NVIDIA GeForce 840M vs NVIDIA GeForce GTX 980M Comparison

The NVIDIA GeForce 840M and the NVIDIA GeForce GTX 980M are both end-of-life mobile GPUs from NVIDIA, but they sit in completely different performance tiers. The GTX 980M is the clear winner in the available head-to-head data, though the aggregate benchmark scores from the database tell a more nuanced story about how these parts are positioned relative to their peers.

Head-to-Head Benchmarks

The two shared benchmarks available for direct comparison are both compute-oriented, and the GTX 980M dominates in each. In the Geekbench OpenCL test, the GTX 980M scores 23,832 points against the 840M's 5,764 points. That is a 75.8% advantage for the GTX 980M, a gap so wide that the 840M is effectively in a different class for GPU compute workloads. The Vulkan test shows a similar pattern: the GTX 980M scores 17,703 versus the 840M's 4,880, a 72.4% lead. These deltas are enormous, and they reflect the fundamental hardware disparity between the two chips.

However, the database's average benchmark scores present a strange anomaly. The 840M has an average benchmark score of 5,322, while the GTX 980M averages 5,308. That places the GTX 980M 0.3% behind the 840M in the aggregate metric, and the 840M is listed as a "nearest rival" to the GTX 980M with a deltaPct of -0.3 from the GTX 980M's perspective. This is almost certainly a quirk of how the average is computed across different test suites — the GTX 980M has ten recorded benchmarks including several Passmark tests, while the 840M has only two Geekbench tests. The GTX 980M's Passmark scores range from 31 in DirectX 12 to 7,338 in G3D, and its Passmark GPU Compute score is 2,816. When you average all of those together with the high Geekbench scores, the number gets dragged down relative to the 840M's clean two-test average. The direct head-to-head results are unambiguous: the GTX 980M wins both tests by over 70%.

The percentile rankings also align: both GPUs sit at the 31st percentile versus all GPUs. That means despite the massive compute advantage of the GTX 980M, both parts are in the lower third of the entire GPU landscape. The nearest rivals list for both cards includes the same set of parts — the GeForce 930A, the GeForce 940M, and the AMD Radeon R7 M445 — with deltaPct values within ±0.9% of each card's average score. The GTX 980M trails the AMD Radeon R7 M445 by 0.9%, while the 840M trails the same AMD part by 0.7%. So in the aggregate database metric, these two NVIDIA parts are effectively indistinguishable from each other and from their low-end peers.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce 840M has a slightly higher average benchmark score of 5,322, versus 5,308 for the GTX 980M. This is a 0.3% difference, with the 840M ahead. However, this aggregate number is misleading because the two cards have very different benchmark portfolios — the 840M only has two Geekbench results, while the GTX 980M has ten results including multiple Passmark tests.

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

A: The GTX 980M scores 23,832 in Geekbench OpenCL, which is 75.8% higher than the 840M's 5,764. The GTX 980M wins this test outright.

Q: What is the performance gap in Vulkan?

A: The GTX 980M scores 17,703 in Geekbench Vulkan, a 72.4% advantage over the 840M's 4,880. Again, the GTX 980M wins decisively.

Q: Do both GPUs support DirectX 12?

A: Yes, both support DirectX 12, but at different feature levels. The 840M supports DirectX 12 (11_0), while the GTX 980M supports DirectX 12 (12_1). Both also support OpenGL 4.6 and Vulkan 1.4.

Q: Which GPU has more shading units?

A: The GTX 980M has 1,536 shading units, which is four times the 384 shading units found on the 840M. The GTX 980M also has 96 texture mapping units versus 16, and 64 ROPs versus 8.

Q: What is the difference in memory bandwidth?

A: The GTX 980M has 160.4 GB/s of memory bandwidth, which is exactly ten times the 16.02 GB/s of the 840M. The GTX 980M uses 8 GB of GDDR5 memory on a 256-bit bus, while the 840M uses 2 GB of DDR3 on a 64-bit bus.

Architecture Differences

Both GPUs are built on the same 28 nm process node at TSMC, but they are very different chips. The 840M uses the GM108S chip with a die size of 77 mm², while the GTX 980M uses the GM204 chip with a die size of 398 mm². The transistor counts reflect this: the 840M packs 1,020 million transistors, and the GTX 980M has 5,200 million. Interestingly, the transistor density is nearly identical — 13.2M per mm² for the 840M and 13.1M per mm² for the GTX 980M — which means the GTX 980M's advantage comes from raw die area, not denser packing.

The architectural generation is also different. The 840M is based on the original Maxwell architecture, while the GTX 980M is Maxwell 2.0. This is reflected in the DirectX feature level: the 840M supports DirectX 12 (11_0), while the GTX 980M supports DirectX 12 (12_1). The GTX 980M's Maxwell 2.0 design delivers higher feature support and better compute throughput per clock.

Neither GPU has any ray tracing cores or tensor cores. Both are pure rasterization and compute parts. The GTX 980M produces 72.13 GPixel/s and 108.2 GTexel/s, while the 840M produces 8.992 GPixel/s and 17.98 GTexel/s. The FP32 compute rating is 3.462 TFLOPS for the GTX 980M versus 863.2 GFLOPS for the 840M — a four-fold difference in raw floating-point throughput. Neither card lists an FP16 rating.

The memory subsystems are entirely different. The 840M uses 2 GB of DDR3 with a 64-bit bus, yielding 16.02 GB/s of bandwidth. The GTX 980M uses 8 GB of GDDR5 with a 256-bit bus, yielding 160.4 GB/s. The memory clock is also higher on the GTX 980M at 1,253 MHz (5 Gbps effective) versus 1,001 MHz (2 Gbps effective) on the 840M.

Specification Differences

The two GPUs differ on nearly every specification that matters. The GTX 980M has 1,536 shading units, 96 TMUs, and 64 ROPs, while the 840M has 384 shading units, 16 TMUs, and 8 ROPs. The GTX 980M has exactly four times the shading units and TMUs, and eight times the ROPs.

The clock speeds are surprisingly close. The 840M has a base clock of 1,029 MHz and a boost clock of 1,124 MHz. The GTX 980M has a base clock of 1,038 MHz and a boost clock of 1,127 MHz — a difference of less than 1% in both cases. The performance gap comes entirely from the much larger execution resources and the faster memory.

The memory configuration is starkly different: 2 GB DDR3 on a 64-bit bus for the 840M versus 8 GB GDDR5 on a 256-bit bus for the GTX 980M. Bandwidth is 16.02 GB/s versus 160.4 GB/s, a ten-fold difference. The GTX 980M also has a higher effective memory speed of 5 Gbps versus 2 Gbps.

The TDP is listed as 33 W for the 840M, while the GTX 980M has no TDP listed in the data. The 840M is an IGP slot width, while the GTX 980M is an MXM Module. The 840M uses a PCIe 3.0 x8 bus interface, while the GTX 980M uses MXM-B (3.0). Neither card requires external power connectors.

The release dates differ by about seven months: the 840M came out on 2014-03-11, and the GTX 980M on 2014-10-06. The 840M's predecessor is the GeForce 700M and its successor is the GeForce 900M, while the GTX 980M's predecessor is the GeForce 800M and its successor is the GeForce 10 Mobile.

Where Each One Wins

The GTX 980M wins every direct benchmark comparison available. In Geekbench OpenCL, it leads by 75.8%. In Geekbench Vulkan, it leads by 72.4%. The GTX 980M also has a much broader benchmark portfolio, with ten recorded results covering DirectX 9, 10, 11, and 12, plus G2D, G3D, and GPU compute tests. Its Passmark G3D score of 7,338 shows strong rasterization performance for its era.

The 840M does not win any of the two head-to-head tests. Its only advantage in the data is the slightly higher average benchmark score of 5,322 versus 5,308, which is a statistical artifact rather than a real performance edge. The 840M does have one practical advantage: a 33 W TDP, which makes it suitable for thin-and-light laptops with integrated-class cooling. The GTX 980M has no listed TDP, but its MXM Module form factor and larger die strongly suggest it requires a much more robust cooling solution.

For use cases, the GTX 980M is the obvious choice for any workload that stresses the GPU: gaming, 3D rendering, GPU compute, or anything that benefits from the 160.4 GB/s of memory bandwidth and 3.462 TFLOPS of FP32 throughput. The 840M is limited to light duties — 2 GB of DDR3 on a 64-bit bus will bottleneck any serious workload. The 840M's 16.02 GB/s of bandwidth is a tenth of the GTX 980M's, and its 8 ROPs will choke on high-resolution output.

The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 980M is the superior GPU by every meaningful performance metric. It wins both head-to-head tests by over 70%, has four times the shading units, eight times the ROPs, and ten times the memory bandwidth. The aggregate average benchmark score slightly favors the 840M, but that is a function of the different test suites recorded for each card, not a reflection of real-world capability.

For a buyer choosing between these two parts, the GTX 980M is the only option for gaming or compute work. Its 8 GB of GDDR5 memory and 256-bit bus provide the headroom that modern games and applications demand. The 840M, with its 2 GB DDR3 and 64-bit bus, is suitable only for basic display output and light GPU-accelerated tasks. The 33 W TDP of the 840M makes it a low-power part for ultraportables, but the GTX 980M's performance advantage is so large that the power efficiency comparison is moot for any serious use. Both are end-of-life parts, and neither ranks above the 31st percentile of all GPUs, but within this pairing, the GTX 980M is the clear choice for anyone who needs actual graphics performance.