AMD Radeon Vega 10 Mobile vs NVIDIA GeForce GTX 670M Comparison

The GeForce GTX 670M and the Radeon Vega 10 Mobile occupy nearly the same performance class, but they achieve parity through radically different designs. The data shows a razor-thin margin: the GTX 670M leads the single Geekbench OpenCL benchmark by a 0.6% delta, scoring 6513 against the Vega 10's 6476. This places both firmly in the 37th-38th percentile of all GPUs, indicating that neither part offers class-leading compute power. The verdict is simple: if you are choosing purely on measured compute output, the two are effectively interchangeable, with the NVIDIA card holding a marginal edge.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, and the result is a statistical tie. The NVIDIA GeForce GTX 670M scores 6513, while the AMD Radeon Vega 10 Mobile scores 6476. The deltaPct of 0.6% in favor of the GTX 670M is minuscule—well within typical run-to-run variance for such benchmarks. This 37-point gap is smaller than the difference between the GTX 670M and its own nearest rival, the GeForce GT 555M, which trails by just 0.3% (6493). Similarly, the Vega 10's closest competitor, the Quadro M5000M, is only 0.1% behind it.

What is more telling is the company each card keeps. The GTX 670M's nearest rivals include the Intel UHD Graphics P750, which beats it by 0.6% (6554), and the Quadro M5000M, which trails by 0.5%. The Vega 10's rival list is nearly identical, with the RTX PRO 5000 72 GB Blackwell sitting 1.1% behind it, the largest margin in either list. This clustering of scores between 6407 and 6554 confirms that both the GTX 670M and Vega 10 are firmly anchored in a mid-tier compute bracket, where no single GPU dominates by more than a few percentage points. In terms of pure benchmark wins, the GTX 670M takes the single available contest; the Vega 10 wins none.

Architecture Differences

The architectural gulf between these two is vast, despite their similar benchmark scores. The GTX 670M is built on NVIDIA's Fermi 2.0 architecture, using the GF114 chip fabricated on a 40 nm process at TSMC. This is a mature, power-hungry design from 2012, packing 1,950 million transistors into a 332 mm² die. The transistor density is a modest 5.9 million per square millimeter. In contrast, the Vega 10 Mobile uses AMD's GCN 5.0 architecture with the Raven-M chip, built on a 14 nm process at GlobalFoundries. This newer design crams 4,940 million transistors into a smaller 210 mm² die, achieving a density of 23.5M transistors per mm²—roughly four times the density of the Fermi part.

The compute resources are distributed differently. The GTX 670M has 336 shading units, 56 texture mapping units, and 24 raster operation pipelines. The Vega 10 Mobile counters with 640 shading units, 40 TMUs, but only 8 ROPs. This explains the fillrate numbers: the GTX 670M produces 8.372 GPixel/s and 33.49 GTexel/s, while the Vega 10 outputs 10.41 GPixel/s and 52.04 GTexel/s. The Vega 10 has a clear advantage in both pixel and texture throughput, despite having fewer TMUs and ROPs, because its clock speeds are far higher. The FP32 compute is also lopsided: the Vega 10 delivers 1.665 TFLOPS, more than double the GTX 670M's 803.7 GFLOPS. The Vega 10 also supports FP16 at 3.331 TFLOPS (2:1), a feature the Fermi architecture lacks entirely.

Memory subsystems could not be more different. The GTX 670M uses 1536 MB of dedicated GDDR5 memory on a 192-bit bus, delivering 72.00 GB/s of bandwidth. The Vega 10 Mobile uses System Shared memory, with its bus width and bandwidth listed as System Dependent. This means the Vega 10's memory performance is entirely contingent on the host system's RAM configuration, whereas the GTX 670M has fixed, dedicated resources. The GTX 670M's memory clock is 750 MHz (3 Gbps effective), while the Vega 10's is simply "System Shared."

Where Each One Wins

The GTX 670M wins the only head-to-head benchmark, but its victory is so narrow that it is practically a draw. The more meaningful wins come from architectural attributes. The GTX 670M's dedicated 72.00 GB/s GDDR5 memory is a significant advantage in scenarios where memory bandwidth is a bottleneck, such as high-resolution texture streaming or heavy multi-tasking. Its 192-bit bus and 1536 MB frame buffer provide predictable, consistent performance regardless of system RAM speed. This makes the GTX 670M a better choice for workloads that require stable, dedicated video memory.

The Vega 10 Mobile wins on raw compute throughput and efficiency. Its 1.665 TFLOPS FP32 performance is more than double the GTX 670M's 803.7 GFLOPS, making it superior for compute-heavy tasks like OpenCL workloads that scale with shader count and clock speed. Its 52.04 GTexel/s texture rate and 10.41 GPixel/s pixel rate are also higher, giving it an edge in fillrate-bound applications. The Vega 10 also supports Vulkan 1.3 and DirectX 12 (12_1), while the GTX 670M is limited to DirectX 12 (11_0) with no Vulkan support listed. For modern API utilization, the Vega 10 is clearly ahead.

The power envelope is a decisive factor. The Vega 10 Mobile has a 10 W TDP, while the GTX 670M draws 75 W. This seven-fold difference means the Vega 10 is far better suited for thin-and-light laptops where battery life and thermal headroom are critical. The GTX 670M requires an MXM module slot and has no power connectors listed, but its 75 W draw demands a more robust cooling solution. The Vega 10 is an IGP, integrated directly into the system, which eliminates the need for a separate graphics card slot.

Specification Differences

The two GPUs differ on nearly every specification field. The process node is 40 nm (TSMC) for NVIDIA versus 14 nm (GlobalFoundries) for AMD. Transistor counts are 1,950 million versus 4,940 million, and die sizes are 332 mm² versus 210 mm². The GTX 670M has 336 shading units, 56 TMUs, and 24 ROPs; the Vega 10 has 640 shading units, 40 TMUs, and 8 ROPs. Clock speeds: the GTX 670M has no listed base or boost clock, only a 750 MHz memory clock; the Vega 10 has a 300 MHz base and 1301 MHz boost clock.

Memory is the starkest contrast: the GTX 670M has 1536 MB GDDR5 on a 192-bit bus with 72.00 GB/s bandwidth, while the Vega 10 uses System Shared memory with System Dependent bandwidth. The GTX 670M's pixel rate is 8.372 GPixel/s and texture rate is 33.49 GTexel/s; the Vega 10's are 10.41 GPixel/s and 52.04 GTexel/s. FP32 compute is 803.7 GFLOPS versus 1.665 TFLOPS, and the Vega 10 adds FP16 at 3.331 TFLOPS. TDP is 75 W versus 10 W. The GTX 670M uses an MXM-B (3.0) bus interface, while the Vega 10 is an IGP. API support differs: DirectX 12 (11_0) versus 12 (12_1), with Vulkan 1.3 only on the Vega 10. Release dates are March 2012 for the GTX 670M and January 2019 for the Vega 10.

FAQ

Q: Which GPU has a higher Geekbench OpenCL score?

A: The NVIDIA GeForce GTX 670M scores 6513, which is 0.6% higher than the AMD Radeon Vega 10 Mobile's 6476.

Q: How does the Vega 10 Mobile compare to the GTX 670M in power consumption?

A: The Vega 10 Mobile has a TDP of 10 W, while the GTX 670M has a TDP of 75 W, making the Vega 10 far more power-efficient.

Q: What is the memory configuration difference between the two?

A: The GTX 670M has 1536 MB of dedicated GDDR5 memory on a 192-bit bus with 72.00 GB/s bandwidth. The Vega 10 Mobile uses System Shared memory with System Dependent bandwidth, meaning it relies on the host system's RAM.

Q: Which GPU supports Vulkan?

A: Only the AMD Radeon Vega 10 Mobile supports Vulkan 1.3. The NVIDIA GeForce GTX 670M has no Vulkan support listed, and its DirectX support is limited to 12 (11_0) compared to the Vega 10's 12 (12_1).

Q: How do their compute capabilities compare?

A: The Vega 10 Mobile delivers 1.665 TFLOPS FP32 and 3.331 TFLOPS FP16 (2:1), while the GTX 670M delivers 803.7 GFLOPS FP32 and no FP16 support.

Q: Which GPU has a higher transistor density?

A: The AMD Radeon Vega 10 Mobile has a density of 23.5M transistors per mm², compared to the GTX 670M's 5.9M per mm², due to its newer 14 nm process versus the 40 nm process of the GTX 670M.

The Verdict

The data supports a clear split based on use case. If you prioritize raw compute throughput, modern API support, and energy efficiency, the AMD Radeon Vega 10 Mobile is the superior choice. Its 1.665 TFLOPS FP32 performance, Vulkan 1.3 support, and 10 W TDP make it the better fit for battery-powered devices and workloads that leverage shader-heavy compute. The doubling of FP32 performance over the GTX 670M is not reflected in the Geekbench OpenCL score, but it is a real architectural advantage.

If you require predictable, dedicated memory bandwidth and a larger framebuffer, the NVIDIA GeForce GTX 670M is the safer option. Its 1536 MB GDDR5 on a 192-bit bus provides 72.00 GB/s of guaranteed bandwidth, which the Vega 10 cannot match without a fast system RAM configuration. The GTX 670M also wins the single available benchmark, though by a negligible margin. Its 75 W TDP suggests it was designed for larger laptops with active cooling, not ultra-portables.

In the end, the benchmark scores say these are equivalent performers. The GTX 670M's 0.6% lead in Geekbench OpenCL is meaningless in real-world terms. The choice comes down to platform: the Vega 10 is an IGP for modern, efficient systems; the GTX 670M is a legacy MXM part for older, bulkier machines. Neither card breaks out of the 37th-38th percentile, so expectations should be modest. The Vega 10 is the forward-looking pick for its feature set and efficiency; the GTX 670M is the pick for those needing dedicated VRAM.