AMD Radeon Vega 8 Mobile vs NVIDIA GeForce GTX 670M Comparison

The Verdict

The data presents a clear but narrow hierarchy. The AMD Radeon Vega 8 Mobile holds a decisive lead in the only shared benchmark, scoring 7,435 in Geekbench OpenCL versus the NVIDIA GeForce GTX 670M’s 6,513 — a 14.2% advantage. This is not a marginal victory; it is a substantial margin that places the AMD part in a different performance tier. For anyone choosing between these two end-of-life mobile graphics solutions, the AMD Radeon Vega 8 Mobile is the superior option based purely on compute performance.

However, the verdict is more nuanced than a simple win-loss record. The AMD part’s average benchmark score of 7,203 places it at the 39th percentile of all GPUs, while the NVIDIA part’s 6,513 average lands at the 38th percentile. These percentile rankings are nearly identical, suggesting that while the AMD part wins the head-to-head, both occupy a similar position in the broader GPU landscape. The NVIDIA GeForce GTX 670M’s nearest rival list includes the AMD Radeon Vega 10 Mobile with a score of 6,476, which is within 0.6% of the GTX 670M’s own score — an ironic twist where the higher-tier Vega 10 is essentially tied with the older NVIDIA part. The AMD Vega 8’s nearest rival list shows it trading blows with the NVIDIA GeForce GTX 750 (7,222, a -0.3% delta) and the GTX 560 SE (7,171, a 0.4% delta), indicating the Vega 8 sits in a competitive cluster of desktop-class parts from a previous generation.

For users prioritizing raw compute throughput, the choice is the AMD Radeon Vega 8 Mobile. For those constrained by system architecture — if the platform only supports MXM modules — the GTX 670M remains the only option despite its lower scores. The data does not support any scenario where the NVIDIA part wins on performance; its sole advantage lies in its form factor and dedicated memory subsystem.

Architecture Differences

The architectural gap between these two parts is generational and profound. The AMD Radeon Vega 8 Mobile is built on GCN 5.0 architecture using a 14 nm process at GlobalFoundries, packing 4,940 million transistors into a 210 mm² die. This yields a transistor density of 23.5 million per square millimeter — a figure that reflects a modern, dense design. The NVIDIA GeForce GTX 670M, by contrast, uses the older Fermi 2.0 architecture on TSMC’s 40 nm process, with 1,950 million transistors spread across a much larger 332 mm² die. Its transistor density of 5.9 million per square millimeter is roughly a quarter of the AMD part’s, illustrating how far process technology advanced between their release dates.

The compute configurations differ radically. The Vega 8 Mobile fields 512 shading units, 32 texture mapping units, and 8 ROPs. The GTX 670M counters with 336 shading units, 56 TMUs, and 24 ROPs. This is a fascinating split: the AMD part has 52% more shading units but 43% fewer TMUs and 67% fewer ROPs. The GTX 670M’s higher TMU and ROP counts suggest it was designed for traditional rasterization workloads, where texture fill and pixel output matter most. The Vega 8’s larger shader array points toward compute-heavy tasks that leverage parallel floating-point operations.

Clock behavior also diverges. The Vega 8 Mobile lists a base clock of 300 MHz and a boost clock of 1,101 MHz, with system-shared memory whose bandwidth is listed as "System Dependent." The GTX 670M lists no base or boost clocks for the core, only a memory clock of 750 MHz (3 Gbps effective). This omission is notable — it means the NVIDIA part’s core clock is not documented in the data, while the AMD part’s boost behavior is explicit. The resulting throughput figures reflect this: the Vega 8 achieves 1,127.4 GFLOPS of FP32 performance and 2.255 TFLOPS of FP16 (with a 2:1 ratio), while the GTX 670M delivers 803.7 GFLOPS of FP32 and no listed FP16 capability. The AMD part’s FP16 support is a modern feature that the Fermi-based NVIDIA part lacks entirely.

API support further separates them. The Vega 8 Mobile supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The GTX 670M supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support listed. This means the AMD part is future-proofed for modern Vulkan titles, while the NVIDIA part cannot run Vulkan workloads as specified.

Head-to-Head Benchmarks

The single head-to-head benchmark is Geekbench OpenCL, and it tells a straightforward story. The AMD Radeon Vega 8 Mobile scores 7,435, while the NVIDIA GeForce GTX 670M scores 6,513. The delta is 14.2% in favor of AMD. This is a significant margin in compute performance, indicating that the Vega 8 processes OpenCL workloads substantially faster than the GTX 670M.

Contextualizing this score against their respective rival pools adds depth. The Vega 8’s 7,435 OpenCL result aligns closely with its average benchmark score of 7,203, suggesting consistent performance across tests. The GTX 670M’s 6,513 OpenCL score matches its average exactly, as it has only one benchmark result. The Vega 8’s nearest rival, the NVIDIA GeForce GTX 750, averages 7,222 — just 0.3% below the Vega 8’s average. This implies the Vega 8 performs at the level of a desktop GTX 750, which is a remarkable achievement for an integrated part. The GTX 670M’s nearest rival, the GT 555M, averages 6,493, a 0.3% delta. The GTX 670M is therefore only marginally faster than its immediate predecessor in the mobile lineup, whereas the Vega 8 competes with desktop parts from NVIDIA’s 700-series.

The benchmark results also reveal a discrepancy in the rival lists. The Vega 8’s rivals include the NVIDIA GeForce GTX 970 with an average score of 7,157 and a 0.6% delta — a part that is nominally a high-end desktop GPU. This suggests the Vega 8’s performance cluster is broad, spanning from mid-range desktop to older high-end parts. The GTX 670M’s rival list includes the Intel UHD Graphics P750 at 6,554, which is 0.6% ahead of the GTX 670M, indicating that even integrated Intel graphics have caught up to this Fermi-era part.

Specification Differences

The two parts differ on nearly every specification field. The AMD Radeon Vega 8 Mobile uses a Raven-M chip on a 14 nm process at GlobalFoundries, while the NVIDIA GeForce GTX 670M uses a GF114 chip on a 40 nm process at TSMC. Transistor counts are 4,940 million versus 1,950 million, and die sizes are 210 mm² versus 332 mm². The AMD part’s transistor density of 23.5M/mm² dwarfs the NVIDIA part’s 5.9M/mm².

Memory configurations are fundamentally different. The Vega 8 Mobile uses system-shared memory with no dedicated size, type, bus width, or bandwidth figures — all listed as "System Shared" or "System Dependent." The GTX 670M has 1,536 MB of dedicated GDDR5 memory on a 192-bit bus, with 72.00 GB/s of bandwidth. This dedicated memory gives the GTX 670M a predictable memory performance profile, whereas the Vega 8’s memory performance depends entirely on the host system’s RAM.

Core counts differ: 512 shading units, 32 TMUs, and 8 ROPs for the AMD part versus 336 shading units, 56 TMUs, and 24 ROPs for the NVIDIA part. Pixel rates are close — 8.808 GPixel/s for the Vega 8 versus 8.372 GPixel/s for the GTX 670M — as are texture rates at 35.23 GTexel/s versus 33.49 GTexel/s. The FP32 throughput is 1,127.4 GFLOPS for AMD versus 803.7 GFLOPS for NVIDIA. The Vega 8 also lists FP16 at 2.255 TFLOPS, a feature entirely absent from the GTX 670M.

Power and form factor differ substantially. The Vega 8 Mobile is rated at 25 W TDP and is an IGP with no power connectors and an IGP bus interface. The GTX 670M is rated at 75 W TDP, uses an MXM Module slot width, has no power connectors, and uses an MXM-B (3.0) bus interface. Release dates are seven years apart: January 7, 2019 for the AMD part versus March 21, 2012 for the NVIDIA part. The AMD part’s predecessor is "GCN 3.0 IGP" and successor is "Navi II IGP," while the NVIDIA part’s predecessor is "GeForce 500M" and successor is "GeForce 700M."

FAQ

Q: Which GPU has higher raw compute performance?

A: The AMD Radeon Vega 8 Mobile delivers 1,127.4 GFLOPS of FP32 performance, compared to the NVIDIA GeForce GTX 670M’s 803.7 GFLOPS. This is a 40% advantage for the AMD part in theoretical peak throughput.

Q: Do both GPUs support the same APIs?

A: No. The AMD Radeon Vega 8 Mobile supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The NVIDIA GeForce GTX 670M supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support listed.

Q: How much memory does each GPU have?

A: The NVIDIA GeForce GTX 670M has 1,536 MB of dedicated GDDR5 memory on a 192-bit bus with 72.00 GB/s bandwidth. The AMD Radeon Vega 8 Mobile uses system-shared memory, with its bandwidth listed as "System Dependent."

Q: Which GPU is more power-efficient?

A: The AMD Radeon Vega 8 Mobile has a 25 W TDP, while the NVIDIA GeForce GTX 670M has a 75 W TDP. The AMD part delivers higher FP32 performance at one-third the power draw.

Q: What do the percentile rankings indicate?

A: The AMD Radeon Vega 8 Mobile sits at the 39th percentile of all GPUs, while the NVIDIA GeForce GTX 670M sits at the 38th percentile. Despite the AMD part’s 14.2% benchmark win, both rank nearly identically among all GPUs.

Q: Can the NVIDIA GeForce GTX 670M run Vulkan games?

A: Based on the data, no. The GTX 670M’s API list includes DirectX 12 (11_0) and OpenGL 4.6, but Vulkan is listed as null. The AMD Radeon Vega 8 Mobile explicitly supports Vulkan 1.3.

Where Each One Wins

The AMD Radeon Vega 8 Mobile wins in every measured benchmark category. It is 14.2% faster in Geekbench OpenCL, delivers 40% higher FP32 throughput, and adds FP16 support at 2.255 TFLOPS. Its 25 W TDP makes it suitable for ultrathin laptops where power efficiency is paramount. The Vega 8’s Vulkan 1.3 support means it can handle modern games and compute applications that require this API, while the GTX 670M cannot. The Vega 8’s modern 14 nm process and GCN 5.0 architecture suggest better architectural efficiency, with a transistor density of 23.5M/mm² versus the GTX 670M’s 5.9M/mm².

The NVIDIA GeForce GTX 670M wins in memory architecture. Its dedicated 1,536 MB of GDDR5 memory on a 192-bit bus provides 72.00 GB/s of fixed bandwidth, which is predictable and independent of system RAM. The Vega 8’s system-shared memory is listed as "System Dependent," meaning its performance varies with the host laptop’s memory configuration. For workloads that are sensitive to memory latency or bandwidth consistency, the GTX 670M offers a more stable platform. The GTX 670M also has a higher ROP count (24 versus 8) and TMU count (56 versus 32), which may benefit traditional rasterization workloads that depend on pixel fill and texture filtering. Its MXM Module form factor means it can be replaced or upgraded in compatible laptops, whereas the Vega 8 is soldered as an IGP.

The use-case split is clear: the AMD Radeon Vega 8 Mobile is the winner for compute-heavy tasks, modern API support, and power-constrained environments. The NVIDIA GeForce GTX 670M retains a niche for systems that require a dedicated memory pool and a replaceable MXM module, accepting lower compute performance and higher power consumption. The data shows no scenario where the GTX 670M outperforms the Vega 8 in raw benchmarks, but its memory architecture and form factor may dictate its use in specific legacy platforms.