AMD Radeon 610M vs AMD Radeon HD 8790M Comparison

The AMD Radeon HD 8790M and AMD Radeon 610M represent two very different approaches to mobile graphics, separated by nearly a decade of architectural evolution. The data shows a tight contest in synthetic benchmarks, with the older discrete GPU edging out the modern integrated solution by a slim margin overall. The HD 8790M wins both head-to-head tests, but the 610M counters with superior API support, a much smaller power envelope, and a more advanced manufacturing process that hints at a different kind of performance story.

Head-to-Head Benchmarks

The two benchmark results paint a picture of near-parity, but the margins tell a more nuanced story. In Geekbench OpenCL, the HD 8790M scores 5017 against the 610M's 4535, a 10.6% advantage for the older card. This is a substantial lead in raw compute throughput, reflecting the HD 8790M's 384 shading units operating at a 691.2 GFLOPS FP32 rate. The 610M, with only 128 shading units and 486.4 GFLOPS, cannot match this raw parallelism, despite its significantly higher clock speeds of 1500 MHz base and 1900 MHz boost.

The Vulkan test tells a different story entirely. Here, the HD 8790M scores 6365, while the 610M trails by a hair at 6353 — a delta of just 0.2%. This near-tie suggests that the 610M's RDNA 2.0 architecture, with its modern instruction set and support for Vulkan 1.4, is far more efficient at translating its lower theoretical compute into real-world graphics API performance. The older GCN 1.0 chip, despite its raw FLOP advantage, cannot stretch its lead in a modern API context.

Contextualizing these scores against their nearest rivals is instructive. The HD 8790M's average benchmark score of 5691 places it just 0.4% behind the Intel Iris Pro Graphics P6300 and 0.5% behind the NVIDIA GeForce GTX 670MX. It is 1.6% ahead of the NVIDIA Quadro M500M. The 610M's average of 5444 puts it 0.4% behind the NVIDIA Quadro M4000, 0.7% behind the AMD Radeon R7 M440, and 1% behind the NVIDIA GeForce GTX 765M, while running 0.5% ahead of the AMD Radeon R7 M365X. Neither GPU dominates its peer group; both sit in the low-30s percentile ranking across all GPUs.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon HD 8790M, with an average benchmark score of 5691, outperforms the AMD Radeon 610M's 5444. This represents a difference of roughly 4.5% in favor of the older discrete card.

Q: Does the newer Radeon 610M offer any performance advantage in specific tests?

A: No, the data shows the 610M loses both head-to-head benchmarks. It trails by 10.6% in Geekbench OpenCL and by a marginal 0.2% in Geekbench Vulkan, despite its more modern architecture.

Q: What is the most significant architectural difference between the two?

A: The HD 8790M uses a 28 nm GCN 1.0 architecture with 384 shading units, while the 610M uses a 6 nm RDNA 2.0 design with 128 shading units. The 610M also features 2 ray tracing cores, which the HD 8790M lacks entirely.

Q: How do their memory systems compare?

A: The HD 8790M has a dedicated 2 GB GDDR5 memory pool on a 128-bit bus, delivering 64.00 GB/s of bandwidth. The 610M uses System Shared memory, with bandwidth described as System Dependent, meaning its performance relies entirely on the host system's RAM.

Q: Which GPU supports more advanced DirectX features?

A: The Radeon 610M supports DirectX 12 Ultimate (12_2), while the HD 8790M only reaches DirectX 12 (11_1). The 610M also has a newer Vulkan version, 1.4 compared to 1.2.170.

Q: What are the physical and power characteristics of each card?

A: The HD 8790M is an MXM Module with no power connectors and no listed TDP, while the 610M is an IGP with a 15 W TDP and also no power connectors. The 610M's process node is 6 nm versus 28 nm for the HD 8790M.

Architecture Differences

The architectural gap between these two GPUs is vast, despite their similar benchmark outcomes. The HD 8790M, built on the 28 nm process at TSMC, uses the GCN 1.0 architecture with a die size of 77 mm² and 950 million transistors. Its design philosophy is brute force: 384 shading units, 24 texture mapping units, and 8 ROPs. This configuration yields a pixel rate of 7.200 GPixel/s and a texture rate of 21.60 GTexel/s.

The 610M, in contrast, is a product of the 6 nm era, fabricated at TSMC with a die size of 100 mm². Its RDNA 2.0 architecture is fundamentally different, emphasizing efficiency and modern feature sets over raw shader count. It has only 128 shading units, 8 TMUs, and 4 ROPs, but it compensates with much higher clocks — 1500 MHz base and 1900 MHz boost, versus the HD 8790M's 850 MHz base and 900 MHz boost. This results in a lower FP32 rate of 486.4 GFLOPS, but the chip supports FP16 at 972.8 GFLOPS (2:1), a feature absent from the older card.

The 610M also includes 2 ray tracing cores, a hardware feature entirely missing from the HD 8790M. Its memory interface is radically different: the HD 8790M uses a dedicated 128-bit GDDR5 bus, while the 610M relies on System Shared memory with a System Dependent bandwidth. The 610M's bus interface is PCIe 4.0 x8, whereas the HD 8790M uses MXM-A (3.0). Both support DirectX and OpenGL 4.6, but the 610M's Vulkan 1.4 and DirectX 12 Ultimate (12_2) support far exceeds the HD 8790M's Vulkan 1.2.170 and DirectX 12 (11_1).

The Verdict

The benchmark data presents a clear trade-off. For users prioritizing raw compute performance in legacy DirectX workloads, the HD 8790M is the stronger choice, as evidenced by its 10.6% lead in Geekbench OpenCL. Its dedicated GDDR5 memory and 64.00 GB/s bandwidth provide a consistent, predictable memory subsystem that does not depend on the host system.

However, the 610M is not far behind in modern API tests, with a 0.2% deficit in Vulkan that is effectively a statistical tie. Its 15 W TDP, versus the unspecified but likely higher power draw of the MXM module, makes it a vastly more efficient solution for thin-and-light laptops. The 610M's support for DirectX 12 Ultimate and ray tracing cores means it is future-proof for software that leverages these features, even if current benchmarks do not show a decisive win. The HD 8790M's end-of-life status and 2013 release date suggest limited driver optimization for contemporary titles, while the 610M, released in 2022, aligns with modern software ecosystems.

Specification Differences

The two GPUs diverge on nearly every technical specification. The HD 8790M uses a 28 nm process, while the 610M uses 6 nm. The HD 8790M has a 77 mm² die with 950 million transistors, versus the 610M's 100 mm² die with undisclosed transistor count. Clock speeds differ significantly: the HD 8790M runs at 850 MHz base and 900 MHz boost, while the 610M runs at 1500 MHz base and 1900 MHz boost. The HD 8790M has 384 shading units, 24 TMUs, and 8 ROPs, while the 610M has 128 shading units, 8 TMUs, and 4 ROPs. The HD 8790M features 2 GB of GDDR5 memory on a 128-bit bus with 64.00 GB/s bandwidth; the 610M uses System Shared memory with System Dependent bandwidth. The 610M has 2 ray tracing cores; the HD 8790M has none. The HD 8790M's pixel rate is 7.200 GPixel/s and texture rate is 21.60 GTexel/s, while the 610M's are 7.600 GPixel/s and 15.20 GTexel/s, respectively. The HD 8790M's FP32 is 691.2 GFLOPS, while the 610M's is 486.4 GFLOPS with an FP16 rating of 972.8 GFLOPS. The 610M has a 15 W TDP; the HD 8790M has none listed. The HD 8790M is an MXM Module with MXM-A (3.0) interface, while the 610M is an IGP with PCIe 4.0 x8. The 610M supports DirectX 12 Ultimate and Vulkan 1.4, whereas the HD 8790M supports DirectX 12 (11_1) and Vulkan 1.2.170.

Where Each One Wins

The HD 8790M wins in raw compute throughput. Its 691.2 GFLOPS FP32 performance and 64.00 GB/s dedicated memory bandwidth give it a distinct edge in applications that are not optimized for modern APIs, as shown by its 10.6% OpenCL victory. It also wins on texture fill rate, with 21.60 GTexel/s versus the 610M's 15.20 GTexel/s, making it potentially better suited for older games that rely heavily on texture sampling. The 2 GB GDDR5 frame buffer, while small by modern standards, is larger than a shared-memory allocation in a system with limited RAM.

The 610M wins on efficiency and modern feature support. Its 15 W TDP is a fraction of what an MXM discrete GPU would require, making it the only viable choice for ultraportable systems. It wins on API compatibility, supporting DirectX 12 Ultimate and Vulkan 1.4, which are essential for current and upcoming titles. The 2 ray tracing cores provide hardware acceleration for ray-traced effects, a feature the HD 8790M cannot emulate. The 610M's higher clock speeds and 7.600 GPixel/s pixel rate also suggest it can drive modern resolutions more effectively in API-optimized scenarios, despite its lower texture rate. For users who need a GPU for everyday computing, light gaming, and content consumption in a low-power package, the 610M's architecture wins. For those running legacy compute workloads that scale with shader count and memory bandwidth, the HD 8790M's data suggests it remains the superior performer.