AMD Radeon HD 8730M vs NVIDIA GeForce GTX 670MX Comparison

# Head-to-Head Benchmarks

The data shows a clear, if narrow, victory for the NVIDIA GeForce GTX 670MX in the only directly comparable benchmark. In Geekbench OpenCL, the NVIDIA part scores 6,125 against the AMD Radeon HD 8730M's 5,955, a delta of -2.8% from AMD's perspective. That translates to the GTX 670MX finishing roughly 170 points ahead, a margin that is measurable but not transformative.

What is more telling is the broader benchmark context. The AMD Radeon HD 8730M's score of 5,955 places it at the 34th percentile of all GPUs, while the GTX 670MX's OpenCL result of 6,125 sits at the 33rd percentile. The percentile difference is negligible — both cards occupy nearly the same position in the overall performance distribution. However, the NVIDIA card also has a second benchmark result, Geekbench Vulkan, where it scores 5,316. That additional data point drags its average benchmark score down to 5,721, which is actually lower than the AMD card's single-score average of 5,955. This is a crucial nuance: in raw compute benchmarks, the GTX 670MX wins, but its average across multiple tests is lower because the Vulkan result is significantly weaker than its OpenCL result.

The nearest-rival data reinforces the closeness of this matchup. The AMD card's closest rival is the NVIDIA Quadro K620M at 5,957, a delta of 0% — essentially a tie. The AMD Radeon HD 8750M is just 0.3% ahead, and the Intel UHD Graphics 730 is 0.4% behind. For the NVIDIA card, its closest rival is the Intel Iris Pro Graphics P6300 at 5,712 (0.2% behind), with the NVIDIA GeForce GTX 550 Ti just 0.2% ahead. The AMD Radeon HD 8790M is 0.5% behind the GTX 670MX. These delta values are all within a few tenths of a percent, confirming that both cards are clustered in a tight performance band where individual benchmark runs could easily flip the order.

The head-to-head data gives the GTX 670MX one win and the AMD card zero. But the single benchmark tells only part of the story. The GTX 670MX's OpenCL advantage of 2.8% is real, yet the AMD card's consistent single-score profile suggests it may be more predictable across workloads. Meanwhile, the GTX 670MX's Vulkan score of 5,316 is roughly 13% lower than its OpenCL score, indicating significant API-dependent performance variation on the NVIDIA side.

# Where Each One Wins

NVIDIA GeForce GTX 670MX wins in raw OpenCL compute. The head-to-head benchmark is unambiguous: 6,125 vs. 5,955, a 2.8% margin. If the workload is OpenCL-based, the NVIDIA card is the safer bet. This is further supported by its shading unit count — 960 vs. 384 — which suggests a substantially wider execution width. The GTX 670MX also has a higher texture rate (48.08 GTexel/s vs. 16.80 GTexel/s) and pixel rate (12.02 GPixel/s vs. 5.600 GPixel/s), indicating it should handle fill-rate-heavy tasks with greater ease.

AMD Radeon HD 8730M wins on consistency and average score. The AMD card's average benchmark score is 5,955, while the NVIDIA card's average is 5,721 — a 4.1% advantage for AMD when considering all available data. This is because the AMD card's only benchmark is its OpenCL result, whereas the NVIDIA card's Vulkan result (5,316) pulls its average down. If a user cares about performance across multiple API types, the AMD card's profile is more stable. The AMD card also has a higher percentile ranking (34th vs. 33rd), though the difference is marginal.

NVIDIA GeForce GTX 670MX wins in memory bandwidth. The GTX 670MX offers 67.20 GB/s of bandwidth versus the AMD card's 28.80 GB/s. That is a 2.33x advantage, enabled by its 192-bit bus and GDDR5 memory, compared to the AMD card's 128-bit DDR3 interface. For memory-bound workloads, the NVIDIA card should pull further ahead than the 2.8% OpenCL delta suggests.

AMD Radeon HD 8730M wins on transistor density efficiency. The AMD chip packs 950 million transistors into 77 mm², yielding a density of 12.3M / mm². The NVIDIA chip uses 3,540 million transistors across 294 mm², for a density of 12.0M / mm². The AMD design achieves slightly better packing, though both use the same 28 nm TSMC process.

# FAQ

Q: Which GPU has the higher OpenCL benchmark score?

A: The NVIDIA GeForce GTX 670MX scores 6,125 in Geekbench OpenCL, while the AMD Radeon HD 8730M scores 5,955. The NVIDIA card leads by 2.8%.

Q: Does the AMD card have a higher average benchmark score than the NVIDIA card?

A: Yes. The AMD Radeon HD 8730M has an average benchmark score of 5,955, while the NVIDIA GeForce GTX 670MX averages 5,721. This is because the NVIDIA card's Vulkan score of 5,316 is substantially lower than its OpenCL score.

Q: How do these cards compare in terms of overall GPU percentile ranking?

A: The AMD Radeon HD 8730M sits at the 34th percentile of all GPUs, while the NVIDIA GeForce GTX 670MX sits at the 33rd percentile. The difference is negligible — roughly one percentile point.

Q: What is the memory bandwidth difference between the two cards?

A: The NVIDIA GeForce GTX 670MX provides 67.20 GB/s of bandwidth using GDDR5 memory on a 192-bit bus. The AMD Radeon HD 8730M provides 28.80 GB/s using DDR3 memory on a 128-bit bus. The NVIDIA card offers more than double the bandwidth.

Q: Which card has more shading units?

A: The NVIDIA GeForce GTX 670MX has 960 shading units, compared to 384 on the AMD Radeon HD 8730M. The NVIDIA card also has 80 texture mapping units and 24 ROPs, versus 24 TMUs and 8 ROPs on the AMD card.

Q: Do both cards support DirectX 12?

A: Yes, both support DirectX 12, but with different feature levels. The AMD card supports DirectX 12 (11_1), while the NVIDIA card supports DirectX 12 (11_0). Both support OpenGL 4.6 and Vulkan, with the NVIDIA card having a slightly newer Vulkan version (1.2.175 vs. 1.2.170).

# Specification Differences

The two cards differ across nearly every major specification. The AMD Radeon HD 8730M uses a 128-bit memory bus with 2 GB of DDR3 memory, while the NVIDIA GeForce GTX 670MX uses a 192-bit bus with 3 GB of GDDR5 memory. This drives a significant bandwidth gap: 28.80 GB/s for AMD versus 67.20 GB/s for NVIDIA.

Clock speeds also diverge. The AMD card runs at a 650 MHz base clock with a 700 MHz boost, while the NVIDIA card runs at a fixed 601 MHz with no boost. The AMD card's memory runs at 900 MHz (1,800 Mbps effective), whereas the NVIDIA card's memory runs at 700 MHz (2.8 Gbps effective). Despite the lower clock, the NVIDIA card's wider bus and GDDR5 type deliver far more bandwidth.

Compute resources differ by a wide margin. The NVIDIA card has 960 shading units, 80 TMUs, and 24 ROPs, compared to 384 shading units, 24 TMUs, and 8 ROPs on the AMD card. This yields a peak FP32 performance of 1,153.9 GFLOPS for NVIDIA versus 537.6 GFLOPS for AMD — more than double. The pixel rate is 12.02 GPixel/s for NVIDIA versus 5.600 GPixel/s for AMD, and the texture rate is 48.08 GTexel/s versus 16.80 GTexel/s.

The bus interface also differs: the NVIDIA card uses PCIe 3.0 x16, while the AMD card uses PCIe 3.0 x8. Power consumption is specified at 75 W for the NVIDIA card, while the AMD card has no TDP listed in the data. The NVIDIA card has no power connectors required, and its display outputs are listed as "Portable Device Dependent."

# Architecture Differences

The two GPUs come from different architectural generations and design philosophies. The AMD Radeon HD 8730M is built on the Graphics Core Next 1.0 architecture, using the Mars chip. The NVIDIA GeForce GTX 670MX uses the Kepler architecture, built on the GK104 chip. Both are manufactured on TSMC's 28 nm process, but the silicon differs dramatically in scale.

The AMD Mars chip contains 950 million transistors on a 77 mm² die, achieving a transistor density of 12.3M / mm². The NVIDIA GK104 chip contains 3,540 million transistors on a 294 mm² die, with a density of 12.0M / mm². The NVIDIA chip is roughly 3.7x larger in terms of transistor count and 3.8x larger in die area. This explains the substantial differences in shading units, TMUs, and ROPs — the NVIDIA chip simply has more hardware to work with.

The AMD card belongs to the Solar System generation (HD 8700M series), with a predecessor named London and a successor named Gem System. The NVIDIA card belongs to the GeForce 600M generation, with a predecessor of GeForce 500M and a successor of GeForce 700M. The AMD card was released on 2013-03-31, while the NVIDIA card was released earlier on 2012-09-30. Both are marked as end-of-life products.

API support shows minor differences. The AMD card supports DirectX 12 (11_1), while the NVIDIA card supports DirectX 12 (11_0) — a subtle difference in feature level. Both support OpenGL 4.6. The Vulkan versions differ slightly: 1.2.170 for AMD and 1.2.175 for NVIDIA. Neither card features dedicated ray tracing or tensor cores, as those were not part of either architecture.

Memory technology is a key architectural differentiator. The AMD card uses DDR3, which is lower-bandwidth but typically more power-efficient per access. The NVIDIA card uses GDDR5, which provides substantially higher bandwidth at the cost of different power characteristics. The NVIDIA card's 3 GB capacity also exceeds the AMD card's 2 GB, which matters for larger textures or datasets.

# The Verdict

The data points to the NVIDIA GeForce GTX 670MX as the stronger performer in raw compute, with a 2.8% lead in OpenCL and a more than 2x advantage in memory bandwidth, pixel rate, and FP32 throughput. For any workload that stresses fill rate or memory bandwidth, the NVIDIA card is the clear choice. Its 960 shading units versus 384, 80 TMUs versus 24, and 24 ROPs versus 8 give it a structural advantage that no benchmark margin can fully obscure.

However, the AMD Radeon HD 8730M is not without merit. Its average benchmark score of 5,955 exceeds the NVIDIA card's 5,721, primarily because the NVIDIA card's Vulkan performance (5,316) is markedly weaker than its OpenCL performance. If a user's application mix includes Vulkan, the AMD card may deliver more consistent results. The AMD card also achieves a slightly higher GPU percentile ranking (34th vs. 33rd) and does so with a smaller, denser chip that consumes less silicon area.

For buyers prioritizing OpenCL compute, the GTX 670MX is the pick. For buyers who value cross-API consistency or who are constrained by the AMD card's lower power profile (no TDP listed, but the smaller die suggests lower draw), the HD 8730M is defensible. The NVIDIA card's 3 GB GDDR5 memory also gives it a capacity and bandwidth edge that matters for modern workloads. The verdict is not a knockout — it is a split decision that favors NVIDIA in most performance metrics but leaves AMD with a legitimate consistency argument. Choose the GTX 670MX for peak throughput; choose the HD 8730M for predictable average performance.