AMD Radeon HD 8730M vs AMD Radeon Vega 10 Mobile Comparison

The Radeon Vega 10 Mobile and the Radeon HD 8730M represent two very different eras of AMD mobile graphics: one is an integrated GPU built into a Raven Ridge-M processor, the other a dedicated Mars-based part from the Solar System HD 8700M family. Both are end-of-life products now, but the database shows a clear performance gap between them, with the Vega 10 leading by 8.7% in the only shared benchmark, Geekbench OpenCL. What makes the comparison interesting is how differently each GPU reaches its result: a modern integrated design on a dense 14 nm node versus an older dedicated chip on a mature 28 nm process with its own memory. The data invites a closer look at whether the dedicated card's independent memory is enough to offset the integrated part's newer architecture.

FAQ

Q: Which GPU is faster in benchmarks?

A: The Radeon Vega 10 Mobile wins the only head-to-head test in the database, Geekbench OpenCL, with a score of 6476 versus 5955 for the HD 8730M, an 8.7% advantage for the integrated GPU.

Q: How do the two compare against the broader GPU landscape?

A: The Vega 10 Mobile sits at the 37th percentile of all GPUs in the database, while the HD 8730M sits at the 34th percentile. Both are therefore below the midpoint, but the Vega 10 is the stronger of the two in relative standing.

Q: Which rivals score closest to each GPU?

A: The Vega 10 Mobile's nearest rivals include the NVIDIA Quadro M5000M at 6481, the GeForce GT 555M at 6493, the GeForce GTX 670M at 6513, and the RTX PRO 5000 72 GB Blackwell at 6407. The HD 8730M's closest competitors are the Quadro K620M at 5957, the Radeon HD 8750M at 5970, the Intel UHD Graphics 730 at 5929, and the Quadro K4000 at 5982.

Q: Do the two GPUs differ in memory configuration?

A: Yes, significantly. The Vega 10 Mobile uses system-shared memory with system-dependent bandwidth, since it is an integrated GPU. The HD 8730M has its own 2 GB of DDR3 on a 128-bit bus, delivering 28.80 GB/s of bandwidth.

Q: Which GPU supports newer graphics APIs?

A: The Vega 10 Mobile. It supports DirectX 12 at feature level 12_1 and Vulkan 1.3, while the HD 8730M supports DirectX 12 at feature level 11_1 and Vulkan 1.2.170. Both support OpenGL 4.6.

Q: How do their raw compute figures compare?

A: The Vega 10 Mobile delivers 1.665 TFLOPS of FP32 compute and 3.331 TFLOPS of FP16 at a 2:1 ratio. The HD 8730M manages 537.6 GFLOPS of FP32, and the database records no FP16 figure for it.

Architecture Differences

These two GPUs are separated by two full Graphics Core Next generations. The Vega 10 Mobile is a GCN 5.0 design, built on the Raven-M chip as part of the Vega IGP family for Raven Ridge-M processors. The HD 8730M is a GCN 1.0 part, the "Mars" chip from the Solar System generation that the HD 8700M series belonged to. GCN 1.0 was AMD's first iteration of the architecture; GCN 5.0 arrived several generations later with refinements that show up clearly in the recorded throughput figures.

The manufacturing story is just as stark. The Vega 10 Mobile was fabricated by GlobalFoundries on a 14 nm process, packing 4,940 million transistors into a 210 mm² die, a density of 23.5M transistors per mm². The HD 8730M was built by TSMC on a 28 nm process with only 950 million transistors across 77 mm², or 12.3M per mm². That is roughly half the transistor density, and the total transistor count differs by more than a factor of five. The catch is that the Vega's 210 mm² die is the full processor die, not graphics silicon alone, so the comparison reflects platform design philosophy more than pure GPU size. Still, the density gap between 14 nm and 28 nm is unmistakable.

Feature support also diverges. The Vega 10 Mobile supports DirectX 12 at feature level 12_1 and Vulkan 1.3, while the HD 8730M stops at feature level 11_1 for DirectX 12 and Vulkan 1.2.170. Both report OpenGL 4.6. Neither GPU has RT cores or tensor cores, so ray tracing and dedicated AI acceleration are absent from both. The Vega 10 Mobile also exposes a native FP16 path at 3.331 TFLOPS with a 2:1 ratio, a capability the database does not record for the older chip.

Lineage context rounds out the picture. The HD 8730M's predecessor was "London" and its successor was the "Gem System". The Vega 10 Mobile succeeded the GCN 3.0 IGP and was itself succeeded by the Navi II IGP.

The Verdict

The recorded data points in one direction: the Vega 10 Mobile is the faster GPU. It wins the only shared benchmark by 8.7%, sits three percentile points higher against the full database, and holds commanding leads in every theoretical throughput metric. Its FP32 compute of 1.665 TFLOPS is more than three times the HD 8730M's 537.6 GFLOPS, its texture rate of 52.04 GTexel/s is more than three times the older chip's 16.80 GTexel/s, and its pixel rate of 10.41 GPixel/s is nearly double the HD 8730M's 5.600 GPixel/s.

The one argument in the HD 8730M's favor is memory autonomy. As a dedicated card on a PCIe 3.0 x8 interface with its own 2 GB of DDR3 and 28.80 GB/s of guaranteed bandwidth, it does not compete with the CPU for memory. The Vega 10 Mobile, as an IGP with system-shared memory and system-dependent bandwidth, lives or dies by the host system's RAM configuration. Notably, the benchmark gap of 8.7% is far smaller than the theoretical compute gap, which raises the question of whether the integrated design is memory-bottlenecked in practice. Even so, on the evidence recorded here, the Vega 10 Mobile is the pick for performance and modern API support; the HD 8730M's only data-backed advantage is its independent dedicated memory.

Specification Differences

• Process node: 14 nm (GlobalFoundries) versus 28 nm (TSMC)
• Transistors: 4,940 million versus 950 million
• Die size: 210 mm² versus 77 mm²
• Transistor density: 23.5M/mm² versus 12.3M/mm²
• Shading units: 640 versus 384
• TMUs: 40 versus 24
• ROPs: 8 on both, no difference
• Base clock: 300 MHz versus 650 MHz
• Boost clock: 1301 MHz versus 700 MHz
• Memory: system-shared versus 2 GB DDR3 on a 128-bit bus with 28.80 GB/s bandwidth
• Bus interface: IGP versus PCIe 3.0 x8
• FP32: 1.665 TFLOPS versus 537.6 GFLOPS
• FP16: 3.331 TFLOPS (2:1) versus not recorded
• Pixel rate: 10.41 GPixel/s versus 5.600 GPixel/s
• Texture rate: 52.04 GTexel/s versus 16.80 GTexel/s
• DirectX: 12 (12_1) versus 12 (11_1)
• Vulkan: 1.3 versus 1.2.170
• TDP: 10 W for the Vega 10 Mobile, not recorded for the HD 8730M
• Release date: 2019-01-07 versus 2013-03-31

Head-to-Head Benchmarks

The database contains one shared benchmark between these two GPUs, but it is a meaningful cross-platform test of OpenCL compute.

In Geekbench OpenCL, the Vega 10 Mobile scored 6476 against the HD 8730M's 5955, an 8.7% win for the integrated GPU. That single result decides the head-to-head: one win for the Vega 10 Mobile, zero for the HD 8730M.

Context from the rival tables adds perspective. The Vega 10 Mobile's 6476 lands it within a fraction of a percent of the Quadro M5000M (6481, a delta of just -0.1%) and the GeForce GT 555M (6493, -0.3%), and within -0.6% of the GeForce GTX 670M. Curiously, it also sits 1.1% ahead of the RTX PRO 5000 72 GB Blackwell's recorded average of 6407, a reminder that database averages can group very different hardware together. The HD 8730M's 5955 is effectively tied with the Quadro K620M (a delta of 0%), sits -0.3% behind the Radeon HD 8750M, 0.4% ahead of Intel's UHD Graphics 730, and -0.5% behind the Quadro K4000. In other words, both GPUs cluster tightly with their immediate peers, but the Vega 10 Mobile's cluster sits a tier higher.

The theoretical numbers frame why the dedicated card keeps it close despite the architecture gap. The Vega's shading-unit count (640 versus 384) and TMU count (40 versus 24) suggest a much larger raw advantage than 8.7%, yet the HD 8730M's fixed 28.80 GB/s of dedicated bandwidth partially compensates against an IGP whose memory speed is entirely system-dependent. The benchmark result is the synthesis of those opposing forces.

Where Each One Wins

Vega 10 Mobile wins for: raw compute throughput (1.665 TFLOPS FP32 and 3.331 TFLOPS FP16 versus 537.6 GFLOPS), pixel fill rate (10.41 versus 5.600 GPixel/s), texture throughput (52.04 versus 16.80 GTexel/s), modern API support (DirectX 12_1 and Vulkan 1.3), clock headroom (a 1301 MHz boost versus 700 MHz), and recorded benchmark performance (the 8.7% Geekbench OpenCL win and a higher database percentile, 37 versus 34). Anyone prioritizing measured speed and feature-level support should choose this GPU, with the caveat that its memory performance depends entirely on the host system.

HD 8730M wins for: memory independence. It is the only one of the two with its own memory: 2 GB of DDR3 on a 128-bit bus with a guaranteed 28.80 GB/s of bandwidth, connected via PCIe 3.0 x8 rather than sharing system RAM. For scenarios where predictable, dedicated memory bandwidth matters more than peak compute, the data supports this older chip. Its higher base clock of 650 MHz versus the Vega's 300 MHz also reflects a different clocking strategy, though the Vega's boost clock more than erases that difference.

On the overall record, the head-to-head score stands at one win to zero in favor of the Vega 10 Mobile, and that, combined with the percentile gap and the API support difference, makes it the stronger GPU in this database's measurements.