AMD Radeon R9 M290X vs NVIDIA GeForce RTX 3090 Comparison

The Verdict

The recorded data positions the NVIDIA GeForce RTX 3090 as the dominant performer in this comparison. Its average benchmark score of 27,565 places it in the 73rd percentile of all GPUs, while the AMD Radeon R9 M290X averages 23,276, sitting in the 68th percentile. The single head-to-head benchmark, Geekbench OpenCL, shows the RTX 3090 delivering a score of 172,758 against 22,028 for the R9 M290X, a delta of 684.3%. That is not a marginal gap; it is a generational chasm.

The RTX 3090's nearest rivals in the database are the NVIDIA GeForce RTX 4070 Mobile (average 27,435, delta 0.5%), the AMD Radeon RX 6700 XT (average 27,425, delta 0.5%), the AMD Radeon Pro Vega 20 (average 27,839, delta -1%), and the AMD Radeon RX 7800M (average 27,883, delta -1.1%). This means the RTX 3090 sits in a tightly packed cluster of high-end parts, trading blows within a 1.1% band. The R9 M290X, by contrast, aligns with the AMD Radeon RX 6600M (average 23,273, delta 0%), the AMD Radeon Pro Vega 16 (average 23,250, delta 0.1%), the NVIDIA P106-100 (average 23,249, delta 0.1%), and the AMD Radeon AI PRO R9700 (average 23,315, delta -0.2%). Its position is similarly tight, but in a much lower performance tier.

The data indicates that the RTX 3090 is the clear choice for any workload requiring maximum compute throughput. The R9 M290X, however, was designed for a different era and a different power envelope. The database shows it is an end-of-life part from 2014, built for portable systems, and its performance level is consistent with that role. The verdict is straightforward: the RTX 3090 wins every measurable category, and the R9 M290X is only relevant for legacy systems or scenarios where its 100 W power draw and MXM form factor are absolute requirements.

Architecture Differences

The two GPUs come from entirely different architectural generations. The RTX 3090 uses the GA102 chip on NVIDIA's Ampere architecture, fabricated on Samsung's 8 nm process. It packs 28,300 million transistors into a 628 mm² die, yielding a transistor density of 45.1 million per square millimeter. The R9 M290X uses the Neptune chip on AMD's GCN 1.0 architecture, built on TSMC's 28 nm process. It contains 2,800 million transistors on a 212 mm² die, with a density of 13.2 million per square millimeter.

Core counts diverge sharply. The RTX 3090 features 10,496 shading units, 328 texture mapping units, 112 raster output units, 82 ray tracing cores, and 328 tensor cores. The R9 M290X has 1,280 shading units, 80 TMUs, and 32 ROPs, with no ray tracing or tensor cores listed. Clock speeds also differ: the RTX 3090 runs at 1395 MHz base and 1695 MHz boost, while the R9 M290X runs at 850 MHz base and 900 MHz boost.

Memory architecture reflects the generational leap. The RTX 3090 carries 24 GB of GDDR6X on a 384-bit bus, delivering 936.2 GB/s of bandwidth. The R9 M290X carries 4 GB of GDDR5 on a 256-bit bus, providing 153.6 GB/s. The RTX 3090's memory clock is listed at 1219 MHz with 19.5 Gbps effective data rate; the R9 M290X's memory clock is 1200 MHz with 4.8 Gbps effective.

Compute rates follow the same pattern. The RTX 3090 achieves 35.58 TFLOPS FP32 and 35.58 TFLOPS FP16 (1:1), with a pixel rate of 189.8 GPixel/s and a texture rate of 556.0 GTexel/s. The R9 M290X delivers 2.304 TFLOPS FP32, with a pixel rate of 28.80 GPixel/s and a texture rate of 72.00 GTexel/s. The FP16 figure for the R9 M290X is not recorded.

API support also differs. The RTX 3090 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The R9 M290X supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The RTX 3090 uses a PCIe 4.0 x16 interface, while the R9 M290X uses PCIe 3.0 x16. Power consumption is not comparable: the RTX 3090 is rated at 350 W TDP with a triple-slot design and a single 12-pin connector, while the R9 M290X is rated at 100 W TDP with an MXM module form factor and no external power connectors.

Head-to-Head Benchmarks

The database contains exactly one direct comparison between these two GPUs: Geekbench OpenCL. The RTX 3090 scores 172,758, and the R9 M290X scores 22,028. The RTX 3090 wins this test with a delta of 684.3%. To put that in perspective, the RTX 3090's score is roughly 7.8 times higher than the R9 M290X's score. This is the largest single-test margin recorded in the head-to-head data, and it aligns with the architectural differences: the RTX 3090 has 8.2 times the shading units, 4.1 times the TMUs, 3.5 times the ROPs, and 6.1 times the FP32 throughput.

The RTX 3090 also shows a much broader benchmark footprint. Its recorded tests include 3DMark Steel Nomad DX12 (5,118), Geekbench Vulkan (53,927), Passmark DirectX 10 (182), DirectX 11 (220), DirectX 12 (110), DirectX 9 (268), G2D (1,063), G3D (26,645), and GPU Compute (15,356). The R9 M290X only has two recorded tests: Geekbench Metal (24,524) and Geekbench OpenCL (22,028). This lack of shared tests beyond OpenCL means the head-to-head analysis is limited to that single data point, but the margin is decisive enough to establish the performance hierarchy.

The average benchmark score for the RTX 3090 is 27,565, which is 18.5% higher than the R9 M290X's average of 23,276. However, the averages are computed from different test suites, so the direct OpenCL comparison is the more reliable indicator. In that test, the RTX 3090's lead is overwhelming, and the percentile rankings confirm the gap: the RTX 3090 sits at the 73rd percentile, the R9 M290X at the 68th, a difference of five percentile points despite the massive score delta.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce RTX 3090 has an average benchmark score of 27,565, while the AMD Radeon R9 M290X has an average of 23,276.

Q: How large is the performance gap in the shared Geekbench OpenCL test?

A: The RTX 3090 scores 172,758, and the R9 M290X scores 22,028, giving the RTX 3090 a 684.3% advantage.

Q: Does the R9 M290X support ray tracing or tensor cores?

A: No. The database lists no ray tracing cores and no tensor cores for the R9 M290X, while the RTX 3090 has 82 RT cores and 328 tensor cores.

Q: What are the memory specifications for each card?

A: The RTX 3090 has 24 GB of GDDR6X on a 384-bit bus with 936.2 GB/s bandwidth. The R9 M290X has 4 GB of GDDR5 on a 256-bit bus with 153.6 GB/s bandwidth.

Q: Which card has a lower power draw?

A: The R9 M290X is rated at 100 W TDP, while the RTX 3090 is rated at 350 W TDP. The R9 M290X also uses an MXM Module form factor with no power connectors, whereas the RTX 3090 is triple-slot with a 12-pin connector.

Q: What is the DirectX support for each GPU?

A: The RTX 3090 supports DirectX 12 Ultimate (12_2), while the R9 M290X supports DirectX 12 (11_1).

Where Each One Wins

The RTX 3090 wins the only direct head-to-head benchmark, Geekbench OpenCL, by a factor of 7.8. It also wins every category where comparative data exists: shading units (10,496 vs 1,280), TMUs (328 vs 80), ROPs (112 vs 32), memory capacity (24 GB vs 4 GB), memory bandwidth (936.2 GB/s vs 153.6 GB/s), FP32 throughput (35.58 TFLOPS vs 2.304 TFLOPS), and pixel/texture rates. Its 73rd percentile ranking versus the R9 M290X's 68th percentile reinforces this dominance.

The R9 M290X, however, wins in several non-performance categories. It draws 100 W versus 350 W, making it suitable for systems with strict power limits. Its MXM Module form factor and lack of power connectors indicate it was designed for portable or compact systems, whereas the RTX 3090's triple-slot, 336 mm length, and 61 mm width require a full desktop chassis. The R9 M290X also has a 28 nm process node, which is older but not a disadvantage in its intended niche.

The R9 M290X's Geekbench Metal score of 24,524 is its only other recorded benchmark, and there is no Metal score for the RTX 3090 in the database. This means the R9 M290X has a measurable result in an Apple-oriented API that the RTX 3090 lacks, but this is a data gap rather than a performance advantage. The RTX 3090's Geekbench Vulkan score of 53,927 and its Passmark G3D score of 26,645 further illustrate its breadth across modern APIs.

For users prioritizing raw compute, the RTX 3090 is the only choice. For users constrained by power, physical space, or legacy platform compatibility, the R9 M290X remains a functional part, but its 2014 release date, 28 nm process, and 2.304 TFLOPS FP32 throughput place it firmly in a lower performance tier. The database shows no scenario where the R9 M290X outperforms the RTX 3090 in shared workloads.