NVIDIA GeForce GTX 580 vs NVIDIA GeForce GTX 680 Comparison

# The Verdict

The GeForce GTX 680 is the clear winner in raw compute performance, but the GeForce GTX 580 holds a surprising edge in overall market positioning. The GTX 680 beats the GTX 580 by 9.6% in Geekbench OpenCL, scoring 16,906 versus 15,283. However, the GTX 580 achieves a higher percentile ranking among all GPUs — 58th percentile versus 55th — despite losing the head-to-head benchmark. This means the GTX 580 sits closer to its nearest rivals in performance, while the GTX 680's average score of 14,150 is dragged down by its weaker Metal and Vulkan results.

For users prioritizing compute performance in OpenCL workloads, the GTX 680 is the pick. It delivers higher raw throughput with a 3.250 TFLOPS FP32 rate, more shading units, and faster texture and pixel rates. For those who need consistent performance across a broader range of benchmark scenarios, the GTX 580's single OpenCL score of 15,283 places it within 0.1% of the AMD Radeon 680M and 0.6% of the RTX 3050 OEM, showing it remains competitive in its class despite being older.

The GTX 680 also wins on efficiency and platform modernity. It consumes 195 W versus 244 W, requires a 450 W PSU instead of 550 W, and supports PCIe 3.0 x16, Vulkan 1.2.175, and DisplayPort 1.2. The GTX 580 is limited to PCIe 2.0 x16, lacks Vulkan support, and only offers mini-HDMI 1.3a. Both cards launched at the same MSRP of 499 USD, but the GTX 680 delivers more performance per watt and newer feature support for that identical launch price.

Architecture Differences

The GTX 580 uses the GF110 chip based on the Fermi 2.0 architecture, built on TSMC's 40 nm process. The GTX 680 uses the GK104 chip based on the Kepler architecture, built on TSMC's 28 nm process. This node shrink allows the GTX 680 to pack 3,540 million transistors into a 294 mm² die, achieving a transistor density of 12.0M per mm². The GTX 580 has 3,000 million transistors on a much larger 520 mm² die, yielding only 5.8M per mm².

The Kepler architecture fundamentally reworks the shader arrangement. The GTX 680 has 1,536 shading units, three times the GTX 580's 512. It also doubles the texture mapping units to 128 versus 64. However, the GTX 580 has more ROPs at 48 versus 32. The GTX 680 compensates with higher clock speeds — a 1006 MHz base and 1058 MHz boost — while the GTX 580 has no listed base or boost clock. Memory runs at 1502 MHz (6 Gbps effective) on the GTX 680 versus 1002 MHz (4 Gbps effective) on the GTX 580.

Memory capacity and bus width differ significantly. The GTX 580 uses 1536 MB of GDDR5 on a 384-bit bus, delivering 192.4 GB/s bandwidth. The GTX 680 uses 2 GB of GDDR5 on a 256-bit bus, delivering 192.3 GB/s — nearly identical bandwidth despite the narrower bus. The GTX 680 adds Vulkan 1.2.175 API support, which the GTX 580 completely lacks. Both support DirectX 12 (11_0) and OpenGL 4.6.

The GTX 680 is physically shorter at 256 mm versus 267 mm for the GTX 580, and it has a defined width of 38 mm. Both are dual-slot cards with 111 mm height. Power delivery differs: the GTX 580 needs 1x 6-pin plus 1x 8-pin connectors, while the GTX 680 only needs 2x 6-pin.

FAQ

Q: Which card has better OpenCL performance?

A: The GTX 680 wins the Geekbench OpenCL test with a score of 16,906, which is 9.6% higher than the GTX 580's 15,283.

Q: Does the GTX 580 support Vulkan?

A: No. The GTX 580 has no Vulkan API support listed, while the GTX 680 supports Vulkan version 1.2.175.

Q: What is the difference in power consumption?

A: The GTX 580 has a TDP of 244 W and requires a 550 W PSU, while the GTX 680 has a TDP of 195 W and a 450 W suggested PSU.

Q: Which card has more memory and bandwidth?

A: The GTX 680 has 2 GB of GDDR5 on a 256-bit bus with 192.3 GB/s bandwidth. The GTX 580 has 1536 MB on a 384-bit bus with 192.4 GB/s bandwidth. Bandwidth is nearly identical, but the GTX 680 has more capacity.

Q: How do their market positions compare?

A: The GTX 580 sits at the 58th percentile of all GPUs, while the GTX 680 sits at the 55th percentile. The GTX 580's nearest rival is the RTX 2060 with a 0% delta, while the GTX 680's nearest rival is the GTX 1070 Ti at -0.9%.

Q: What display outputs does each card offer?

A: The GTX 580 has 2x DVI and 1x mini-HDMI 1.3a. The GTX 680 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2.

Specification Differences

The two cards differ across nearly every major specification. The GTX 680 is built on a 28 nm process versus 40 nm for the GTX 580. Transistor count is 3,540 million versus 3,000 million, but die size is smaller at 294 mm² versus 520 mm². Transistor density is 12.0M/mm² versus 5.8M/mm².

The GTX 680 has a base clock of 1006 MHz and boost clock of 1058 MHz; the GTX 580 has no listed base or boost clocks. Memory clocks are 1502 MHz (6 Gbps effective) versus 1002 MHz (4 Gbps effective). Memory size is 2 GB versus 1536 MB, bus width is 256-bit versus 384-bit, and bandwidth is 192.3 GB/s versus 192.4 GB/s.

Shading units are 1,536 versus 512. TMUs are 128 versus 64. ROPs are 32 versus 48. Pixel rate is 33.86 GPixel/s versus 24.70 GPixel/s. Texture rate is 135.4 GTexel/s versus 49.41 GTexel/s. FP32 performance is 3.250 TFLOPS versus 1.581 TFLOPS.

TDP is 195 W versus 244 W. Power connectors are 2x 6-pin versus 1x 6-pin + 1x 8-pin. Suggested PSU is 450 W versus 550 W. Bus interface is PCIe 3.0 x16 versus PCIe 2.0 x16. Display outputs are 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 versus 2x DVI, 1x mini-HDMI 1.3a. The GTX 680 supports Vulkan 1.2.175; the GTX 580 does not. Length is 256 mm versus 267 mm, and the GTX 680 has a width of 38 mm while the GTX 580 has none listed.

Head-to-Head Benchmarks

In the only direct comparison available, the GTX 680 wins the Geekbench OpenCL test decisively. The GTX 680 scores 16,906, which is 9.6% ahead of the GTX 580's 15,283. This margin aligns with the raw compute advantage: the GTX 680 delivers 3.250 TFLOPS FP32, more than double the GTX 580's 1.581 TFLOPS. The GTX 680 also has triple the shading units and double the texture rate, which explains its superior OpenCL throughput.

However, the GTX 580's single benchmark score of 15,283 is remarkably close to several modern GPUs. It trails the RTX 2060 by only 0% (a 7-point difference), beats the AMD Radeon 680M by 0.1%, and leads the RTX 3050 OEM by 0.6%. The GTX 680's OpenCL score of 16,906 is not directly compared to these rivals, but its average score of 14,150 includes weaker results from other tests. The GTX 680's Metal score is 7,897 and its Vulkan score is 17,646, but the GTX 580 has no Metal or Vulkan scores listed.

The GTX 680's average benchmark score of 14,150 is actually lower than its OpenCL score, indicating that its Metal result drags down the average. The GTX 580's average equals its OpenCL score at 15,283, since that's its only benchmark. This explains the percentile discrepancy: despite winning the head-to-head, the GTX 680 ranks lower overall because its average is pulled down by the Metal test.

Where Each One Wins

The GTX 680 wins in raw compute performance, efficiency, and modern feature support. It is 9.6% faster in OpenCL, has double the FP32 throughput, and delivers higher pixel and texture rates. Its 195 W TDP with a 450 W PSU requirement makes it easier to integrate into existing systems than the GTX 580's 244 W TDP and 550 W PSU requirement. The GTX 680 supports PCIe 3.0, Vulkan 1.2.175, HDMI 1.4a, and DisplayPort 1.2, making it the better choice for users who need current API support or multiple display output types.

The GTX 580 wins in market positioning and consistency. Its 58th percentile ranking versus 55th for the GTX 680 means it performs more strongly relative to its contemporary peers. The GTX 580's nearest rival is the RTX 2060 with a 0% delta, showing it remains competitive with a much newer card. The GTX 580 also has a larger memory bus at 384-bit and more ROPs at 48, which can benefit certain memory-heavy workloads despite the nearly identical bandwidth.

For OpenCL compute tasks, the GTX 680 is the clear choice — it delivers 1,623 more points in the benchmark. For users who care about overall GPU ranking or who need Vulkan support, the GTX 680 again takes the lead. The GTX 580's only advantage lies in its higher percentile ranking and its closer proximity to modern rivals in the Geekbench OpenCL test. Users with a 550 W PSU and 8-pin connector can run the GTX 580, but the GTX 680 offers more performance with less power draw. The data supports the GTX 680 as the better overall card, with the GTX 580 representing a solid alternative for those prioritizing its specific strengths in ROP throughput and memory bus width.