AMD Radeon HD 7790

The AMD Radeon HD 7790, built on the GCN 2.0 architecture with the Bonaire chip, occupies a specific niche in the benchmark hierarchy. Its single Geekbench Metal score of 17,666 places it at the 59th percentile of all GPUs, indicating a solid mid-range standing. The data reveals a card that is remarkably tightly clustered with its nearest competitors, all within a 1.1% performance delta, suggesting that architectural differences at this performance tier translate into minimal real-world variation.

How It Compares

Against the AMD Radeon Pro 460, the HD 7790 holds a marginal 0.5% lead. This is a statistical tie in practical terms, as the 17,666 score versus 17,575 falls well within normal benchmark variance. The implication is that users transitioning between these two cards would experience virtually identical raw compute performance in Metal-accelerated workloads, despite any generational or feature-set differences not captured by this single metric.

The AMD FirePro W7000 edges out the HD 7790 by 0.7%, with the rival scoring 17,790. This slim deficit means the HD 7790 trails by roughly 124 points, a negligible gap that underscores how workstation-oriented and consumer-oriented cards from the same era converged in raw throughput. The data suggests the HD 7790's GCN 2.0 architecture was competitive with professional-tier offerings of its time.

Relative to the AMD Radeon Pro 560, the HD 7790 is 1.0% faster, posting 17,666 against 17,497. This 169-point advantage is the largest margin among the four rivals, yet it remains under two percent. The benchmark results indicate that the HD 7790 holds a slight but consistent edge over this mobile-derived Pro variant, likely reflecting differences in clock behavior or memory configuration that the scores capture.

The NVIDIA Tesla K40c is the closest competitor, with the HD 7790 leading by 1.1%. The Tesla scores 17,468, meaning the HD 7790 outperforms it by 198 points. This is notable because the Tesla K40c is a compute-oriented accelerator, yet the HD 7790's consumer-focused design still manages to surpass it in this particular Metal benchmark, hinting at efficient shader utilization in the GCN architecture.

Power and Cooling

The HD 7790 carries a thermal design power (TDP) of 85 W, which is modest for a dual-slot card. This power envelope allows for a straightforward cooling solution, and the dual-slot form factor provides adequate surface area for heat dissipation without requiring exotic cooler designs. The data shows a card that should remain thermally manageable in typical desktop chassis.

The suggested power supply is 250 W, which is a low bar for modern systems. However, the card requires a single 6-pin power connector, meaning the PSU must have that specific connector available. The 85 W TDP combined with the 250 W PSU recommendation suggests the card draws well within standard PCIe slot and auxiliary power limits, making it compatible with older or entry-level power supplies that meet the connector requirement.

The bus interface is PCIe 3.0 x16, which is fully compatible with modern motherboards, though the card's performance ceiling means it will not saturate even PCIe 2.0 bandwidth in most scenarios. The physical dimensions are 183 mm (7.2 inches) in length, a compact size that fits in most mid-tower cases, while the dual-slot width occupies two expansion brackets.

Benchmark Performance

The Geekbench Metal score of 17,666 serves as the sole performance metric in the data. At the 59th percentile of all GPUs, the HD 7790 sits in the upper-middle range, outperforming more than half of all recorded graphics cards. This percentile placement, combined with the tight rival clustering, indicates a card that delivers consistent, predictable performance for its era.

The 1.1% lead over the NVIDIA Tesla K40c (17,468) is the most significant delta in the rival set. This means the HD 7790 delivers roughly 1.1% more Metal compute performance than the Tesla, a surprising result given the Tesla's enterprise positioning. Conversely, the 0.7% deficit to the AMD FirePro W7000 (17,790) shows the HD 7790 is 0.7% slower than that workstation card, placing it in a narrow performance band between 17,468 and 17,790.

The deltaPct values of 0.5, -0.7, 1.0, and 1.1 against the four rivals paint a picture of extreme parity. The total spread from the slowest rival (Tesla K40c at 17,468) to the fastest (FirePro W7000 at 17,790) is just 322 points, or 1.8%. Benchmark results indicate that the HD 7790's performance is firmly anchored in this cluster, with no single competitor achieving a decisive advantage. The 1.792 TFLOPS FP32 throughput and 56.00 GTexel/s texture rate provide the theoretical underpinning for these observed scores.

FAQ

Q: How does the HD 7790 compare to the AMD Radeon Pro 460?

A: The HD 7790 is 0.5% faster, scoring 17,666 versus the Pro 460's 17,575 in Geekbench Metal.

Q: What is the memory bandwidth of the HD 7790?

A: The card features 96.00 GB/s bandwidth, derived from 1024 MB of GDDR5 memory on a 128-bit bus running at 6 Gbps effective.

Q: Does the HD 7790 support modern graphics APIs?

A: Yes, the data lists DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170 support.

Q: What power supply is recommended for the HD 7790?

A: A 250 W PSU is suggested, with the card requiring a single 6-pin power connector.

Q: Is the HD 7790 faster than the NVIDIA Tesla K40c?

A: Yes, by 1.1%, with scores of 17,666 and 17,468 respectively.

Q: What is the pixel fill rate of the HD 7790?

A: The card achieves 16.00 GPixel/s, based on its 16 ROPs.

Who Should Consider It

The HD 7790's 59th percentile ranking and 17,666 Geekbench Metal score suggest it is suited for 1080p gaming at medium to high settings, though the data does not include specific gaming benchmarks. The 1.792 TFLOPS FP32 performance indicates capability for older titles and less demanding modern games, but users targeting high refresh rates or maximum detail at 1440p would likely find the card insufficient.

For users with a 250 W power supply and a single 6-pin connector, the HD 7790 presents a low-conflict upgrade path. The 85 W TDP means it will not strain aging power supplies, and the 183 mm length fits most compact cases. However, the 1.1% lead over the Tesla K40c and 1.0% lead over the Pro 560 are marginal, so users already owning those cards would see no meaningful benefit from switching.

The card's 16 ROPs and 56 TMUs, combined with 896 shading units, suggest it is best paired with 1080p displays. At this resolution, the 96.00 GB/s bandwidth and 16.00 GPixel/s pixel rate are adequate for the era's standard settings. Users with 4K displays should look elsewhere, as the memory configuration and compute throughput would struggle with the increased pixel workload.

Memory Subsystem

The HD 7790 is equipped with 1024 MB of GDDR5 memory, a modest capacity even for its release period. The 128-bit bus width is narrow by modern standards, but the 1500 MHz memory clock (6 Gbps effective) delivers 96.00 GB/s of bandwidth. This figure is the key constraint for high-resolution operation, as texture streaming and framebuffer access at 1440p or 4K would quickly exhaust both the capacity and the bandwidth.

For 1080p gaming, 1024 MB is workable for older titles, but modern games with high-resolution texture packs would exceed this limit, causing stuttering or texture pop-in. The 96.00 GB/s bandwidth is sufficient for the 56.00 GTexel/s texture rate, meaning the card can feed its TMUs without bottlenecking in most scenarios. However, the 16.00 GPixel/s pixel rate suggests that fill-rate-bound scenes at higher resolutions will reveal the memory subsystem's limitations.

The effective memory speed of 6 Gbps is notable for the 128-bit bus, as it partially compensates for the narrow width. Compared to the rival cluster, all of which score within 1.1%, the memory configuration appears competitive for the performance tier, but the 1024 MB capacity is the most limiting factor for longevity.

Ray Tracing and Feature Set

The HD 7790 does not include dedicated ray tracing or tensor cores, as indicated by the null values in the FACT PACK. This places it firmly in the pre-DXR era, where ray tracing was not a consumer feature. Instead, the card relies on its 896 shading units for all compute and graphics workloads, which is reflected in the 1.792 TFLOPS FP32 throughput.

The API support is surprisingly robust for the card's age, with DirectX 12 (12_0) and Vulkan 1.2.170 support listed. This means the card can run modern API titles, though the lack of hardware acceleration for ray tracing means any such effects would be software-emulated, resulting in poor performance. The OpenGL 4.6 support ensures compatibility with legacy applications and emulators.

The display outputs include 1x DVI, 1x HDMI 1.4a, and 2x mini-DisplayPort 1.2, providing flexible multi-monitor options. However, the absence of RT/tensor cores means the card offers no advantage in AI-accelerated workloads or ray-traced games. The GCN 2.0 architecture's strength lies in traditional rasterization, where the 16 ROPs and 56 TMUs deliver the measured 16.00 GPixel/s and 56.00 GTexel/s rates. Users requiring ray tracing would need a significantly more modern card, as the HD 7790's feature set is anchored to its 2013 release era.