NVIDIA GeForce GTX 590 vs NVIDIA GeForce GTX 950 Comparison

The NVIDIA GeForce GTX 950 and NVIDIA GeForce GTX 590 represent two distinct design philosophies from the same manufacturer, separated by four years of architectural evolution. The GTX 950, a Maxwell 2.0 product from the GeForce 900 generation, faces off against the GTX 590, a dual-GPU Fermi 2.0 flagship from the GeForce 500 generation. The data reveals a modern mid-range card outperforming an older dual-chip behemoth, with the GTX 950 securing a 22.1% victory in the only shared benchmark, Geekbench OpenCL, scoring 15,662 against the GTX 590's 12,830. This result is particularly striking given the GTX 590's much larger physical footprint and power demands, suggesting that architectural efficiency can trump raw silicon scale in legacy applications.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 950 has a higher average benchmark score of 13,189, compared to the NVIDIA GeForce GTX 590's 12,830. This places the GTX 950 in the 53rd percentile of all GPUs, while the GTX 590 sits in the 52nd percentile.

Q: How significant is the performance gap in the head-to-head OpenCL test?

A: The GTX 950 leads by 22.1% in the Geekbench OpenCL benchmark, scoring 15,662 versus the GTX 590's 12,830. This is the only direct benchmark comparison available, and it represents a substantial margin of victory for the newer card.

Q: What are the closest rivals to each card based on average score?

A: The GTX 950's nearest rival is the NVIDIA GeForce GTX 480, which has an average score of 13,300 and is 0.8% behind. For the GTX 590, the closest competitor is the AMD Radeon Pro 455 with an average score of 12,831, showing a 0% delta, making it essentially a statistical tie.

Q: How do the memory configurations differ between the two cards?

A: The GTX 950 features 2 GB of GDDR5 memory on a 128-bit bus with 105.8 GB/s of bandwidth, while the GTX 590 has 1536 MB of GDDR5 on a wider 384-bit bus, delivering 164.0 GB/s. Despite the GTX 590's higher bandwidth, the GTX 950's greater capacity and newer memory clock (6.6 Gbps effective versus 3.4 Gbps) contribute to its overall performance advantage.

Q: Which card has the higher transistor density, and what does this imply?

A: The GTX 950 has a transistor density of 12.9 million transistors per mm², more than double the GTX 590's 5.8M / mm². The GTX 950 packs 2,940 million transistors into a 228 mm² die, whereas the GTX 590 uses 3,000 million transistors across a massive 520 mm² die, highlighting the efficiency gains of the newer 28 nm process over the older 40 nm node.

Q: What is the difference in their DirectX API support?

A: The GTX 950 supports DirectX 12 (12_1), while the GTX 590 is limited to DirectX 12 (11_0). This means the GTX 950 can fully utilize feature level 12_1 capabilities, whereas the GTX 590 operates at the older 11_0 feature level despite nominally supporting DirectX 12.

Architecture Differences

The architectural chasm between these two GPUs is defined by their manufacturing processes and core designs. The GTX 950 is built on TSMC's 28 nm process node, enabling a compact 228 mm² die that houses 2,940 million transistors. In contrast, the GTX 590 uses a 40 nm process, resulting in a sprawling 520 mm² die with 3,000 million transistors. The density difference is stark: 12.9 million transistors per square millimeter for the GTX 950 versus just 5.8 million for the GTX 590, demonstrating how process shrinks allow more logic in less space.

The GTX 950 employs the Maxwell 2.0 architecture, which represents a significant departure from the older Fermi 2.0 architecture found in the GTX 590. This architectural shift is evident in the core configuration: the GTX 950 packs 768 shading units, 48 texture mapping units (TMUs), and 32 render output units (ROPs). The GTX 590, despite being a dual-GPU card, offers only 512 shading units per GPU (though the data lists 512 total), 64 TMUs, and 48 ROPs. The Maxwell design achieves higher throughput with fewer, more efficient cores, as reflected in the GTX 950's pixel rate of 38.02 GPixel/s and texture rate of 57.02 GTexel/s, both exceeding the GTX 590's respective 19.46 GPixel/s and 38.91 GTexel/s.

Another fundamental difference lies in the memory subsystem. The GTX 950 uses a 128-bit memory interface with 2 GB of GDDR5 running at 6.6 Gbps effective, yielding 105.8 GB/s of bandwidth. The GTX 590 counters with a 384-bit bus and 1536 MB of GDDR5 at 3.4 Gbps effective, providing 164.0 GB/s. The GTX 590's wider bus gives it higher peak bandwidth, but the GTX 950's smaller capacity and faster clock speed suggest better latency management. The GTX 950 also supports PCIe 3.0 x16, while the GTX 590 is limited to PCIe 2.0 x16, offering double the interconnect bandwidth for data transfer.

Head-to-Head Benchmarks

The sole head-to-head benchmark between these two cards is the Geekbench OpenCL test, and the results are decisively in favor of the GTX 950. The GTX 950 scores 15,662, while the GTX 590 manages 12,830, giving the newer card a 22.1% advantage. This is a substantial margin that underscores the efficiency of Maxwell's compute architecture over Fermi's older design. The GTX 950's higher FP32 throughput of 1.825 TFLOPS, compared to the GTX 590's 1,244.2 GFLOPS, directly contributes to this OpenCL performance gap.

In the broader benchmark context, the GTX 950 also demonstrates superiority in other tests, though these are not head-to-head comparisons. The GTX 950 achieves a Geekbench Metal score of 7,172 and a Geekbench Vulkan score of 16,734, showing strong modern API support. The GTX 590 has no Metal or Vulkan benchmarks listed, and it lacks Vulkan API support entirely according to the data. The GTX 950's average benchmark score of 13,189 versus the GTX 590's 12,830 reinforces the head-to-head result, even when accounting for different test suites.

The nearest rival data provides additional context for these scores. The GTX 950's average score places it just 0.8% behind the NVIDIA GeForce GTX 480 (13,300) and 1% behind the AMD Radeon Pro 555X (13,321), while being 0.9% ahead of the NVIDIA Tesla M2090 (13,075). The GTX 590's average score is virtually tied with the AMD Radeon Pro 455 (12,831) and only 0.4% ahead of the NVIDIA GeForce GTX 670 (12,773). This suggests that the GTX 950 competes in a slightly higher performance tier than the GTX 590, despite the latter's dual-GPU configuration and higher launch MSRP.

Specification Differences

The specification sheets reveal a generational leap in efficiency and capability. The GTX 950 has a base clock of 1024 MHz and a boost clock of 1188 MHz, whereas the GTX 590 has no base or boost clock listed, indicating it relies on fixed clock behavior. Memory clocks differ significantly: the GTX 950 runs at 1653 MHz (6.6 Gbps effective), while the GTX 590 operates at 854 MHz (3.4 Gbps effective). The GTX 950's memory interface is 128-bit with 2 GB capacity, while the GTX 590 uses a 384-bit bus with 1536 MB.

Power requirements present a dramatic contrast. The GTX 950 has a TDP of 90 W and requires a single 6-pin power connector with a suggested 250 W power supply. The GTX 590, by comparison, has a 365 W TDP, needs two 8-pin connectors, and demands a 750 W power supply. Physical dimensions also differ: the GTX 950 measures 202 mm (8 inches) in length, while the GTX 590 spans 279 mm (11 inches) with a height of 111 mm (4.4 inches) and width of 40 mm (1.6 inches). The GTX 950 offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2 outputs, whereas the GTX 590 provides 3x DVI and 1x mini-DisplayPort.

API support diverges as well: the GTX 950 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while the GTX 590 supports DirectX 12 (11_0) and OpenGL 4.6 but has no Vulkan support. The GTX 950's release date is 2015-08-19, with the GTX 590 released earlier on 2011-03-23. The launch MSRP for the GTX 950 is 159 USD, while the GTX 590 launched at 699 USD.

Where Each One Wins

The GTX 950 wins decisively in compute performance, as evidenced by its 22.1% lead in the OpenCL benchmark and its higher FP32 throughput. Its modern architecture supports Vulkan 1.4 and DirectX 12 (12_1), making it far more suitable for contemporary applications and games that leverage these APIs. The GTX 950's lower power draw (90 W versus 365 W) and smaller physical footprint (202 mm versus 279 mm length) make it a practical choice for compact builds or systems with modest power supplies. Its higher pixel rate (38.02 GPixel/s) and texture rate (57.02 GTexel/s) indicate superior rasterization throughput, which benefits traditional rendering workloads.

The GTX 590's advantages are more limited but still notable. It offers higher memory bandwidth at 164.0 GB/s, a 55% improvement over the GTX 950's 105.8 GB/s, which could benefit workloads that are heavily bandwidth-bound. The GTX 590 also has more TMUs (64 versus 48) and ROPs (48 versus 32), suggesting theoretical advantages in texture-heavy or fill-rate-limited scenarios. Its larger memory bus (384-bit versus 128-bit) provides a foundation for high-throughput data access, though this is partially offset by lower effective clock speeds. The GTX 590's 3x DVI outputs might appeal to users with legacy multi-monitor setups, though the GTX 950's DisplayPort 1.2 support is more future-proof.

In terms of benchmark coverage, the GTX 950 has three recorded benchmarks (Metal, OpenCL, Vulkan), while the GTX 590 has only one (OpenCL). This disparity means the GTX 950's average score of 13,189 is derived from a broader set of tests, potentially reflecting more varied workloads. The GTX 950's percentile ranking of 53 versus the GTX 590's 52 indicates a marginal overall standing advantage, but the head-to-head result is what truly separates them.

The Verdict

The data overwhelmingly favors the NVIDIA GeForce GTX 950 for general-purpose and modern computing tasks. Its 22.1% lead in the shared OpenCL benchmark, combined with support for Vulkan 1.4 and DirectX 12 (12_1), makes it the clear choice for users running contemporary software. The GTX 950's efficiency is remarkable: it achieves higher performance with a 90 W TDP versus the GTX 590's 365 W, and it does so in a card that is 77 mm shorter. For anyone building a system today or upgrading from older hardware, the GTX 950 offers better compute performance, broader API compatibility, and significantly lower power demands.

The GTX 590's sole advantage lies in memory bandwidth, where its 164.0 GB/s exceeds the GTX 950's 105.8 GB/s. This could make it marginally better for specific bandwidth-intensive workloads like certain scientific simulations or large data transfers, provided the software can utilize its older Fermi architecture effectively. However, the GTX 590's lack of Vulkan support and limited DirectX 12 feature level (11_0) severely constrain its modern usability. Its dual-GPU design also introduces complexity without delivering proportional performance, as the benchmark data shows it trailing a single mid-range Maxwell chip.

For most users, the GTX 950 is the rational choice based on benchmark results alone. It ranks in a higher percentile (53 versus 52), has a higher average score (13,189 versus 12,830), and wins the only direct comparison. The GTX 590's higher launch MSRP of 699 USD compared to the GTX 950's 159 USD further highlights the value proposition, though pricing should not be the primary consideration. Ultimately, the GTX 590 remains a historical curiosity—a powerful dual-GPU card from 2011 that has been surpassed by a more efficient 2015 design. The GTX 950 demonstrates that architectural progress and process node improvements can overcome raw hardware specifications, delivering superior performance at a fraction of the power and size.