NVIDIA GeForce GTX 1070 vs NVIDIA GeForce GTX 1650 SUPER Comparison

The NVIDIA GeForce GTX 1650 SUPER and the NVIDIA GeForce GTX 1070 represent two distinct approaches to GPU design within the same GeForce 10-series family. The GTX 1070, an earlier Pascal-generation card, relies on a larger die and wider memory bus, while the GTX 1650 SUPER uses a newer Turing architecture on a smaller, more efficient process. Benchmark data reveals a clear split: the GTX 1070 dominates traditional rasterization workloads, while the GTX 1650 SUPER asserts a significant lead in a specific Vulkan compute test. The following analysis breaks down exactly where each card wins, the architectural reasons behind those results, and what the numbers say about who should choose which.

Where Each One Wins

The head-to-head benchmark results show a decisive overall victory for the NVIDIA GeForce GTX 1070, which wins 9 out of 10 comparisons. The GTX 1070’s wins span compute, legacy DirectX APIs, and modern 3DMark tests, establishing it as the stronger general-purpose performer. Its largest margins come in the 3DMark Steel Nomad DX12 test, where it scores 1082 against the GTX 1650 SUPER’s 352, a 67.5% advantage. This indicates a substantial edge in sustained, modern DX12 load.

The GTX 1650 SUPER’s sole victory is in the Geekbench Vulkan test, where it scores 50519 against the GTX 1070’s 22121. That represents a 128.4% lead for the GTX 1650 SUPER, a remarkable inversion of the overall trend. This suggests that the Turing architecture’s Vulkan driver or compute scheduling gives it a unique advantage in this specific workload. Outside of that single test, the GTX 1650 SUPER trails in every other category, including OpenCL (43875 vs 44700, a 1.8% deficit) and all PassMark tests.

For users prioritizing raw frame rates in conventional games, the GTX 1070 is the clear choice. The data shows it leads by 24.6% in PassMark G3D (13498 vs 10179) and by 26.6% in GPU compute (6102 vs 4477), making it more capable for both gaming and general compute tasks. The GTX 1650 SUPER’s win is narrow in scope but massive in magnitude—its Vulkan score is more than double the GTX 1070’s, which could matter for applications that leverage Vulkan compute heavily.

Architecture Differences

The two cards come from different architectural generations. The GTX 1650 SUPER is built on the Turing architecture, using the TU116 chip fabricated on a 12 nm process at TSMC. The GTX 1070 uses the Pascal architecture with the GP104 chip on a 16 nm process. This process difference is significant: the newer 12 nm node allows the GTX 1650 SUPER to pack 6,600 million transistors into a 284 mm² die, resulting in a transistor density of 23.2M / mm². The GTX 1070 has more transistors overall at 7,200 million, but its larger 314 mm² die yields a slightly lower density of 22.9M / mm².

Core configuration differs sharply. The GTX 1070 has 1920 shading units, 120 texture mapping units, and 64 ROPs. The GTX 1650 SUPER is cut down to 1280 shading units, 80 TMUs, and 32 ROPs. This gives the GTX 1070 a theoretical pixel rate of 107.7 GPixel/s and a texture rate of 202.0 GTexel/s, versus 55.20 GPixel/s and 138.0 GTexel/s for the GTX 1650 SUPER. The FP32 compute also favors the GTX 1070 at 6.463 TFLOPS compared to 4.416 TFLOPS for the GTX 1650 SUPER.

Memory architecture is another key divider. The GTX 1650 SUPER uses 4 GB of GDDR6 on a 128-bit bus, yielding 192.0 GB/s of bandwidth. The GTX 1070 uses 8 GB of GDDR5 on a 256-bit bus, providing 256.3 GB/s. The GTX 1070’s larger memory pool and 33% higher bandwidth give it a clear advantage in texture-heavy and high-resolution scenarios. Notably, the GTX 1650 SUPER’s FP16 throughput is 8.832 TFLOPS (2:1 ratio), while the GTX 1070’s is just 101.0 GFLOPS (1:64), indicating a dramatic difference in half-precision compute capability, though this does not translate into benchmark wins for the GTX 1650 SUPER outside Vulkan.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 test delivers the largest absolute gap. The GTX 1070 scores 1082 versus the GTX 1650 SUPER’s 352, a 67.5% delta. This test is a modern, demanding DX12 workload, and the result strongly indicates the GTX 1070’s superior fill rate and memory bandwidth matter more than architectural efficiency in this scenario.

The Geekbench Vulkan test reverses the narrative completely. The GTX 1650 SUPER scores 50519, beating the GTX 1070’s 22121 by 128.4%. This is the single largest percentage win for either card and suggests that the Turing architecture’s asynchronous compute or Vulkan command processing is exceptionally well-optimized here. It is a stark reminder that architectural generation can dominate raw hardware specs in specific APIs.

PassMark results show consistent GTX 1070 leadership across legacy APIs. In DirectX 9, the GTX 1070 scores 197 versus 148 (a 24.9% lead). DirectX 10 sees the GTX 1070 at 82 versus 50 (a 39% lead). DirectX 11 shows the GTX 1070 at 100 versus 73 (a 27% lead). DirectX 12 narrows the gap to just 6.2%, with the GTX 1070 scoring 48 against 45. This pattern suggests that while the GTX 1070 is faster across the board, the GTX 1650 SUPER becomes relatively more competitive as the API gets newer.

In OpenCL, the GTX 1070 wins narrowly at 44700 versus 43875, a 1.8% margin. This is the closest result outside of DirectX 12, indicating that compute performance is more balanced than gaming performance. The PassMark G2D test shows the GTX 1070 ahead at 846 versus 749 (an 11.5% lead), while PassMark G3D has the GTX 1070 at 13498 versus 10179 (a 24.6% lead), and GPU compute at 6102 versus 4477 (a 26.6% lead). The GTX 1070’s wins in G3D and compute are substantial, reinforcing its position as the more capable all-around card.

Specification Differences

The two cards differ on nearly every hardware specification except for their bus interface (PCIe 3.0 x16) and API support (DirectX 12_1, OpenGL 4.6, Vulkan 1.4). Key differences include:

• Process Node: GTX 1650 SUPER uses 12 nm; GTX 1070 uses 16 nm.
• Transistors: GTX 1650 SUPER has 6,600 million; GTX 1070 has 7,200 million.
• Die Size: GTX 1650 SUPER is 284 mm²; GTX 1070 is 314 mm².
• Base Clock: GTX 1650 SUPER runs at 1530 MHz; GTX 1070 at 1506 MHz.
• Boost Clock: GTX 1650 SUPER boosts to 1725 MHz; GTX 1070 to 1683 MHz.
• Memory: GTX 1650 SUPER has 4 GB GDDR6; GTX 1070 has 8 GB GDDR5.
• Memory Bus: GTX 1650 SUPER is 128-bit; GTX 1070 is 256-bit.
• Bandwidth: GTX 1650 SUPER offers 192.0 GB/s; GTX 1070 offers 256.3 GB/s.
• Shading Units: 1280 for the GTX 1650 SUPER, 1920 for the GTX 1070.
• TMUs: 80 vs 120.
• ROPs: 32 vs 64.
• TDP: GTX 1650 SUPER is 100 W; GTX 1070 is 150 W.
• Power Connectors: GTX 1650 SUPER uses 1x 6-pin; GTX 1070 uses 1x 8-pin.
• Suggested PSU: 300 W for the GTX 1650 SUPER, 450 W for the GTX 1070.
• Dimensions: GTX 1650 SUPER is 229 mm long; GTX 1070 is 267 mm long.

The GTX 1650 SUPER is smaller, cooler, and more power-efficient, while the GTX 1070 offers more memory, more cores, and higher bandwidth. These differences directly explain the benchmark outcomes, with the GTX 1070’s hardware advantages proving decisive in most tests.

FAQ

Q: Which card has a higher average benchmark score?

A: The GTX 1650 SUPER has an average benchmark score of 11047, while the GTX 1070 sits at 9780. This is despite the GTX 1070 winning 9 of 10 head-to-head tests, showing that the GTX 1650 SUPER’s Vulkan win heavily inflates its average.

Q: Why does the GTX 1650 SUPER win the Vulkan test by such a large margin?

A: The GTX 1650 SUPER scores 50519 in Geekbench Vulkan versus the GTX 1070’s 22121, a 128.4% lead. This likely stems from its Turing architecture’s newer Vulkan implementation, which offers better compute scheduling, though the pack provides no further technical detail.

Q: How much faster is the GTX 1070 in DirectX 11?

A: The GTX 1070 scores 100 in PassMark DirectX 11, compared to the GTX 1650 SUPER’s 73, representing a 27% advantage for the GTX 1070.

Q: What is the memory bandwidth difference?

A: The GTX 1070 has a 256.3 GB/s bandwidth on a 256-bit bus, while the GTX 1650 SUPER has 192.0 GB/s on a 128-bit bus. The GTX 1070’s bandwidth is roughly 33% higher.

Q: Which card has a lower TDP?

A: The GTX 1650 SUPER has a TDP of 100 W, while the GTX 1070 is rated at 150 W. The GTX 1650 SUPER also requires a 300 W PSU versus the 450 W suggested for the GTX 1070.

Q: Do both cards support the same APIs?

A: Yes, both support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. Their API feature sets are identical, despite the architectural differences.

The Verdict

The benchmark data is unambiguous for general use: the NVIDIA GeForce GTX 1070 is the superior card for most workloads. It wins 9 out of 10 head-to-head tests, with decisive leads in 3DMark Steel Nomad (67.5%), PassMark G3D (24.6%), and PassMark GPU compute (26.6%). Its larger memory pool of 8 GB, higher bandwidth of 256.3 GB/s, and greater core count make it the stronger choice for gaming, rendering, and traditional compute. Users who prioritize frame rates in modern DX12 games or run compute-heavy applications should choose the GTX 1070.

The GTX 1650 SUPER is not without merit, however. Its 128.4% lead in the Geekbench Vulkan test is remarkable and suggests that applications built on Vulkan compute could run significantly faster on this card. Additionally, its lower TDP of 100 W, smaller physical footprint (229 mm vs 267 mm), and 12 nm process node make it a more efficient and compact option. For users with tight power budgets or small form factor cases, the GTX 1650 SUPER offers a compelling trade-off, sacrificing the GTX 1070’s broad performance for a specific Vulkan advantage and lower power draw. The data does not support choosing the GTX 1650 SUPER for general gaming, but for Vulkan-specific workloads, it is demonstrably faster.