AMD Radeon RX 6700S vs NVIDIA GeForce RTX 3070 Ti Mobile Comparison

The NVIDIA GeForce RTX 3070 Ti Mobile and AMD Radeon RX 6700S are both end-of-life mobile graphics solutions from the same generation, released on the same date in January 2022. The benchmark data shows a consistent performance advantage for the NVIDIA part, which wins all ten head-to-head comparisons, yet the AMD card holds its own in the broader market percentile rankings. The RTX 3070 Ti Mobile sits at the 69th percentile of all GPUs, while the RX 6700S is close behind at the 67th percentile, indicating that despite the NVIDIA card’s higher raw scores, the AMD solution remains a viable competitor in its own performance tier.

FAQ

Q: Which GPU is faster in the 3DMark Steel Nomad DX12 test?

A: The NVIDIA GeForce RTX 3070 Ti Mobile scores 2572, which is 77.3% higher than the AMD Radeon RX 6700S’s 1451. This is the largest performance gap between the two cards in any benchmark.

Q: How does the average benchmark score compare between the two?

A: The NVIDIA card has an average benchmark score of 23518, while the AMD card averages 22154. This puts the RTX 3070 Ti Mobile roughly 6.2% ahead on average, a margin that aligns with its 10-0 win record in individual tests.

Q: In which test is the performance difference the smallest?

A: The DirectX 9 PassMark test shows the closest result, with the NVIDIA card scoring 196 versus 195 for the AMD card. That is only a 0.5% difference, making it effectively a tie in legacy DX9 workloads.

Q: Does the AMD card win any benchmark categories?

A: No. The head-to-head data shows 10 wins for the NVIDIA GeForce RTX 3070 Ti Mobile and 0 wins for the AMD Radeon RX 6700S. The AMD card comes closest in the DirectX 9 test but still trails by a single point.

Q: How do the two GPUs compare to their nearest rivals?

A: The RTX 3070 Ti Mobile is within 0.5% of the RTX 3080 Mobile and 0.9% ahead of the AMD Radeon AI PRO R9700. The RX 6700S is essentially tied with the RX 5700 (0.1% behind) and 1.4% ahead of the GTX 1060 6 GB.

Q: What is the percentile ranking difference between the two cards?

A: The NVIDIA card ranks at the 69th percentile of all GPUs, while the AMD card ranks at the 67th percentile. This 2-percentile gap reflects the modest overall performance difference despite the NVIDIA card’s dominance in the specific tests listed.

Architecture Differences

The two mobile GPUs come from fundamentally different architectures and manufacturing processes. The NVIDIA GeForce RTX 3070 Ti Mobile uses the GA104 chip built on Ampere architecture, fabricated by Samsung on an 8 nm process. The AMD Radeon RX 6700S uses the Navi 23 chip with RDNA 2.0 architecture, built by TSMC on a more advanced 7 nm node. This process advantage allows the AMD chip to pack 46.7 million transistors per square millimeter, slightly denser than the NVIDIA chip’s 44.4 million. However, the NVIDIA GA104 is a much larger and more complex chip overall, with 17,400 million transistors spread across a 392 mm² die, compared to 11,060 million transistors on the AMD’s 237 mm² die.

The compute resources differ dramatically. The NVIDIA card features 5888 shading units, 184 texture mapping units, and 96 ROPs, alongside 46 dedicated ray tracing cores and 184 tensor cores. The AMD card has far fewer shading units at 1792, with 112 TMUs and 64 ROPs, and only 28 ray tracing cores. The AMD card has no tensor cores listed, which is a key architectural difference since NVIDIA’s tensor cores enable dedicated AI acceleration features that AMD lacks in this implementation. Clock speeds tell a different story: the AMD card runs much higher, with a 1700 MHz base and 2000 MHz boost, while the NVIDIA card has a 915 MHz base and 1410 MHz boost. The AMD card also lists a game clock of 1890 MHz, a figure the NVIDIA card does not provide.

Memory configuration also diverges significantly. Both cards have 8 GB of GDDR6 memory and the same 1750 MHz memory clock with 14 Gbps effective speed. However, the NVIDIA card uses a 256-bit memory bus, which yields 448.0 GB/s of bandwidth, exactly double the AMD card’s 224.0 GB/s from its 128-bit bus. The AMD card compensates partially with its higher clocks and RDNA 2.0’s efficiency, but the bandwidth gap is substantial. The bus interface also differs: NVIDIA uses PCIe 4.0 x16 while AMD uses PCIe 4.0 x8, halving the available interconnect bandwidth for the AMD solution.

Head-to-Head Benchmarks

The RTX 3070 Ti Mobile’s most decisive victory comes in the 3DMark Steel Nomad DX12 test, where it scores 2572 versus the RX 6700S’s 1451. That 77.3% advantage is more than double the next-largest gap in the dataset, showing that the NVIDIA card has a particular strength in modern DX12 workloads at high load. The Geekbench OpenCL results follow a similar but less extreme pattern, with the NVIDIA card at 106022 versus 70965, a 49.4% lead that suggests its raw compute throughput is substantially higher.

Vulkan performance also favors NVIDIA, with a 100027 score against 76150, a 31.4% difference. This is notable because Vulkan often scales well with memory bandwidth, and the NVIDIA card’s 448.0 GB/s versus 224.0 GB/s likely contributes to this result. The PassMark suite shows smaller but still consistent advantages. In DirectX 12, the NVIDIA card leads 73 to 58, a 25.9% margin. DirectX 10 shows a 42.5% lead for NVIDIA (124 versus 87), while DirectX 11 has a narrower 14.2% gap (153 versus 134). The DirectX 9 test is the closest, with NVIDIA ahead by just 0.5% (196 versus 195), indicating that legacy API performance is essentially equivalent between the two architectures.

The compute-focused results reinforce NVIDIA’s overall lead. PassMark GPU Compute shows 7504 versus 5990, a 25.3% advantage for the NVIDIA card. The PassMark G3D score has NVIDIA at 17727 against 15066, a 17.7% gap, while the 2D graphics test shows a modest 6% lead for NVIDIA (783 versus 739). Across all ten tests, the pattern is clear: NVIDIA wins every category, with the largest margins in modern APIs and compute workloads, and the smallest in legacy DX9. The average benchmark scores of 23518 for NVIDIA and 22154 for AMD put the overall performance difference at about 6%, which is a meaningful but not overwhelming edge for the RTX 3070 Ti Mobile.

Specification Differences

The two cards differ across nearly every major specification category. The process node and foundry are different: the NVIDIA card uses Samsung’s 8 nm process, while the AMD card uses TSMC’s 7 nm process. Transistor counts are 17,400 million for NVIDIA versus 11,060 million for AMD, with die sizes of 392 mm² and 237 mm² respectively. The transistor density is slightly higher on the AMD chip at 46.7M per mm² versus 44.4M per mm² for NVIDIA.

Clock speeds show the AMD card running significantly faster, with a 1700 MHz base and 2000 MHz boost versus the NVIDIA card’s 915 MHz base and 1410 MHz boost. The AMD card also has a game clock of 1890 MHz, which the NVIDIA card does not list. Memory bandwidth is a major differentiator: NVIDIA has 448.0 GB/s from a 256-bit bus, while AMD has 224.0 GB/s from a 128-bit bus, both using GDDR6 at the same 1750 MHz memory clock.

Compute unit counts differ substantially. NVIDIA has 5888 shading units, 184 TMUs, and 96 ROPs, while AMD has 1792 shading units, 112 TMUs, and 64 ROPs. Ray tracing cores number 46 for NVIDIA versus 28 for AMD. NVIDIA has 184 tensor cores, while AMD has none listed. The pixel rate is 135.4 GPixel/s for NVIDIA versus 128.0 GPixel/s for AMD, and the texture rate is 259.4 GTexel/s versus 224.0 GTexel/s. FP32 performance is 16.60 TFLOPS for NVIDIA versus 7.168 TFLOPS for AMD, though the AMD card achieves 14.34 TFLOPS FP16 with a 2:1 ratio, while NVIDIA does 16.60 TFLOPS FP16 at a 1:1 ratio.

Power consumption differs by 35 watts, with the NVIDIA card rated at 115 W TDP and the AMD card at 80 W TDP. The bus interface is PCIe 4.0 x16 for NVIDIA versus PCIe 4.0 x8 for AMD. Both cards share the same API support, including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, and both are end-of-life with the same release date of January 2022. The AMD card has a slot width listed as "IGP" (integrated graphics processor), while the NVIDIA card has no slot width listed.

The Verdict

The data points strongly toward the NVIDIA GeForce RTX 3070 Ti Mobile for users who prioritize raw performance across modern workloads. It wins all ten head-to-head benchmarks, with particularly large margins in DX12 (77.3% in 3DMark Steel Nomad) and compute (25.3% in PassMark GPU Compute). Its 448.0 GB/s of memory bandwidth and 16.60 TFLOPS FP32 performance provide a substantial foundation for demanding titles and GPU compute tasks. The card also holds a higher average benchmark score of 23518 versus 22154 and a higher percentile ranking of 69 versus 67. For any scenario where maximum frame rates or compute throughput is the primary goal, the RTX 3070 Ti Mobile is the clear choice based on this data.

The AMD Radeon RX 6700S, despite losing every benchmark, should not be dismissed entirely. Its 80 W TDP is significantly lower than the NVIDIA card’s 115 W, making it a more power-efficient option for slim laptops where thermal and battery constraints matter. The AMD card also uses a more advanced 7 nm process and runs at much higher clocks (2000 MHz boost versus 1410 MHz), which could translate to better efficiency in lighter workloads. Its average benchmark score of 22154 is only 6% behind the NVIDIA card, and it ranks just 2 percentiles lower, suggesting that in real-world usage in less demanding games, the performance gap may be less noticeable than the extreme differences in synthetic tests suggest.

For legacy DirectX 9 workloads, the two cards are effectively tied, with the NVIDIA card ahead by just 0.5%. The AMD card also offers a 2:1 FP16 ratio (14.34 TFLOPS) that could benefit certain compute workloads that favor half-precision, though the NVIDIA card still has higher absolute FP16 throughput at 16.60 TFLOPS. Ultimately, the choice depends on whether the user values the NVIDIA card’s outright performance lead or the AMD card’s lower power draw and process efficiency. The benchmark data favors NVIDIA, but the power consumption figures give AMD a compelling argument for battery-conscious portable systems.