AMD Radeon R9 370X vs NVIDIA GeForce GTX 690 Comparison

The NVIDIA GeForce GTX 690 and AMD Radeon R9 370X represent two very different approaches to GPU design from different eras of the 28 nm process node. The GTX 690 is a dual-GPU flagship from the GeForce 600 generation, while the R9 370X is a mid-range single-chip card from the later Pirate Islands lineup. Benchmark data shows a split decision: the R9 370X dominates in OpenCL compute workloads, while the GTX 690 holds a massive lead in Vulkan graphics performance. Their average benchmark scores differ by only 6.9% in favor of the GTX 690, but the underlying performance profiles could hardly be more distinct.

Head-to-Head Benchmarks

The most striking result in the head-to-head comparison is the Geekbench OpenCL score, where the AMD Radeon R9 370X posts 25,893 points against the GTX 690's 17,399. That is a delta of -32.8% from the perspective of the NVIDIA card, meaning the R9 370X outperforms the GTX 690 by roughly 48.8% in this compute-oriented test. This is a substantial margin for a card with fewer shading units (1,280 vs. 1,536) and lower peak FP32 throughput (2.637 TFLOPS vs. 3.130 TFLOPS). The result suggests that the GCN 1.0 architecture's compute scheduling and driver optimization for OpenCL workloads are significantly more efficient than the Kepler architecture's approach, despite the latter's higher theoretical specifications.

The Vulkan benchmark tells the opposite story with even greater severity. The GTX 690 scores 16,675 points, while the R9 370X manages only 9,018 points. The deltaPct of 84.9% indicates the NVIDIA card is 84.9% faster than the AMD card in this test. This is a massive generational gap in graphics API performance. The GTX 690's Kepler architecture, despite being older, delivers nearly double the Vulkan performance of the R9 370X. This could be attributed to driver maturity, architectural efficiency in draw call processing, or differences in how each GPU handles the Vulkan workload's specific characteristics.

When examining the average benchmark scores, the GTX 690 edges ahead with 17,037 points versus 15,862 points for the R9 370X. This places the GTX 690 at the 60th percentile of all GPUs, while the R9 370X sits at the 58th percentile. The nearest rivals for the GTX 690 include the AMD Radeon RX 7600 XT at 17,083 points (0.3% ahead) and the NVIDIA GeForce RTX 3070 at 17,208 points (1% ahead), showing that the GTX 690's average score is competitive with much newer mainstream cards. The R9 370X's nearest rivals include the AMD Radeon RX 7700 at 15,852 points (0.1% behind) and the NVIDIA GeForce RTX 3060 Ti at 16,129 points (1.7% ahead), indicating it sits in a similar performance tier despite its age.

The wins are split evenly at one each, but the magnitude of the victories is not symmetrical. The R9 370X's OpenCL win is significant but not overwhelming, while the GTX 690's Vulkan win is absolutely decisive. This suggests that workload selection is paramount when choosing between these two cards. For compute-heavy tasks leveraging OpenCL, the R9 370X is clearly superior; for graphics-centric workloads using Vulkan, the GTX 690 is in a different league.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA GeForce GTX 690 has an average benchmark score of 17,037, which is higher than the AMD Radeon R9 370X's 15,862. This puts the GTX 690 at the 60th percentile of all GPUs, compared to the 58th percentile for the R9 370X.

Q: How large is the Vulkan performance gap between the two cards?

A: In the Geekbench Vulkan test, the GTX 690 scores 16,675 points versus 9,018 points for the R9 370X. The deltaPct of 84.9% means the NVIDIA card is 84.9% faster, making this the single largest performance disparity in the head-to-head comparison.

Q: Does the R9 370X beat the GTX 690 in any benchmark?

A: Yes, in the Geekbench OpenCL test, the R9 370X scores 25,893 points compared to the GTX 690's 17,399 points. This represents a -32.8% deltaPct from the GTX 690's perspective, meaning the AMD card is roughly 48.8% faster in this compute workload.

Q: How does the GTX 690 compare to modern GPUs in its nearest rivals list?

A: The GTX 690's average score of 17,037 is only 0.3% lower than the AMD Radeon RX 7600 XT (17,083) and 1% lower than the NVIDIA GeForce RTX 3070 (17,208). It also edges out the AMD Radeon HD 7970M (17,019) by 0.1% and the NVIDIA Tesla M4 (16,932) by 0.6%.

Q: What is the transistor density difference between the two chips?

A: The AMD Radeon R9 370X's Trinidad chip has a transistor density of 13.2 million transistors per square millimeter, which is higher than the GTX 690's GK104 chip at 12.0 million per square millimeter. This is despite the R9 370X having fewer total transistors (2,800 million vs. 3,540 million) and a smaller die size (212 mm² vs. 294 mm²).

Q: Which card has the higher boost clock speed?

A: The AMD Radeon R9 370X has a boost clock of 1030 MHz, while the NVIDIA GeForce GTX 690 has a boost clock of 1019 MHz. The R9 370X also has a higher base clock at 980 MHz versus 915 MHz for the GTX 690.

Architecture Differences

The two cards are built on fundamentally different GPU architectures. The GTX 690 uses the GK104 chip based on NVIDIA's Kepler architecture, while the R9 370X uses the Trinidad chip based on AMD's GCN 1.0 architecture. Both are fabricated by TSMC on a 28 nm process node, but the similarities end there.

The GK104 chip contains 3,540 million transistors on a die size of 294 mm², resulting in a transistor density of 12.0 million per square millimeter. The Trinidad chip is smaller and less complex, with 2,800 million transistors on a 212 mm² die, giving it a higher density of 13.2 million per square millimeter. This density advantage suggests AMD packed transistors more efficiently, but the GTX 690's larger chip provides more raw resources.

In terms of compute resources, the GTX 690 has 1,536 shading units, 128 texture mapping units, and 32 ROPs. The R9 370X has fewer shading units at 1,280, fewer TMUs at 80, but the same number of ROPs at 32. This configuration gives the GTX 690 a texture rate of 130.4 GTexel/s, significantly higher than the R9 370X's 82.40 GTexel/s. The pixel rates are nearly identical, with the GTX 690 at 32.61 GPixel/s and the R9 370X at 32.96 GPixel/s.

Clock speeds favor the AMD card. The R9 370X runs at a base clock of 980 MHz and boosts to 1030 MHz, while the GTX 690 runs at 915 MHz base and 1019 MHz boost. Despite the clock advantage, the GTX 690's higher FP32 throughput of 3.130 TFLOPS exceeds the R9 370X's 2.637 TFLOPS, highlighting the NVIDIA card's superior per-clock efficiency in raw compute.

Memory subsystems are similar in width but differ in speed. Both cards use 2 GB of GDDR5 memory on a 256-bit bus. The GTX 690's memory runs at 1502 MHz (6 Gbps effective), yielding a bandwidth of 192.3 GB/s. The R9 370X's memory runs at 1400 MHz (5.6 Gbps effective), providing 179.2 GB/s of bandwidth. The GTX 690's 7.3% bandwidth advantage aligns with its higher average benchmark score.

API support shows distinct differences. The GTX 690 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The R9 370X supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The AMD card has a slightly higher DirectX feature level, while the NVIDIA card has a marginally newer Vulkan version. Neither card supports hardware ray tracing or tensor cores, as indicated by the null values for those fields.

Power and physical specifications differ substantially. The GTX 690 has a TDP of 300 W and requires two 8-pin power connectors, with a suggested PSU of 700 W. The R9 370X has a much lower TDP of 180 W, needs only two 6-pin connectors, and recommends a 450 W PSU. The GTX 690 is also longer at 279 mm (11 inches) compared to 221 mm (8.7 inches) for the R9 370X, though both are dual-slot cards with the same height of 111 mm.

The Verdict

The data presents a clear bifurcation. For users whose primary workloads are OpenCL compute applications, the AMD Radeon R9 370X is the superior choice. Its OpenCL score of 25,893 versus 17,399 for the GTX 690 represents a decisive 48.8% advantage that outweighs its lower average benchmark score. The R9 370X also offers lower power consumption (180 W TDP vs. 300 W), a smaller physical footprint, and a less demanding power supply requirement (450 W vs. 700 W).

For graphics-intensive workloads, particularly those leveraging Vulkan, the NVIDIA GeForce GTX 690 is overwhelmingly dominant. Its Vulkan score of 16,675 is 84.9% higher than the R9 370X's 9,018, a gap that no other benchmark metric comes close to matching. The GTX 690 also holds advantages in raw texture rate (130.4 GTexel/s vs. 82.40 GTexel/s), FP32 throughput (3.130 TFLOPS vs. 2.637 TFLOPS), and memory bandwidth (192.3 GB/s vs. 179.2 GB/s).

The average benchmark scores reflect this mixed picture, with the GTX 690 leading 17,037 to 15,862. This 6.9% average advantage, combined with the 60th vs. 58th percentile ranking, suggests the GTX 690 is the more well-rounded performer. However, the R9 370X's OpenCL supremacy cannot be ignored for compute-centric users.

Neither card is a universal winner. The GTX 690 is the choice for Vulkan gaming and graphics workloads, while the R9 370X is the choice for OpenCL compute tasks and users prioritizing efficiency. The 2012 release of the GTX 690 versus the 2015 release of the R9 370X shows that architectural efficiency and driver maturity can overcome generational gaps, but only in specific workload domains.

Specification Differences

| Specification | NVIDIA GeForce GTX 690 | AMD Radeon R9 370X |

|---|---|---|

| Architecture | Kepler | GCN 1.0 |

| Chip | GK104 | Trinidad |

| Generation | GeForce 600 | Pirate Islands (R9 300) |

| Transistors | 3,540 million | 2,800 million |

| Die Size | 294 mm² | 212 mm² |

| Transistor Density | 12.0M / mm² | 13.2M / mm² |

| Base Clock | 915 MHz | 980 MHz |

| Boost Clock | 1019 MHz | 1030 MHz |

| Memory Clock | 1502 MHz (6 Gbps) | 1400 MHz (5.6 Gbps) |

| Memory Bandwidth | 192.3 GB/s | 179.2 GB/s |

| Shading Units | 1536 | 1280 |

| TMUs | 128 | 80 |

| Pixel Rate | 32.61 GPixel/s | 32.96 GPixel/s |

| Texture Rate | 130.4 GTexel/s | 82.40 GTexel/s |

| FP32 | 3.130 TFLOPS | 2.637 TFLOPS |

| TDP | 300 W | 180 W |

| Power Connectors | 2x 8-pin | 2x 6-pin |

| Suggested PSU | 700 W | 450 W |

| Display Outputs | 3x DVI, 1x mini-DisplayPort | 2x DVI, 1x HDMI, 1x DisplayPort |

| DirectX | 12 (11_0) | 12 (11_1) |

| Vulkan | 1.2.175 | 1.2.170 |

| Length | 279 mm (11 inches) | 221 mm (8.7 inches) |

| Launch MSRP | 999 USD | 199 USD |