NVIDIA Quadro GP100 vs NVIDIA Quadro P6000 Comparison

Head-to-Head Benchmarks

The single direct comparison available in the data is the Geekbench OpenCL test, and the result is decisive. The NVIDIA Quadro GP100 scores 87,445 points against the Quadro P6000's 66,382 points. That translates to a 31.7% advantage for the GP100 in this compute-oriented workload. This is not a marginal lead; it is a substantial gap that places the two cards in different performance tiers despite their shared Pascal architecture and launch date.

Context from the nearest rivals reinforces the strength of each card's position. The GP100's 87,445 average score puts it 0.4% ahead of the AMD Radeon PRO W7600 (87,108) and 2.1% ahead of the NVIDIA CMP 40HX (85,637). However, it trails the NVIDIA RTX A4500 by 4% (91,671) and the RTX A4500 Mobile by 4% (91,134). The P6000, meanwhile, sits at a lower overall level: its 66,382 OpenCL score and 69,986 average benchmark score place it just 0.2% ahead of the AMD Radeon Pro WX 8200 (69,870) and 1.4% behind the NVIDIA CMP 90HX (69,000). The P6000 also lags 1.2% behind the AMD Radeon RX 6600 LE (70,829) and 0.2% behind the NVIDIA RTX A3000 Mobile (70,140).

The GP100's win count in direct head-to-head benchmarks is 1, while the P6000 has 0 wins. The only test available is OpenCL, and the GP100 wins it outright. The P6000 does have a second benchmark entry — a Geekbench Vulkan score of 73,590 — but no corresponding GP100 Vulkan result exists in the data, so no direct comparison can be made on that API. The average benchmark score for the GP100 is 87,445, matching its single OpenCL score. The P6000's average is 69,986, pulled down by its lower OpenCL result. The percentile rankings tell a similar story: the GP100 sits at the 93rd percentile among all GPUs, while the P6000 sits at the 90th.

Architecture Differences

Both cards are built on the Pascal architecture, fabricated by TSMC on a 16 nm process node, and both share a transistor density of 25.1M per square millimeter. The similarities end there. The GP100 uses the GP100 chip, which packs 15,300 million transistors onto a 610 mm² die. The P6000 uses the GP102 chip, with 11,800 million transistors on a 471 mm² die. The GP100's chip is substantially larger and denser in absolute terms, which directly feeds its compute advantage.

Memory architecture is where the two diverge most sharply. The GP100 features 16 GB of HBM2 memory on a 4096-bit bus, yielding a massive 732.2 GB/s of bandwidth. The P6000 counters with 24 GB of GDDR5X on a 384-bit bus, delivering 432.8 GB/s. The GP100's bandwidth advantage is roughly 69% higher, a critical factor for memory-bound compute workloads. The P6000, however, offers 50% more capacity, which benefits large datasets that exceed the GP100's 16 GB frame buffer.

The compute unit counts are close but not identical. The GP100 has 3,584 shading units, 224 texture mapping units, and 96 ROPs. The P6000 has 3,840 shading units, 240 TMUs, and 96 ROPs. Despite having fewer shaders, the GP100 achieves a higher FP16 throughput: 20.69 TFLOPS at a 2:1 ratio, versus the P6000's 197.4 GFLOPS at a 1:64 ratio — a 100x difference. In FP32, however, the P6000 leads with 12.63 TFLOPS against the GP100's 10.34 TFLOPS. The P6000 also leads in pixel rate (157.9 GPixel/s vs 138.5 GPixel/s) and texture rate (394.8 GTexel/s vs 323.2 GTexel/s).

Clock speeds differ notably. The P6000 runs at a 1506 MHz base and 1645 MHz boost, while the GP100 runs at 1304 MHz base and 1443 MHz boost. Memory clocks also differ: the GP100's memory runs at 715 MHz (1430 Mbps effective), while the P6000's runs at 1127 MHz (9 Gbps effective). The P6000's higher clocks explain its lead in FP32, pixel, and texture throughput despite the GP100's larger die and bandwidth.

Both cards share identical display outputs (1x DVI, 4x DisplayPort 1.4a), bus interface (PCIe 3.0 x16), slot width (dual-slot), power connector (1x 8-pin), and dimensions (267 mm length, 111 mm height). Their API support is nearly identical, with both supporting DirectX 12 (12_1) and OpenGL 4.6. The P6000 supports Vulkan 1.4, while the GP100 supports Vulkan 1.3. Power draw differs slightly: the P6000 is rated at 250 W TDP with a 600 W suggested PSU, while the GP100 is rated at 235 W TDP with a 550 W suggested PSU.

The Verdict

The data points to a clear split in workloads. For raw compute performance, particularly in OpenCL, the GP100 is the superior card. Its 31.7% lead in the only head-to-head benchmark is decisive, and its FP16 throughput of 20.69 TFLOPS dwarfs the P6000's 197.4 GFLOPS. This makes the GP100 the obvious choice for compute-heavy tasks like machine learning inference, scientific simulation, or any workload that can leverage half-precision arithmetic. Its 732.2 GB/s memory bandwidth is also a massive advantage for memory-bound kernels.

The P6000, however, is not without its strengths. It offers 24 GB of memory versus the GP100's 16 GB, which is a 50% capacity increase. For workloads that require fitting entire datasets into VRAM — such as large 3D scenes, high-resolution texture sets, or complex visualizations — the P6000's extra capacity is a practical advantage. The P6000 also leads in FP32 throughput (12.63 TFLOPS vs 10.34 TFLOPS), pixel rate (157.9 GPixel/s vs 138.5 GPixel/s), and texture rate (394.8 GTexel/s vs 323.2 GTexel/s). For traditional graphics rendering, the P6000 is the stronger card on paper.

The overall percentile rankings favor the GP100 — 93rd percentile versus the P6000's 90th — and its average benchmark score of 87,445 is 24.9% higher than the P6000's 69,986. Yet the P6000's Vulkan score of 73,590 shows it has capable API-specific performance that the GP100 cannot be directly measured against in the provided data.

Specification Differences

• Chip: GP100 (GP100) vs GP102 (P6000)
• Transistors: 15,300 million vs 11,800 million
• Die size: 610 mm² vs 471 mm²
• Base clock: 1304 MHz vs 1506 MHz
• Boost clock: 1443 MHz vs 1645 MHz
• Memory size: 16 GB vs 24 GB
• Memory type: HBM2 vs GDDR5X
• Memory bus width: 4096 bit vs 384 bit
• Memory bandwidth: 732.2 GB/s vs 432.8 GB/s
• Memory clock: 715 MHz (1430 Mbps effective) vs 1127 MHz (9 Gbps effective)
• Shading units: 3584 vs 3840
• TMUs: 224 vs 240
• Pixel rate: 138.5 GPixel/s vs 157.9 GPixel/s
• Texture rate: 323.2 GTexel/s vs 394.8 GTexel/s
• FP32: 10.34 TFLOPS vs 12.63 TFLOPS
• FP16: 20.69 TFLOPS (2:1) vs 197.4 GFLOPS (1:64)
• TDP: 235 W vs 250 W
• Suggested PSU: 550 W vs 600 W
• Vulkan support: 1.3 vs 1.4
• Launch MSRP: The P6000 launched at 5,999 USD; the GP100 has no listed launch MSRP.

FAQ

Q: Which card is faster in OpenCL?

A: The NVIDIA Quadro GP100, with a score of 87,445 versus the P6000's 66,382 — a 31.7% advantage.

Q: Which card has more memory?

A: The NVIDIA Quadro P6000 has 24 GB of GDDR5X, while the GP100 has 16 GB of HBM2.

Q: Which card has higher memory bandwidth?

A: The GP100 leads with 732.2 GB/s over a 4096-bit bus, compared to the P6000's 432.8 GB/s over a 384-bit bus.

Q: Which card has better FP32 performance?

A: The P6000 leads in FP32 with 12.63 TFLOPS, while the GP100 delivers 10.34 TFLOPS.

Q: Which card has better FP16 performance?

A: The GP100 is dramatically ahead, with 20.69 TFLOPS at a 2:1 ratio, versus the P6000's 197.4 GFLOPS at a 1:64 ratio.

Q: Are these cards the same generation?

A: Yes, both are from the Quadro Pascal (Px000) generation, released on 2016-09-30, and both are end-of-life products.

Where Each One Wins

Choose the NVIDIA Quadro GP100 for: Compute-heavy workloads that rely on OpenCL, half-precision arithmetic, or massive memory bandwidth. Its 31.7% OpenCL lead over the P6000 is substantial, and its FP16 throughput of 20.69 TFLOPS is an order of magnitude higher. The 732.2 GB/s bandwidth is more than 69% higher than the P6000's, making it the better fit for data-intensive compute tasks. Its 235 W TDP also means lower power draw and a 550 W PSU requirement versus the P6000's 250 W and 600 W.

Choose the NVIDIA Quadro P6000 for: Graphics rendering and workloads that need more VRAM capacity. The 24 GB frame buffer is 50% larger than the GP100's 16 GB, which matters when loading large scenes or datasets that must fit entirely on the card. The P6000 also wins on FP32 (12.63 vs 10.34 TFLOPS), pixel rate (157.9 vs 138.5 GPixel/s), and texture rate (394.8 vs 323.2 GTexel/s). Its Vulkan score of 73,590 shows solid API-specific performance, and its launch MSRP was 5,999 USD, though no comparable price data exists for the GP100.