GPU Comparison

# NVIDIA Quadro P2000 vs NVIDIA GeForce GTX 670MX

The NVIDIA Quadro P2000 and NVIDIA GeForce GTX 670MX represent two very different generations of NVIDIA mobile and workstation graphics. The P2000 is a Pascal-era workstation card from 2017, while the GTX 670MX is a Kepler-era mobile GPU from 2012. Benchmark data shows the P2000 holding a commanding lead in every shared test, with a 228.6% advantage in Geekbench OpenCL and a 343.3% advantage in Geekbench Vulkan. The P2000 also carries a higher overall average benchmark score of 6049 versus 5721 for the GTX 670MX, placing it in the 35th percentile of all GPUs compared to the 33rd percentile for the older card.

FAQ

Q: Which GPU is faster in compute workloads?

A: The Quadro P2000 wins decisively. In Geekbench OpenCL, it scores 20125 versus 6125 for the GTX 670MX, a 228.6% difference. In Geekbench Vulkan, the P2000 scores 23566 versus 5316, a 343.3% gap.

Q: How do their memory configurations compare?

A: The P2000 has 5 GB of GDDR5 on a 160-bit bus with 140.2 GB/s bandwidth. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus but only 67.20 GB/s bandwidth. Despite the narrower bus, the P2000 delivers over double the memory bandwidth.

Q: Are these cards still relevant for modern workloads?

A: Both are end-of-life products. The P2000 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while the GTX 670MX supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The P2000 has more up-to-date API support.

Q: What is the power draw difference?

A: Both cards are rated at 75 W TDP. The P2000 has a suggested PSU of 250 W and requires no power connectors. The GTX 670MX also uses no power connectors, but no suggested PSU is listed.

Q: Which card has better pixel and texture throughput?

A: The P2000 dominates: 59.20 GPixel/s pixel rate and 94.72 GTexel/s texture rate versus 12.02 GPixel/s and 48.08 GTexel/s for the GTX 670MX. The P2000 is nearly 5x faster in pixel fill and about 2x faster in texture fill.

Q: How do their transistor counts and die sizes compare?

A: The P2000 uses a GP106 chip with 4,400 million transistors on a 200 mm² die at 16 nm. The GTX 670MX uses GK104 with 3,540 million transistors on a larger 294 mm² die at 28 nm. The P2000 achieves 22.0M transistors per mm² versus 12.0M for the older card.

Architecture Differences

The P2000 is built on NVIDIA's Pascal architecture using the GP106 chip, manufactured by TSMC on a 16 nm process. This represents a major node shrink from the 28 nm process used for the GTX 670MX's GK104 Kepler chip. The P2000 packs 4,400 million transistors into a 200 mm² die, achieving a transistor density of 22.0M per mm². The GTX 670MX has fewer transistors (3,540 million) spread over a larger die (294 mm²), resulting in a density of only 12.0M per mm².

Clock speeds tell a similar story of generational progress. The P2000 runs at a 1076 MHz base clock with a 1480 MHz boost, while the GTX 670MX is locked at a flat 601 MHz for both base and boost. Memory clocks also differ substantially: the P2000 uses 1752 MHz memory with 7 Gbps effective speed, whereas the GTX 670MX runs at 700 MHz with 2.8 Gbps effective.

Feature-wise, the P2000 supports DirectX 12 (12_1) while the GTX 670MX is limited to DirectX 12 (11_0). Vulkan support differs too: the P2000 lists Vulkan 1.4 versus 1.2.175 for the GTX 670MX. Both support OpenGL 4.6. The P2000 offers 4x DisplayPort 1.4a outputs, while the GTX 670MX's display outputs are listed as "Portable Device Dependent," reflecting its laptop-oriented design.

The Verdict

The data is unambiguous: the Quadro P2000 is the superior GPU in every measurable way. It wins both head-to-head benchmark tests, with a 228.6% advantage in OpenCL and a 343.3% advantage in Vulkan. Its average benchmark score of 6049 is 5.7% higher than the GTX 670MX's 5721, and it sits at a higher percentile rank (35th versus 33rd).

For anyone choosing between these two for a workstation or professional application, the P2000 is the clear pick. It offers modern architecture, better API support, twice the memory bandwidth, and dramatically higher compute performance. The GTX 670MX, by contrast, is a 2012-era mobile part that has been surpassed in nearly every specification. Its only advantages are a wider memory bus (192-bit versus 160-bit) and more texture mapping units (80 versus 64), but these do not translate into wins in any benchmark.

The P2000's 5 GB memory capacity also gives it a practical edge over the 3 GB GTX 670MX for larger datasets. While both cards are end-of-life, the P2000 is the one with enough modern features to remain useful in lighter professional workloads.

Specification Differences

The two GPUs differ across nearly every specification category:

• Process Node: P2000 at 16 nm versus GTX 670MX at 28 nm (both TSMC)
• Transistors: 4,400 million versus 3,540 million
• Die Size: 200 mm² versus 294 mm²
• Transistor Density: 22.0M / mm² versus 12.0M / mm²
• Base Clock: 1076 MHz versus 601 MHz
• Boost Clock: 1480 MHz versus 601 MHz
• Memory Clock: 1752 MHz (7 Gbps effective) versus 700 MHz (2.8 Gbps effective)
• Memory Size: 5 GB versus 3 GB
• Memory Bus: 160-bit versus 192-bit
• Memory Bandwidth: 140.2 GB/s versus 67.20 GB/s
• Shading Units: 1024 versus 960
• TMUs: 64 versus 80
• ROPs: 40 versus 24
• Pixel Rate: 59.20 GPixel/s versus 12.02 GPixel/s
• Texture Rate: 94.72 GTexel/s versus 48.08 GTexel/s
• FP32: 3.031 TFLOPS versus 1,153.9 GFLOPS
• FP16: 47.36 GFLOPS (1:64) versus null
• DirectX Support: 12 (12_1) versus 12 (11_0)
• Vulkan Support: 1.4 versus 1.2.175
• Display Outputs: 4x DisplayPort 1.4a versus Portable Device Dependent
• Slot Width: Single-slot versus null
• Dimensions: 196 mm x 111 mm versus null

Both share the same TDP (75 W), use no power connectors, and have PCIe 3.0 x16 interfaces.

Head-to-Head Benchmarks

The head-to-head results are lopsided. In Geekbench OpenCL, the P2000 scores 20125 against 6125 for the GTX 670MX, a delta of 228.6%. This means the P2000 delivers more than three times the OpenCL compute performance. For GPU compute tasks such as rendering, simulation, or data processing, this is a massive gap.

Geekbench Vulkan shows an even larger disparity. The P2000 scores 23566 versus 5316, a delta of 343.3%. That is over four times the Vulkan performance. Vulkan is increasingly used in both professional visualization and modern game engines, so this advantage has practical implications for any workload leveraging that API.

The P2000 wins both head-to-head tests, giving it a 2-0 record. The GTX 670MX has zero wins. The average benchmark scores reinforce this: 6049 for the P2000 versus 5721 for the GTX 670MX. The P2000's nearest rivals are the NVIDIA GeForce MX230 (6077, -0.5%), NVIDIA RTX A400 (6078, -0.5%), AMD Radeon 760M (6019, +0.5%), and AMD Radeon RX 6400 (6001, +0.8%). The GTX 670MX's nearest rivals are the Intel Iris Pro Graphics P6300 (5712, +0.2%), NVIDIA GeForce GTX 550 Ti (5731, -0.2%), AMD Radeon HD 8790M (5691, +0.5%), and NVIDIA Quadro M500M (5604, +2.1%).

Where Each One Wins

The Quadro P2000 wins in every category that matters for modern workloads. Compute performance is its strongest suit, with the OpenCL and Vulkan scores showing a 3-4x advantage. Memory bandwidth is another clear win: 140.2 GB/s versus 67.20 GB/s means the P2000 can feed its shaders and texture units much more effectively. The higher pixel rate (59.20 GPixel/s versus 12.02 GPixel/s) and texture rate (94.72 GTexel/s versus 48.08 GTexel/s) make it substantially better for any fill-rate-bound tasks.

The P2000's 5 GB memory capacity versus 3 GB gives it more headroom for larger textures, bigger datasets, or multiple displays. Its DisplayPort 1.4a outputs support modern monitors, whereas the GTX 670MX is tied to whatever ports the laptop manufacturer included. The Pascal architecture also brings better power efficiency, delivering far more performance within the same 75 W TDP envelope.

The GTX 670MX has two specification-level advantages: a wider 192-bit memory bus (versus 160-bit) and more TMUs (80 versus 64). However, neither translates to a benchmark win. The wider bus is paired with much slower memory clocks, resulting in less than half the bandwidth. The extra TMUs are hamstrung by the low 601 MHz clock speed, producing a texture rate of only 48.08 GTexel/s against the P2000's 94.72 GTexel/s.

In practical terms, the P2000 is the choice for anyone needing workstation-grade compute, modern API support, or high-bandwidth memory. The GTX 670MX is best suited for legacy systems or applications where its specific mobile form factor is required, but even then, the performance gap is so large that it is hard to recommend. For professional use, the P2000's higher percentile rank (35th versus 33rd) and superior benchmark scores make it the only rational pick between the two.