AMD Radeon HD 8790M vs NVIDIA GeForce GTX 980M Comparison

The AMD Radeon HD 8790M and NVIDIA GeForce GTX 980M are two mobile GPUs from different eras and entirely different performance tiers. The benchmark data shows a clear and massive performance gulf, with the GTX 980M winning every single head-to-head comparison by a significant margin. This analysis breaks down where each chip stands based strictly on the available data.

Where Each One Wins

The data is unambiguous: the NVIDIA GeForce GTX 980M wins every benchmark category where both are tested. In the two shared tests, the GTX 980M delivers 23832 in Geekbench OpenCL versus 5017 for the HD 8790M, a delta of -78.9% for the AMD part. In Geekbench Vulkan, the GTX 980M scores 17703 against 6365, a -64% gap. There is no test in the data where the AMD Radeon HD 8790M comes out ahead.

The HD 8790M's only claim to relevance is its existence as a lower-tier option. Its average benchmark score is 5691, which places it at the 33rd percentile of all GPUs. The GTX 980M, despite its dominant wins, has a lower average score of 5308 and sits at the 31st percentile. This is a statistical quirk driven by the different benchmark suites available for each card. The GTX 980M's benchmark pool includes demanding tests like Passmark DirectX 12 (score 31) and 3DMark Steel Nomad DX12 (score 649), which drag its average down. The HD 8790M only has two Geekbench scores, both relatively strong for its class.

For practical use-case splits, the GTX 980M is the clear winner for any modern or demanding workload. Its 3DMark Steel Nomad DX12 score of 649 and Passmark G3D score of 7338 indicate it can handle contemporary DirectX 12 titles. The HD 8790M, with no DirectX 12 benchmark scores in the data and a peak Geekbench OpenCL score of 5017, is suited only for older or lighter applications. The GTX 980M also wins decisively in compute-heavy tasks, as shown by its Geekbench OpenCL score being nearly five times higher.

Architecture Differences

The two GPUs come from completely different architectural lineages. The AMD Radeon HD 8790M uses the Mars chip built on GCN 1.0 architecture, part of the Solar System generation (HD 8700M). The NVIDIA GeForce GTX 980M uses the GM204 chip on Maxwell 2.0 architecture, from the GeForce 900M generation.

Both are fabricated on the same 28 nm process at TSMC, but the silicon differs drastically in scale. The GTX 980M's GM204 packs 5,200 million transistors on a 398 mm² die, yielding a transistor density of 13.1M / mm². The HD 8790M's Mars chip is far smaller, with 950 million transistors on 77 mm², giving a density of 12.3M / mm². The GTX 980M has roughly 5.5 times more transistors and a die over five times larger.

The compute resources reflect this disparity. The GTX 980M has 1536 shading units, 96 texture mapping units (TMUs), and 64 raster operations pipelines (ROPs). The HD 8790M has just 384 shading units, 24 TMUs, and 8 ROPs. The GTX 980M offers exactly four times the shading units, four times the TMUs, and eight times the ROPs.

Memory configurations are also wildly different. The GTX 980M features 8 GB of GDDR5 on a 256-bit bus, delivering 160.4 GB/s of bandwidth. The HD 8790M has 2 GB of GDDR5 on a 128-bit bus, with 64.00 GB/s of bandwidth. The GTX 980M has 2.5 times the memory bandwidth and four times the memory capacity.

Clock speeds favor the NVIDIA part as well. The GTX 980M runs at a base clock of 1038 MHz with a boost of 1127 MHz, while the HD 8790M runs at 850 MHz base and 900 MHz boost. Memory clocks differ too: the GTX 980M's memory runs at 1253 MHz (5 Gbps effective), while the HD 8790M's runs at 1000 MHz (4 Gbps effective).

API support shows a generational gap. The GTX 980M supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The HD 8790M supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. Both use an MXM Module slot, but the GTX 980M uses the MXM-B (3.0) interface while the HD 8790M uses the smaller MXM-A (3.0) interface.

The Verdict

The data dictates a simple choice: the NVIDIA GeForce GTX 980M is the superior GPU in every measurable way. Benchmark results show it is 78.9% ahead in Geekbench OpenCL and 64% ahead in Geekbench Vulkan. Its architectural advantages are overwhelming, with four times the shading units and TMUs, eight times the ROPs, four times the memory capacity, and 2.5 times the memory bandwidth.

Anyone choosing between these two for a laptop upgrade or purchase should pick the GTX 980M without hesitation, assuming the system can physically accommodate the MXM-B module. The GTX 980M's higher transistor count, larger die, and newer Maxwell 2.0 architecture make it capable of handling modern gaming and compute tasks that the GCN 1.0-based HD 8790M simply cannot manage.

The only scenario where the HD 8790M makes sense is if the hardware is already installed and the user has very light, legacy workloads. Its average benchmark score of 5691 is slightly higher than the GTX 980M's 5308, but this is misleading. The HD 8790M's score comes from only two Geekbench tests, while the GTX 980M's average includes strenuous modern workloads like Passmark DirectX 12 (score 31) and 3DMark Steel Nomad DX12 (score 649). The GTX 980M is the clear choice for anyone who wants real performance.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon HD 8790M has an average benchmark score of 5691, while the NVIDIA GeForce GTX 980M has an average of 5308. However, the GTX 980M's average is dragged down by its inclusion of demanding modern tests like 3DMark Steel Nomad DX12 and Passmark DirectX 12.

Q: How much faster is the GTX 980M in Geekbench OpenCL?

A: The GTX 980M scores 23832 in Geekbench OpenCL, which is 78.9% higher than the HD 8790M's score of 5017.

Q: Do both GPUs use the same process node?

A: Yes, both the AMD Radeon HD 8790M and the NVIDIA GeForce GTX 980M are fabricated on a 28 nm process at TSMC.

Q: What is the memory bandwidth difference between the two?

A: The GTX 980M offers 160.4 GB/s of memory bandwidth, while the HD 8790M offers 64.00 GB/s. The GTX 980M has 2.5 times the bandwidth.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GTX 980M has 1536 shading units, which is exactly four times the 384 shading units found on the AMD Radeon HD 8790M.

Q: What is the DirectX version support difference?

A: The GTX 980M supports DirectX 12 (12_1), while the HD 8790M supports DirectX 12 (11_1). The GTX 980M's support for the 12_1 feature level is more modern.

Head-to-Head Benchmarks

The two head-to-head benchmark results paint a stark picture. In Geekbench OpenCL, the NVIDIA GeForce GTX 980M scores 23832 against the AMD Radeon HD 8790M's 5017. That is a delta of -78.9%, meaning the HD 8790M performs at roughly one-fifth the level of the GTX 980M in this compute test. This massive gap is consistent with the architectural differences: the GTX 980M has 1536 shading units versus 384, and its FP32 throughput is 3.462 TFLOPS versus 691.2 GFLOPS.

In Geekbench Vulkan, the GTX 980M scores 17703, while the HD 8790M scores 6365. The delta is -64%, still a dominant win for the NVIDIA part. While the Vulkan gap is narrower than the OpenCL gap, the GTX 980M still outperforms the HD 8790M by a factor of nearly three. The HD 8790M's Vulkan score of 6365 is actually higher than its OpenCL score of 5017, suggesting the GCN 1.0 architecture handles Vulkan relatively better than OpenCL. However, this relative strength does not close the absolute gap.

The GTX 980M also has benchmark scores in tests that the HD 8790M does not have data for. It scores 7338 in Passmark G3D, 2816 in Passmark GPU Compute, and 649 in 3DMark Steel Nomad DX12. These scores, while not directly comparable to the HD 8790M, reinforce the GTX 980M's position as the far more capable GPU. The HD 8790M has no scores in any DirectX-specific or modern 3DMark test, further limiting its utility for current gaming.

The wins tally confirms the sweep: the GTX 980M wins 2 out of 2 head-to-head benchmarks, while the HD 8790M wins 0. The data does not support any conclusion other than total NVIDIA dominance in this pairing.

Specification Differences

The specification sheets for these two GPUs differ in nearly every meaningful way. Here are the fields where the two parts diverge:

• Chip: AMD uses "Mars"; NVIDIA uses "GM204".
• Architecture: AMD uses GCN 1.0; NVIDIA uses Maxwell 2.0.
• Generation: AMD is from "Solar System (HD 8700M)"; NVIDIA is from "GeForce 900M".
• Transistors: AMD has 950 million; NVIDIA has 5,200 million.
• Die Size: AMD is 77 mm²; NVIDIA is 398 mm².
• Transistor Density: AMD is 12.3M / mm²; NVIDIA is 13.1M / mm².
• Base Clock: AMD is 850 MHz; NVIDIA is 1038 MHz.
• Boost Clock: AMD is 900 MHz; NVIDIA is 1127 MHz.
• Memory Clock: AMD is 1000 MHz (4 Gbps effective); NVIDIA is 1253 MHz (5 Gbps effective).
• Memory Size: AMD has 2 GB; NVIDIA has 8 GB.
• Memory Bus Width: AMD is 128 bit; NVIDIA is 256 bit.
• Memory Bandwidth: AMD is 64.00 GB/s; NVIDIA is 160.4 GB/s.
• Shading Units: AMD has 384; NVIDIA has 1536.
• TMUs: AMD has 24; NVIDIA has 96.
• ROPs: AMD has 8; NVIDIA has 64.
• Pixel Rate: AMD is 7.200 GPixel/s; NVIDIA is 72.13 GPixel/s.
• Texture Rate: AMD is 21.60 GTexel/s; NVIDIA is 108.2 GTexel/s.
• FP32: AMD is 691.2 GFLOPS; NVIDIA is 3.462 TFLOPS.
• Bus Interface: AMD uses MXM-A (3.0); NVIDIA uses MXM-B (3.0).
• DirectX Support: AMD supports 12 (11_1); NVIDIA supports 12 (12_1).
• Vulkan Support: AMD supports 1.2.170; NVIDIA supports 1.4.
• Release Date: AMD released 2013-03-31; NVIDIA released 2014-10-06.
• Predecessor: AMD's predecessor is "London"; NVIDIA's is "GeForce 800M".
• Successor: AMD's successor is "Gem System"; NVIDIA's is "GeForce 10 Mobile".

Both GPUs share the same process node (28 nm, TSMC), slot width (MXM Module), power connector situation (None), display outputs (Portable Device Dependent), and production status (End-of-life). Neither has a launch MSRP listed in the data.