 1:1 ratio with FP32 (1,211.4 GFLOPS each), while the NVIDIA part has no FP16 capability listed.

Feature support also differs. The AMD Radeon 550 supports DirectX 12 (12_0) and Vulkan 1.3, while the NVIDIA GeForce 920M is limited to DirectX 12 (11_0) and Vulkan 1.2.175. The AMD card is a dual-slot desktop card with display outputs including 1x DVI, 1x HDMI 2.0b, and 1x DisplayPort 1.4a, whereas the NVIDIA part is an integrated GPU (IGP) with portable-device-dependent outputs. The AMD GPU also lists a suggested PSU of 250 W, while the NVIDIA chip has no such requirement.

Head-to-Head Benchmarks

The direct comparison is stark. In Geekbench OpenCL, the AMD Radeon 550 scores 10,339 against the NVIDIA GeForce 920M’s 3,725. This is a 64% delta, meaning the AMD card is approximately 2.8x faster in raw OpenCL compute. The gap is even larger in Geekbench Vulkan, where the AMD Radeon 550 scores 11,257 versus 2,849 for the NVIDIA part, a 74.7% delta or roughly 4x the performance.

These results are consistent with the underlying specifications. The AMD card’s higher shading unit count (512 vs 384), dramatically higher clock speeds (1183 MHz boost vs 954 MHz), and far superior memory bandwidth (56.00 GB/s vs 14.40 GB/s) all contribute to its dominance. The NVIDIA GeForce 920M’s only comparative strengths are its slightly better percentile ranking (20th vs 18th) and higher average benchmark score (3,287 vs 2,801), but these are aggregate metrics that include tests where the AMD card’s suite shows weak results, such as PassMark DirectX 10 scoring only 7 and DirectX 9 scoring 35.

The AMD Radeon 550 also benefits from a much higher pixel rate (18.93 GPixel/s vs 7.632 GPixel/s) and texture rate (37.86 GTexel/s vs 30.53 GTexel/s). Its FP32 throughput of 1,211.4 GFLOPS is nearly double the NVIDIA chip’s 732.7 GFLOPS. Every direct benchmark comparison favors the AMD card, with no test in the head-to-head data showing a win for the NVIDIA GeForce 920M.

Specification Differences

The two GPUs differ across nearly every major specification field. The NVIDIA GeForce 920M uses a 28 nm process from TSMC, while the AMD Radeon 550 uses a 14 nm process from GlobalFoundries. Transistor count is 1,020 million for NVIDIA versus 2,200 million for AMD. Die size is 87 mm² versus 103 mm². Transistor density is 11.7M per mm² versus 21.4M per mm².

Clock speeds differ significantly: the NVIDIA chip has a base and boost clock of 954 MHz, while the AMD card has a base clock of 1100 MHz and a boost clock of 1183 MHz. Memory clocks are 900 MHz (1800 Mbps effective) for NVIDIA versus 1750 MHz (7 Gbps effective) for AMD. Memory bandwidth is 14.40 GB/s versus 56.00 GB/s.

The shading unit count is 384 for NVIDIA versus 512 for AMD. TMU count is the same at 32 for both, but ROPs differ: 8 for NVIDIA versus 16 for AMD. Pixel rate is 7.632 GPixel/s versus 18.93 GPixel/s. Texture rate is 30.53 GTexel/s versus 37.86 GTexel/s. FP32 performance is 732.7 GFLOPS versus 1,211.4 GFLOPS. The AMD card also lists FP16 at 1,211.4 GFLOPS (1:1), while NVIDIA has no FP16 entry.

TDP is 33 W for NVIDIA versus 50 W for AMD. The NVIDIA part is an IGP with no power connectors, while the AMD card is dual-slot with no power connectors but a suggested PSU of 250 W. Bus interface is PCIe 3.0 x8 for both. Display outputs are portable-device-dependent for NVIDIA versus 1x DVI, 1x HDMI 2.0b, and 1x DisplayPort 1.4a for AMD. DirectX support is 12 (11_0) versus 12 (12_0). Vulkan support is 1.2.175 versus 1.3. The AMD card has a physical length of 145 mm (5.7 inches), while the NVIDIA part has no listed dimensions. Release dates are 2015-03-12 for NVIDIA versus 2017-04-19 for AMD.

The Verdict

The benchmark data is unambiguous: the AMD Radeon 550 is the faster GPU by a wide margin in every direct comparison. It wins both head-to-head tests, with a 64% lead in OpenCL and a 74.7% lead in Vulkan. The AMD card’s superior architecture, newer process node, higher clock speeds, doubled ROP count, and nearly 4x memory bandwidth make it the clear choice for any workload that leverages these features.

The NVIDIA GeForce 920M does maintain a higher average benchmark score (3,287 vs 2,801) and a slightly better percentile ranking (20th vs 18th). However, these figures are influenced by the AMD card’s weak performance in legacy DirectX tests, such as a PassMark DirectX 10 score of 7 and a DirectX 9 score of 35. In modern compute and graphics APIs, the NVIDIA part is simply outclassed.

For users choosing between these two, the data suggests the AMD Radeon 550 is the superior option for OpenCL and Vulkan workloads. The NVIDIA GeForce 920M may be acceptable for legacy DirectX applications, but its performance ceiling is far lower. The AMD card also offers newer API support, including DirectX 12 (12_0) and Vulkan 1.3, making it more future-proof for modern software.

Where Each One Wins

AMD Radeon 550: The AMD card wins decisively in Geekbench OpenCL (10,339 vs 3,725) and Geekbench Vulkan (11,257 vs 2,849). It is the clear choice for compute-heavy tasks, modern game APIs, and any application that utilizes GDDR5 memory bandwidth. Its 2x FP32 throughput (1,211.4 GFLOPS vs 732.7 GFLOPS), higher pixel rate (18.93 GPixel/s vs 7.632 GPixel/s), and doubled ROPs make it better suited for resolution scaling and fill-rate-bound scenarios. The AMD card also supports newer API versions, including DirectX 12 (12_0) and Vulkan 1.3, which are critical for recent titles.

NVIDIA GeForce 920M: The NVIDIA part has no direct benchmark wins in the head-to-head data. Its only statistical advantages are a higher average benchmark score (3,287 vs 2,801) and a higher percentile ranking (20th vs 18th). These suggest it may be more consistent across a broader range of legacy tests, but no specific test in the provided data shows it outperforming the AMD Radeon 550. The NVIDIA card also has a lower TDP (33 W vs 50 W) and no suggested PSU requirement, making it a more power-efficient option for thin, portable devices. However, based purely on performance metrics, the NVIDIA GeForce 920M is the weaker GPU in all measured scenarios.