NVIDIA GeForce GTX 980 Ti vs NVIDIA P104-100 Comparison

The Verdict

The recorded data presents a clear split between these two NVIDIA cards. The NVIDIA P104-100 emerges as the overall leader in the database's average benchmark score, posting 32982 points against the GeForce GTX 980 Ti's 28020 points. That is a substantial 17.7% gap in aggregate performance. The P104-100 also holds a higher percentile ranking, sitting at the 77th percentile of all GPUs, while the GTX 980 Ti sits at the 73rd percentile.

However, the verdict is not a simple sweep. The GTX 980 Ti wins the Vulkan compute test, and it offers a larger memory pool. The P104-100 is a mining-focused card with no display outputs, which makes it entirely unsuitable for a desktop user who needs to connect a monitor. The GTX 980 Ti, with its full array of display outputs, is the only one of the two that can function as a traditional graphics card. The data suggests the P104-100 is the stronger raw compute performer, but the GTX 980 Ti is the more versatile piece of hardware. Buyers should choose the P104-100 only if they need pure compute throughput and have a separate display adapter. The GTX 980 Ti is the pick for anyone needing an actual, usable GPU.

Architecture Differences

These two cards come from different architectural generations. The P104-100 is built on the Pascal architecture using the GP104 chip, fabricated on a 16 nm process at TSMC. The GTX 980 Ti uses the Maxwell 2.0 architecture with the GM200 chip, manufactured on a 28 nm node, also at TSMC. The process node difference is stark: 16 nm versus 28 nm. This explains the transistor density figures. The P104-100 packs 7,200 million transistors into a 314 mm² die, yielding a density of 22.9M transistors per mm². The GTX 980 Ti has more transistors overall at 8,000 million, but they are spread across a much larger 601 mm² die, resulting in a density of just 13.3M / mm².

The memory subsystems also differ fundamentally. The P104-100 uses 4 GB of GDDR5X memory on a 256-bit bus, achieving 320.3 GB/s of bandwidth. The GTX 980 Ti uses 6 GB of GDDR5 on a wider 384-bit bus, achieving 336.6 GB/s. Despite the older memory type, the GTX 980 Ti has a slight bandwidth advantage. The P104-100 runs its memory at 1251 MHz (10 Gbps effective), while the GTX 980 Ti runs at 1753 MHz (7 Gbps effective). The P104-100 also has a unique oddity: it supports FP16 at 104.0 GFLOPS with a 1:64 ratio, while the GTX 980 Ti has no recorded FP16 capability.

Head-to-Head Benchmarks

The database records three direct benchmark comparisons. The first is the 3DMark Steel Nomad DX12 test. Here, the P104-100 scores 1413 against the GTX 980 Ti's 1321, a 7% win for the P104-100. This is a modest but clear victory in a modern DirectX 12 workload.

The second test is Geekbench OpenCL. This is the largest margin in the entire comparison. The P104-100 scores 52368, while the GTX 980 Ti scores 43513. That is a 20.4% advantage for the P104-100. This result indicates that in raw, general-purpose compute tasks, the Pascal card is significantly more efficient and faster.

The third test is Geekbench Vulkan, and this is where the GTX 980 Ti fights back. The GTX 980 Ti scores 47724, beating the P104-100's 45165. The delta is -5.4% from the P104-100's perspective, meaning the GTX 980 Ti wins by roughly 5.4%. This is intriguing: the older Maxwell architecture outperforms Pascal in this specific Vulkan compute workload. The data shows a 2-to-1 win tally in favor of the P104-100, but the Vulkan result proves the GTX 980 Ti is not obsolete in every compute scenario.

Specification Differences

The two cards differ in nearly every major specification category. The process node is a key divider: the P104-100 uses 16 nm, the GTX 980 Ti uses 28 nm. The chip sizes differ, with the P104-100 at 314 mm² and the GTX 980 Ti at 601 mm². Transistor counts are 7,200 million versus 8,000 million.

Clock speeds show the P104-100 running much higher. Its base clock is 1607 MHz with a boost of 1733 MHz. The GTX 980 Ti has a base clock of 1000 MHz and a boost of 1076 MHz. The memory clocks differ as well: 1251 MHz (10 Gbps effective) for the P104-100 versus 1753 MHz (7 Gbps effective) for the GTX 980 Ti.

The compute resources are split. The P104-100 has 1920 shading units, 120 TMUs, and 64 ROPs. The GTX 980 Ti has more of each: 2816 shading units, 176 TMUs, and 96 ROPs. Yet the P104-100 still achieves higher pixel and texture rates: 110.9 GPixel/s and 208.0 GTexel/s, versus 103.3 GPixel/s and 189.4 GTexel/s for the GTX 980 Ti. The FP32 throughput also favors the P104-100 at 6.655 TFLOPS versus 6.060 TFLOPS.

Power and connectivity diverge sharply. The GTX 980 Ti has a TDP of 250 W and requires a 600 W suggested PSU with a 1x 6-pin plus 1x 8-pin connector setup. The P104-100 has no recorded TDP, but its suggested PSU is only 200 W with a single 8-pin connector. The bus interface also differs: the P104-100 uses PCIe 1.0 x4, while the GTX 980 Ti uses PCIe 3.0 x16. This is a surprising limitation for the P104-100.

Display outputs are the most consequential difference. The P104-100 has no outputs at all. The GTX 980 Ti has 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. Dimensions are identical in length at 267 mm (10.5 inches), but the GTX 980 Ti is taller at 111 mm and wider at 40 mm, while those measurements are not recorded for the P104-100.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA P104-100, with an average score of 32982, is notably ahead of the GTX 980 Ti's 28020.

Q: Does the GTX 980 Ti win any benchmark test?

A: Yes. In the Geekbench Vulkan test, the GTX 980 Ti scores 47724, beating the P104-100's 45165 by 5.4%.

Q: Why can the GTX 980 Ti be used as a normal graphics card while the P104-100 cannot?

A: The GTX 980 Ti has display outputs (1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2). The P104-100 has no display outputs, making it unusable for direct monitor connection.

Q: Which card has more memory and wider memory bus?

A: The GTX 980 Ti has 6 GB of GDDR5 on a 384-bit bus. The P104-100 has 4 GB of GDDR5X on a 256-bit bus.

Q: What is the biggest performance gap between the two cards?

A: The Geekbench OpenCL test shows the P104-100 leading by 20.4%, scoring 52368 versus 43513.

Q: Which card sits in a higher performance percentile?

A: The P104-100 is at the 77th percentile of all GPUs, while the GTX 980 Ti is at the 73rd percentile.

Where Each One Wins

The P104-100 wins in raw compute throughput. Its 20.4% lead in Geekbench OpenCL demonstrates a significant advantage in general-purpose compute tasks. It also wins in DirectX 12 gaming workloads, as shown by its 7% lead in the 3DMark Steel Nomad test. The higher FP32 rate of 6.655 TFLOPS supports this. The P104-100 is also more power-frugal in terms of system requirements, needing only a 200 W suggested PSU compared to the GTX 980 Ti's 600 W. Its smaller die and denser transistor layout on the 16 nm node make it a more efficient design. For anyone running compute-heavy workloads that rely on OpenCL or DirectX 12, the P104-100 is the stronger choice, provided a separate display solution exists.

The GTX 980 Ti wins in specific scenarios. It takes the Vulkan compute test by 5.4%, showing that its Maxwell architecture still has strengths in certain API workloads. It offers 6 GB of memory versus 4 GB, which matters for larger datasets. Its memory bandwidth is also slightly higher at 336.6 GB/s versus 320.3 GB/s. The GTX 980 Ti has a proper PCIe 3.0 x16 interface, far superior to the P104-100's PCIe 1.0 x4. Most importantly, it has display outputs, making it the only card here that can drive a monitor. The GTX 980 Ti also has a recorded launch MSRP of 649 USD, which serves as a reference point for its original market positioning. The GTX 980 Ti is the winner for users who need a functional GPU with display connectivity, Vulkan compute performance, and a larger memory buffer.