GPU Comparison

The NVIDIA GeForce 820A and NVIDIA GeForce 820M are both end-of-life mobile graphics solutions from NVIDIA, but they represent fundamentally different approaches to the same problem. The data shows that despite the 820A's lower specifications in nearly every category, it is the faster GPU in the available benchmark, which raises immediate questions about what really drives performance in this class. This analysis will dissect the benchmark results, architectural choices, and specification differences to determine which part is better suited for specific workloads.

Head-to-Head Benchmarks

The only available head-to-head benchmark is the Geekbench OpenCL test, and the results are surprising. The NVIDIA GeForce 820A scores 2983 points, while the NVIDIA GeForce 820M scores 2784 points. This gives the 820A a 7.1% lead in the direct comparison. This is a counterintuitive outcome because the 820M has a dramatically more powerful compute configuration on paper. The 820M houses 384 shading units, while the 820A has only 96, a four-fold difference. Similarly, the 820M has a peak FP32 throughput of 622.1 GFLOPS, compared to just 240.0 GFLOPS for the 820A. Based on raw compute, the 820M should be significantly faster, yet the data shows the opposite.

Looking at the nearest rivals for each card provides more context for this anomaly. The 820A’s score of 2983 places it 0.5% ahead of the NVIDIA GeForce GTX 860M, which has an average score of 2967. It is also 1% ahead of the GeForce GTX 750 Ti (2953) and 2.1% ahead of the Quadro P600 (2923). Even more striking, the 820A is 2.4% faster than the GeForce RTX 4060 Ti 8 GB, which scores 2913. This is a remarkable result for a GPU with such a low transistor count and memory bandwidth. The 820M, in contrast, sits in a different performance tier. Its score of 2784 is 0.4% behind the GeForce RTX 3060 Ti GDDR6X (2795) and 0.6% behind the AMD Radeon 550 (2801). It does edge out the Intel UHD Graphics 610 by 1.2% (2752) and the GeForce RTX 5070 SUPER by 3.5% (2690).

The data suggests that the 820A's 7.1% victory is not a small margin; it is a decisive win that defies the specification sheet. The benchmark results indicate that the 820A's software stack, driver optimizations, or some other factor must be compensating for its hardware limitations. The 820M's higher theoretical compute is not translating into real-world OpenCL performance, making the 820A the clear winner in this single test.

Where Each One Wins

The benchmark data provides only one test, so the use-case split is stark. The 820A wins the sole head-to-head benchmark, giving it a 1-0 win record. This means that in the Geekbench OpenCL workload, which is a proxy for general-purpose compute performance, the 820A is the superior choice. Its 7.1% lead indicates that for tasks like OpenCL-accelerated image processing, video encoding, or other compute offloads, the 820A will deliver better performance.

The 820M does not win any benchmarks in the head-to-head comparison. However, its specification sheet suggests it may have strengths that are not captured by this single test. The 820M has 2 GB of memory compared to the 820A's 1 GB, and its memory bandwidth is 28.83 GB/s versus 16.02 GB/s. This suggests that in memory-bandwidth-intensive scenarios, such as higher-resolution textures or more complex scene geometry, the 820M could potentially perform better, even if it loses in raw compute. The 820M's higher pixel rate of 6.480 GPixel/s and texture rate of 25.92 GTexel/s also hint at better rasterization throughput. These advantages are not reflected in the OpenCL score, but they are the only data-driven reasons to favor the 820M in any workload.

For the 820A, the wins are clear. It has a higher OpenCL score and a higher percentile rank (19th vs 18th for the 820M). The data implies that for compute-heavy tasks, the 820A is the go-to part. For the 820M, its only potential wins are hypothetical, based on its superior memory subsystem and texture/pixel fill rates. Without a benchmark to confirm these advantages, the 820M must be considered the weaker choice based on the available evidence.

FAQ

Q: Which GPU has the higher benchmark score?

A: The NVIDIA GeForce 820A has the higher score in the Geekbench OpenCL test, with 2983 points compared to the 820M's 2784 points. This gives the 820A a 7.1% advantage.

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

A: The 820A's score places it ahead of several other GPUs. It is 0.5% faster than the GeForce GTX 860M, 1% faster than the GTX 750 Ti, 2.1% faster than the Quadro P600, and 2.4% faster than the RTX 4060 Ti 8 GB.

Q: Is the 820M competitive with the 820A?

A: No, the 820M is behind in the direct comparison. The 820M's score of 2784 is 7.1% lower than the 820A's. In the nearest rival list, the 820M is 0.4% slower than the RTX 3060 Ti GDDR6X and 0.6% slower than the AMD Radeon 550.

Q: What are the memory specifications for each GPU?

A: The 820A has 1024 MB of DDR3 memory on a 64-bit bus, resulting in 16.02 GB/s of bandwidth. The 820M has 2 GB of DDR3 memory on a 128-bit bus, resulting in 28.83 GB/s of bandwidth.

Q: Do both GPUs support the same APIs?

A: Both support DirectX 12 (11_0) and OpenGL 4.6. However, the 820M also supports Vulkan 1.2.175, while the 820A does not have a listed Vulkan version.

Q: What is the TDP difference between the two?

A: The 820A has a TDP of 15 W, while the 820M has a TDP of 45 W. This indicates a significant difference in power draw and thermal output.

Specification Differences

The specification sheets for these two GPUs reveal a clear contrast in almost every measurable category. The most obvious difference is in the memory subsystem. The 820A is equipped with 1024 MB of DDR3 memory on a 64-bit bus, yielding a bandwidth of 16.02 GB/s. The 820M, in contrast, has 2 GB of DDR3 memory on a 128-bit bus, providing 28.83 GB/s of bandwidth. This means the 820M has double the memory capacity and nearly double the bandwidth.

The compute units also differ greatly. The 820A has 96 shading units, 16 TMUs, and 8 ROPs. The 820M has 384 shading units, 32 TMUs, and 16 ROPs. This translates to a significant difference in fill rates. The 820A has a pixel rate of 2.500 GPixel/s and a texture rate of 10.00 GTexel/s. The 820M has a pixel rate of 6.480 GPixel/s and a texture rate of 25.92 GTexel/s. The FP32 performance is also distinct, with the 820A at 240.0 GFLOPS and the 820M at 622.1 GFLOPS.

The power and physical characteristics are different as well. The 820A has a TDP of 15 W and is an IGP (Integrated Graphics Processor) that is slot-dependent on the portable device. The 820M has a TDP of 45 W and is an MXM Module. The bus interface also differs, with the 820A using PCIe 2.0 x16 and the 820M using PCIe 3.0 x16. The release dates are not identical, with the 820A launching earlier than the 820M.

Architecture Differences

The two GPUs are built on fundamentally different architectures, which likely explains the performance discrepancy. The 820A is based on the GF117 chip using the Fermi 2.0 architecture. The 820M, on the other hand, uses the GK107 chip with the Kepler architecture.

The chip characteristics are notably different. The 820A's GF117 chip contains 585 million transistors on a 116 mm² die, resulting in a transistor density of 5.0M / mm². The 820M's GK107 chip contains 1,270 million transistors on a 118 mm² die, giving it a much higher transistor density of 10.8M / mm². Both are manufactured on a 28 nm process at TSMC, but the Kepler chip packs more than twice the transistors into a similar die area.

These architectural differences have direct implications. Fermi 2.0 is an older design, while Kepler is a more modern architecture with a different scheduling and execution model. The Kepler architecture in the 820M is designed to be more efficient in terms of performance per transistor, yet the benchmark results show it losing to the older Fermi chip. The 820M's support for Vulkan 1.2.175, while the 820A has no listed Vulkan version, is a feature difference that stems from the newer architecture. The 820A's lack of Vulkan support could be a limitation for modern applications, but the OpenCL benchmark suggests that the Fermi 2.0 implementation is more effective for this particular workload.

The Verdict

Based on the data, the NVIDIA GeForce 820A is the pick for users prioritizing compute performance. Its 7.1% lead in the Geekbench OpenCL test is the only direct performance comparison available, and it wins decisively. The 820A's 2983 score, which is competitive with much higher-tier GPUs like the GTX 860M and RTX 4060 Ti, makes it a surprisingly potent option for its class. Its lower TDP of 15 W also makes it a more power-efficient choice for portable devices.

The NVIDIA GeForce 820M, despite its superior theoretical specifications, falls short in the available benchmark. Its 2784 score is lower, and it loses to the 820A in the head-to-head test. However, the 820M's advantage in memory capacity (2 GB vs 1 GB) and bandwidth (28.83 GB/s vs 16.02 GB/s) could make it the better choice for workloads that are heavily memory-bound, such as texture-heavy gaming at higher resolutions. Its higher pixel and texture rates also suggest better rasterization potential. For users who need more VRAM and can tolerate a higher TDP of 45 W, the 820M offers these hardware advantages. But for raw compute power as measured by OpenCL, the data clearly favors the 820A. The verdict is that the 820A is the better compute performer, while the 820M is a more well-rounded part with a stronger memory subsystem, but it does not win any benchmark tests in this data set.