GPU Comparison

The NVIDIA GeForce 820A and NVIDIA GeForce GT 720M are both end-of-life mobile graphics solutions, but the benchmark data shows a clear split in their strengths. The 820A wins the only available head-to-head benchmark, the Geekbench OpenCL test, with a 14.5% lead over the GT 720M. However, the GT 720M counters with a newer architecture, higher core counts, and more memory, making it a more capable option for certain workloads. The data indicates the 820A is the faster part in general compute, while the GT 720M offers a feature set that the 820A lacks.

Where Each One Wins

The GeForce 820A secures the single recorded benchmark victory. In the Geekbench OpenCL test, the 820A scores 2983, while the GT 720M scores 2605. This is a 14.5% advantage for the 820A, placing it in a higher performance tier. The 820A’s nearest rivals include the GeForce GTX 860M (score 2967, only 0.5% behind), the GeForce GTX 750 Ti (score 2953, 1% behind), and the Quadro P600 (score 2923, 2.1% behind). This proximity to significantly more powerful desktop and mobile parts suggests the 820A’s compute performance is unexpectedly robust for its class.

The GT 720M, despite losing the head-to-head, has its own domain of superiority based on its specifications. It offers a 2 GB memory capacity, double the 1024 MB of the 820A. For workloads that are memory-capacity bound, such as holding larger textures or datasets, the GT 720M has a definitive advantage. Its shading unit count is also double that of the 820A (192 vs. 96), which can translate to better performance in tasks that scale with shader count, even if the overall compute score lags. The GT 720M’s nearest rivals, including the GeForce GT 1030 (score 2662, with the 720M being 2.1% behind), show it is competitive within its own performance bracket.

The performance percentile ranking reinforces this split. The 820A sits in the 19th percentile of all GPUs, while the GT 720M sits in the 17th percentile. This 2-percentage-point difference in overall standing is consistent with the 14.5% lead in the OpenCL test, confirming that the 820A is the more performant part on average. However, the GT 720M’s architectural advantages in core count and memory size mean it is not without merit for specific use cases.

Architecture Differences

The two GPUs are built on fundamentally different architectures, which explains their divergent characteristics. The GeForce 820A uses the GF117 chip based on the Fermi 2.0 architecture, while the GeForce GT 720M uses the GK208 chip based on the Kepler 2.0 architecture. Both are manufactured by TSMC on the same 28 nm process node, but the transistor counts differ dramatically. The GF117 packs 585 million transistors on a 116 mm² die, resulting in a transistor density of 5.0M / mm². In contrast, the GK208 contains 1,020 million transistors on a smaller 87 mm² die, achieving a much higher density of 11.7M / mm².

This density difference is significant. The GT 720M crams nearly twice the transistors into a 25% smaller die area. This allows the Kepler-based chip to feature 192 shading units, compared to just 96 on the Fermi-based 820A. Both parts have 16 texture mapping units and 8 ROPs, but the GT 720M’s higher clock speeds, a base of 719 MHz and boost of 758 MHz, versus the 820A’s unspecified base and boost clocks, enable it to achieve higher pixel and texture fill rates. Specifically, the GT 720M posts a 3.032 GPixel/s pixel rate and a 12.13 GTexel/s texture rate, compared to the 820A’s 2.500 GPixel/s and 10.00 GTexel/s.

Memory configurations also differ. The GT 720M uses 2 GB of DDR3 on a 64-bit bus at 800 MHz (1600 Mbps effective), yielding a bandwidth of 12.80 GB/s. The 820A uses 1 GB of DDR3 on the same 64-bit bus but runs at 1001 MHz (2 Gbps effective), achieving a higher bandwidth of 16.02 GB/s. This means the 820A has more memory bandwidth, while the GT 720M has more capacity. In terms of API support, the GT 720M adds Vulkan 1.2.175 support, while the 820A lists none. Both support DirectX 12 (11_0) and OpenGL 4.6.

Head-to-Head Benchmarks

The only direct benchmark comparison available is the Geekbench OpenCL test, and it is a clear win for the GeForce 820A. The 820A scores 2983, while the GT 720M scores 2605, giving the 820A a 14.5% advantage. This is a substantial margin in the world of low-power mobile GPUs, indicating that the Fermi architecture in the 820A is better optimized for the OpenCL compute workload than the Kepler architecture in the GT 720M.

To put the 820A’s score in context, its nearest rival is the GeForce GTX 860M, which averages 2967. The 820A beats that part by 0.5%. It also edges out the GeForce GTX 750 Ti (2953) by 1% and the Quadro P600 (2923) by 2.1%. Even more surprisingly, it outperforms the GeForce RTX 4060 Ti 8 GB, which scores 2913, by 2.4%. This suggests that in this specific compute benchmark, the 820A punches far above its weight class.

The GT 720M’s score of 2605 places it near a different set of rivals. It trails the GeForce GT 440 (2645) by 1.5%, the Intel Arc Pro B50 (2660) by 2.1%, and the GeForce GT 1030 (2662) by 2.1%. It only beats the Quadro K2000M (2555) by 2%. The data shows the GT 720M is competitive with entry-level desktop parts, but it is clearly behind the 820A in raw compute throughput. The 14.5% delta between the two is the single largest performance gap recorded in the head-to-head data.

FAQ

Q: Which GPU has the higher benchmark score?

A: The NVIDIA GeForce 820A scores 2983 on the Geekbench OpenCL test, which is 14.5% higher than the NVIDIA GeForce GT 720M’s score of 2605.

Q: Does the GT 720M have any advantage in memory?

A: Yes, the GT 720M has 2 GB of DDR3 memory, which is double the 1024 MB found on the 820A. However, the 820A has higher memory bandwidth at 16.02 GB/s compared to the GT 720M’s 12.80 GB/s.

Q: What are the key architectural differences?

A: The 820A uses the GF117 chip with Fermi 2.0 architecture, featuring 96 shading units. The GT 720M uses the GK208 chip with Kepler 2.0 architecture, featuring 192 shading units. The GT 720M also has a higher transistor count of 1,020 million versus 585 million.

Q: Which GPU has better pixel and texture fill rates?

A: The GT 720M has higher fill rates, with a pixel rate of 3.032 GPixel/s and a texture rate of 12.13 GTexel/s, compared to the 820A’s 2.500 GPixel/s and 10.00 GTexel/s.

Q: How does the 820A compare to its nearest rivals?

A: The 820A scores 0.5% higher than the GeForce GTX 860M, 1% higher than the GeForce GTX 750 Ti, 2.1% higher than the Quadro P600, and 2.4% higher than the GeForce RTX 4060 Ti 8 GB.

Q: Does the GT 720M support Vulkan?

A: Yes, the GT 720M supports Vulkan 1.2.175, while the 820A does not list Vulkan support in its specifications.

Specification Differences

The two GPUs differ on several key specification fields. The most glaring difference is in the chip and architecture: the 820A uses the GF117 chip based on Fermi 2.0, while the GT 720M uses the GK208 chip based on Kepler 2.0. Transistor counts differ substantially, with the 820A at 585 million and the GT 720M at 1,020 million. The die size also varies, with the 820A measuring 116 mm² and the GT 720M measuring 87 mm², leading to a transistor density of 5.0M / mm² for the 820A and 11.7M / mm² for the GT 720M.

Clock speeds are only specified for the GT 720M, which has a base clock of 719 MHz and a boost clock of 758 MHz; the 820A does not list base or boost clocks. Memory clocks also differ, with the 820A running at 1001 MHz (2 Gbps effective) and the GT 720M at 800 MHz (1600 Mbps effective). Memory capacity is 1024 MB for the 820A versus 2 GB for the GT 720M, while bandwidth is 16.02 GB/s for the 820A and 12.80 GB/s for the GT 720M. The shading unit count is 96 for the 820A and 192 for the GT 720M. Pixel rate is 2.500 GPixel/s for the 820A and 3.032 GPixel/s for the GT 720M, and texture rate is 10.00 GTexel/s versus 12.13 GTexel/s. FP32 performance is 240.0 GFLOPS for the 820A and 291.1 GFLOPS for the GT 720M. The TDP is 15 W for the 820A and 33 W for the GT 720M. The slot width is IGP for the 820A and MXM Module for the GT 720M. The bus interface is PCIe 2.0 x16 for the 820A and PCIe 2.0 x8 for the GT 720M. The GT 720M also supports Vulkan 1.2.175, which the 820A does not. The GT 720M was released on December 25, 2013, while the 820A was released on March 17, 2014.

The Verdict

The data points to a clear choice for users prioritizing compute performance: the GeForce 820A. It wins the only benchmark recorded, and its 14.5% lead over the GT 720M is decisive. The 820A’s ability to outscore rivals like the GeForce GTX 860M and GeForce RTX 4060 Ti 8 GB in the OpenCL test makes it the superior part for general compute and OpenCL-accelerated workloads. Its lower TDP of 15 W, compared to the GT 720M’s 33 W, also makes it a more power-efficient option for thin-and-light notebooks.

However, the GT 720M is not without a case. Its double the memory capacity (2 GB vs. 1 GB) makes it the better choice for applications that require larger frame buffers, such as higher-resolution textures or multi-monitor setups. Its higher shading unit count (192 vs. 96) and higher fill rates (3.032 GPixel/s and 12.13 GTexel/s) suggest it may render certain graphical effects more efficiently, despite losing the overall compute test. The GT 720M also offers Vulkan support, which the 820A lacks, making it the more future-proof option for games and applications that leverage that API.

For a user who values raw compute speed and power efficiency, the GeForce 820A is the recommended choice. For a user who needs more memory capacity and wants Vulkan support with a denser, more modern architecture, the GeForce GT 720M is the better fit. The benchmark data is unambiguous about the 820A’s speed, but the GT 720M’s specifications provide capabilities that the 820A simply cannot match.