AMD Radeon 840M vs AMD Radeon 890M Comparison

Head-to-Head Benchmarks

The recorded database contains benchmark results for the AMD Radeon 890M, while the AMD Radeon 840M has no individual benchmark scores listed. The 890M's average benchmark score is 9210, placing it at the 45th percentile among all GPUs. Its nearest rivals in the database are tightly clustered: the AMD Radeon Vega 8 averages 9221 (0.1% ahead of the 890M), the NVIDIA GeForce GTX 960 averages 9273 (0.7% ahead), the NVIDIA GeForce GTX 465 averages 9294 (0.9% ahead), and the NVIDIA GeForce GTX 850M averages 9302 (1% ahead). This means the 890M sits within a narrow competitive band, trailing its closest rivals by less than 1% in every case.

Looking at individual workloads, the 890M delivers its strongest result in Geekbench Vulkan with a score of 40808. The Geekbench OpenCL result follows at 37254. In Passmark tests, the 890M scores 8076 in G3D, 4157 in GPU compute, 979 in G2D, 97 in DirectX 9, 73 in DirectX 11, 37 in DirectX 12, and 33 in DirectX 10. The 3DMark Steel Nomad DX12 test yields a score of 590. These figures show a clear pattern: compute-oriented workloads produce substantially higher scores than legacy DirectX rasterization tests, which is consistent with a modern architecture prioritizing general-purpose and modern API performance.

The 840M, by contrast, has no benchmark entries in the database. Its percentile rank of 50 is based on the complete GPU population, but without recorded test scores, direct numerical comparison between the two products is not possible from the available data. The specification sheets indicate a large gap in execution resources: the 840M uses 256 shading units, 16 texture mapping units, and 8 render output units, while the 890M uses 1024 shading units, 64 texture mapping units, and 32 render output units. Those are 4x differences in each category. The 890M also has 16 ray tracing cores versus 4 on the 840M, again a 4x difference.

Where Each One Wins

The 890M is the only one of the two with recorded benchmark data, so any statement about workload superiority must be inferred from its architecture and the scores it posts. The 890M's pixel rate is 92.80 GPixel/s, and its texture rate is 185.6 GTexel/s. The 840M posts 23.20 GPixel/s and 46.40 GTexel/s, respectively. These rates indicate that the 890M is positioned for higher-resolution rendering and texture-heavy scenes, while the 840M is a more modest performer suited to lighter graphical loads.

FP32 throughput reinforces this split. The 890M delivers 5.939 TFLOPS of FP32 compute, while the 840M delivers 1,484.8 GFLOPS. The 890M's FP16 throughput is also 5.939 TFLOPS with a 1:1 ratio to FP32, meaning there is no half-rate penalty for FP16 workloads. The 840M likewise has a 1:1 FP16 ratio at 1,484.8 GFLOPS. For compute tasks such as machine learning inference, image processing, or physics simulation, the 890M has a clear advantage in raw throughput. For basic desktop acceleration, video playback, or lightweight gaming at modest settings, the 840M's smaller configuration remains functional but cannot match the 890M's headroom.

The ray tracing capabilities differ by a factor of four as well. The 890M has 16 ray tracing cores, the 840M has 4. In games or applications that use DirectX 12 Ultimate features, the 890M will handle ray-traced effects with more capacity. The 840M can still run such workloads, given its DirectX 12 Ultimate (12_2) API support, but at lower throughput. Both parts share the same 15 W TDP, which is notable: the 890M delivers four times the shading units, TMUs, ROPs, and RT cores within the same power envelope, indicating that it is a significantly more efficient design when normalized for power.

Architecture Differences

Both GPUs are built on TSMC's 4 nm process node and use the RDNA 3.5 architecture. They belong to the same generation, identified in the database as Navi III IGP (Strix Point Mobile). The 890M uses the Strix Point chip, while the 840M uses the Krackan Point chip. The 890M's die size is 233 mm² with 34,000 million transistors, yielding a transistor density of 145.9M per mm². The 840M's transistor count and die size are listed as unknown in the database.

The architecture is identical in terms of instruction set and feature support. Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both use a 4 nm process from TSMC. The core clock behavior is also identical: both have a base clock of 400 MHz and a boost clock of 2900 MHz. This means the architectural difference is purely one of scale, not of clock speed or process technology. The 890M simply has more execution units operating at the same frequencies.

The memory subsystem is shared between the two in concept: both use system shared memory, with the bus width also system shared and bandwidth listed as system dependent. Neither GPU has dedicated VRAM, so performance will depend heavily on the host system's memory configuration, including channel count, speed, and capacity. The 890M's larger number of compute units will place higher demands on memory bandwidth, so it likely benefits more from a dual-channel or high-speed memory configuration, though the database does not specify such a configuration.

The 840M's smaller configuration, with 256 shading units and 8 ROPs, suggests it is designed for basic integrated graphics duties in thin-and-light laptops, where the Krackan Point chip likely pairs with lower-power system designs. The 890M, with 1024 shading units and 32 ROPs, is a more capable integrated solution intended for demanding integrated graphics workloads, possibly in larger laptops or handheld gaming devices. Both are classified as IGP (integrated graphics processor) with a slot width of "IGP" and no power connectors.

Specification Differences

The two products differ in several specification fields, while sharing others. The shading unit count is the most significant divergence: 256 on the 840M versus 1024 on the 890M. Texture mapping units are 16 versus 64. Render output units are 8 versus 32. Ray tracing cores are 4 versus 16. The pixel rate is 23.20 GPixel/s on the 840M and 92.80 GPixel/s on the 890M. The texture rate is 46.40 GTexel/s versus 185.6 GTexel/s. FP32 performance is 1,484.8 GFLOPS versus 5.939 TFLOPS. FP16 performance mirrors those figures at 1,484.8 GFLOPS and 5.939 TFLOPS, both with 1:1 ratios.

The chip names differ: Krackan Point for the 840M and Strix Point for the 890M. The 890M has documented transistor count (34,000 million), die size (233 mm²), and transistor density (145.9M / mm²), while the 840M lists these as unknown. The release dates differ as well: the 840M was released on 2025-02-28, while the 890M was released on 2024-07-14. Both share the same predecessor, Navi II IGP, and neither has a successor listed. Both are marked as Active in production status.

The TDP is identical at 15 W for both. The bus interface is PCIe 4.0 x8 for both. Display outputs are listed as portable device dependent for both. The API support is identical: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Neither has a launch MSRP in the database. The base clock of 400 MHz and boost clock of 2900 MHz are the same for both. Memory is system shared for both, with bus width and bandwidth also system shared or system dependent, respectively.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The AMD Radeon 890M delivers 5.939 TFLOPS of FP32 compute, while the AMD Radeon 840M delivers 1,484.8 GFLOPS. The 890M is approximately 4 times faster in this metric.

Q: What is the average benchmark score for the Radeon 890M, and how does it compare to its nearest rivals?

A: The 890M's average benchmark score is 9210. It trails the AMD Radeon Vega 8 by 0.1% (score 9221), the NVIDIA GeForce GTX 960 by 0.7% (score 9273), the NVIDIA GeForce GTX 465 by 0.9% (score 9294), and the NVIDIA GeForce GTX 850M by 1% (score 9302).

Q: Do the two GPUs share the same architecture and process node?

A: Yes, both use the RDNA 3.5 architecture and are fabricated on TSMC's 4 nm process node. They belong to the same generation, Navi III IGP (Strix Point Mobile), but use different chips: Krackan Point for the 840M and Strix Point for the 890M.

Q: Are the clock speeds different between the two GPUs?

A: No, both have a base clock of 400 MHz and a boost clock of 2900 MHz. The performance difference comes from the 890M having 4 times the shading units, TMUs, ROPs, and RT cores, not from higher clock speeds.

Q: Which GPU has more ray tracing cores?

A: The AMD Radeon 890M has 16 ray tracing cores, while the AMD Radeon 840M has 4. This is a 4x difference in ray tracing capacity.

Q: What do the Passmark DirectX scores indicate about the 890M's legacy API performance?

A: The 890M scores 97 in DirectX 9, 73 in DirectX 11, 37 in DirectX 12, and 33 in DirectX 10 on Passmark. The DirectX 9 score is the highest among these, while DirectX 10 is the lowest, suggesting the architecture is better optimized for modern APIs like Vulkan (score 40808 on Geekbench) and OpenCL (score 37254) than for older DirectX pipelines.