NVIDIA CMP 90HX vs NVIDIA GeForce RTX 3090 Ti Comparison

Head-to-Head Benchmarks

The only directly comparable benchmark recorded in the database for both cards is Geekbench OpenCL, and the result is decisively one-sided. The NVIDIA GeForce RTX 3090 Ti scores 174,441 points, while the NVIDIA CMP 90HX scores 69,000 points. That is a delta of 152.8%, meaning the RTX 3090 Ti delivers roughly two and a half times the raw compute throughput in this workload. There is no benchmark in the database where the CMP 90HX comes out ahead; the win count stands at 1 for the RTX 3090 Ti and 0 for the CMP 90HX.

Context from the nearest rival lists reinforces how wide this gap is. The RTX 3090 Ti's average benchmark score sits at 131,938, placing it in the 95th percentile among all GPUs. Its closest rivals in the database are the NVIDIA L4 at 131,072 (0.7% slower), the NVIDIA RTX 4000 Ada Generation at 135,218 (2.4% faster), the NVIDIA A10M at 135,230 (2.4% faster), and the AMD Radeon PRO W6800 at 135,396 (2.6% faster). So the RTX 3090 Ti is essentially neck-and-neck with a cluster of professional workstation cards, all within a few percentage points of each other. The CMP 90HX, by contrast, has an average benchmark score of 69,000, placing it in the 90th percentile. Its nearest rivals are the Intel Arc A770 at 68,809 (0.3% slower), the AMD Radeon Instinct MI25 at 68,562 (0.6% slower), the AMD Radeon Pro WX 8200 at 69,870 (1.2% faster), and the NVIDIA Quadro P6000 at 69,986 (1.4% faster). The CMP 90HX is competitive with those cards, but that entire group sits roughly half the performance level of the RTX 3090 Ti's peer group.

Looking at the individual specifications that drive this result, the RTX 3090 Ti has 10,752 shading units, 336 texture mapping units, and 112 render output units. The CMP 90HX has 6,400 shading units, 200 TMUs, and 80 ROPs. The RTX 3090 Ti also carries 84 RT cores and 336 tensor cores, versus 50 RT cores and 200 tensor cores on the CMP 90HX. In FP32 compute, the RTX 3090 Ti reaches 40.00 TFLOPS, while the CMP 90HX manages 21.89 TFLOPS. The FP16 figures mirror this exactly: 40.00 TFLOPS for the RTX 3090 Ti and 21.89 TFLOPS for the CMP 90HX, both at a 1:1 ratio. These are not subtle differences; they represent a fundamental gap in execution resources.

The memory subsystem tells a similar story. The RTX 3090 Ti has 24 GB of GDDR6X on a 384-bit bus, producing 1.01 TB/s of bandwidth. The CMP 90HX has 10 GB of GDDR6X on a 320-bit bus, producing 760.3 GB/s. That is a 33% bandwidth advantage for the RTX 3090 Ti, alongside more than double the capacity. The pixel rate for the RTX 3090 Ti is 208.3 GPixel/s versus 136.8 GPixel/s for the CMP 90HX, a 52% lead. The texture rate is 625.0 GTexel/s versus 342.0 GTexel/s, a 83% lead. Every measurable throughput metric in the database favors the RTX 3090 Ti by a wide margin.

The Verdict

The data supports only one conclusion for general-purpose compute and graphics workloads: the NVIDIA GeForce RTX 3090 Ti is the superior card by a substantial margin. Its Geekbench OpenCL score is 152.8% higher than the CMP 90HX, its FP32 throughput is nearly double, its memory bandwidth is a third higher, and its render output and texture rates are far ahead. The RTX 3090 Ti also sits at the 95th percentile among all GPUs, while the CMP 90HX sits at the 90th percentile, but that percentile gap understates the actual score difference because the CMP 90HX's nearest rivals are clustered around the 69,000 mark, a full tier below the RTX 3090 Ti's rivals near 135,000.

The CMP 90HX is not without merit in its own context. It is a dual-slot card with a 320 W TDP, compared to the RTX 3090 Ti's triple-slot design and 450 W TDP. The CMP 90HX also uses two 8-pin power connectors and recommends a 700 W power supply, whereas the RTX 3090 Ti uses a single 16-pin connector and recommends 850 W. For a system builder prioritizing physical footprint and power draw over raw performance, the CMP 90HX is the lighter option. However, there is no benchmark in the database where the CMP 90HX wins, and its only recorded score is less than half of the RTX 3090 Ti's score in the same test.

The RTX 3090 Ti is also the only one of the two with display outputs. It offers 1x HDMI 2.1 and 3x DisplayPort 1.4a, while the CMP 90HX has no outputs at all. That makes the CMP 90HX unsuitable for any interactive or display-driven task, regardless of its compute capabilities. The RTX 3090 Ti is a complete product in that sense; the CMP 90HX is a specialized compute-only device.

Where Each One Wins

The NVIDIA GeForce RTX 3090 Ti wins in every category where the database records a comparison. For compute-heavy workloads such as OpenCL acceleration, the RTX 3090 Ti's 174,441 score versus 69,000 is the decisive evidence. For graphics rendering, the RTX 3090 Ti's 112 ROPs and 208.3 GPixel/s pixel rate, alongside 336 TMUs and 625.0 GTexel/s texture rate, make it far more capable at rasterization. Its 84 RT cores and 336 tensor cores also provide hardware acceleration for ray tracing and AI inference that the CMP 90HX's 50 RT cores and 200 tensor cores cannot match in volume.

The CMP 90HX wins in efficiency-related dimensions that are not directly benchmarked but are recorded in the specification data. It draws 320 W versus 450 W, fits in a dual-slot form factor versus triple-slot, and is shorter at 285 mm versus 336 mm. It also uses PCIe 1.0 x4 as its bus interface, which is a legacy limitation, but that is offset by its lower power requirements. For a mining-focused or headless compute deployment where power density and physical space are the primary constraints, the CMP 90HX offers a viable alternative. But the performance delta is so large that any workload sensitive to compute throughput will strongly prefer the RTX 3090 Ti.

FAQ

Q: Which card has a higher Geekbench OpenCL score?

A: The NVIDIA GeForce RTX 3090 Ti scores 174,441, while the NVIDIA CMP 90HX scores 69,000. The RTX 3090 Ti leads by 152.8%.

Q: What is the memory capacity difference?

A: The RTX 3090 Ti has 24 GB of GDDR6X memory, while the CMP 90HX has 10 GB. The RTX 3090 Ti also has a wider 384-bit bus versus 320-bit, and higher bandwidth at 1.01 TB/s versus 760.3 GB/s.

Q: Do both cards support the same APIs?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: Does the CMP 90HX have any display outputs?

A: No, the CMP 90HX has no display outputs. The RTX 3090 Ti has 1x HDMI 2.1 and 3x DisplayPort 1.4a.

Q: How do their power requirements differ?

A: The RTX 3090 Ti has a 450 W TDP and recommends an 850 W power supply, using a single 16-pin connector. The CMP 90HX has a 320 W TDP and recommends a 700 W power supply, using two 8-pin connectors.

Q: Which card has a higher FP32 compute rating?

A: The RTX 3090 Ti is rated at 40.00 TFLOPS, while the CMP 90HX is rated at 21.89 TFLOPS. The FP16 ratings are identical to the FP32 ratings for both cards at a 1:1 ratio.

Architecture Differences

Both cards are built on the NVIDIA Ampere architecture using the GA102 chip, fabricated on an 8 nm process at Samsung. They share the same transistor count of 28,300 million and the same die size of 628 mm², with a transistor density of 45.1M per mm². The underlying silicon is identical; the differences come from how the chip is configured and the surrounding board design.

The RTX 3090 Ti is a fully enabled configuration with 10,752 shading units, 336 TMUs, 112 ROPs, 84 RT cores, and 336 tensor cores. The CMP 90HX is a cut-down configuration with 6,400 shading units, 200 TMUs, 80 ROPs, 50 RT cores, and 200 tensor cores. This means the CMP 90HX disables roughly 40% of the shader cores and 40% of the tensor cores relative to the RTX 3090 Ti, along with a smaller reduction in ROPs and TMUs.

Clock speeds also differ. The RTX 3090 Ti runs at a base clock of 1560 MHz and a boost clock of 1860 MHz. The CMP 90HX runs at 1500 MHz base and 1710 MHz boost. Memory clocks differ as well: the RTX 3090 Ti operates at 1313 MHz with 21 Gbps effective data rate, while the CMP 90HX operates at 1188 MHz with 19 Gbps effective. These clock differences compound the core count disparity, pushing the RTX 3090 Ti's FP32 throughput to 40.00 TFLOPS versus 21.89 TFLOPS for the CMP 90HX.

The bus interface is another major architectural divergence. The RTX 3090 Ti uses PCIe 4.0 x16, while the CMP 90HX uses PCIe 1.0 x4. That is a legacy interface that severely limits host-to-device data transfer rates, which is consistent with the CMP 90HX's intended use case as a compute-only card for mining, where data transfers are minimal after initial workload setup. The RTX 3090 Ti's full PCIe 4.0 x16 interface allows for much faster data exchange with the host system.

Both cards support the same API set: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Neither card has a listed codename in the database. The RTX 3090 Ti is part of the GeForce 30-series generation, while the CMP 90HX is listed under the generation "Mining GPUs." The RTX 3090 Ti has a predecessor in GeForce 20 and a successor in GeForce 40; the CMP 90HX has neither.

Specification Differences

The following fields differ between the two cards in the database:

• Shading Units: 10,752 (RTX 3090 Ti) vs 6,400 (CMP 90HX)
• TMUs: 336 vs 200
• ROPs: 112 vs 80
• RT Cores: 84 vs 50
• Tensor Cores: 336 vs 200
• Base Clock: 1560 MHz vs 1500 MHz
• Boost Clock: 1860 MHz vs 1710 MHz
• Memory Clock: 1313 MHz / 21 Gbps effective vs 1188 MHz / 19 Gbps effective
• Memory Size: 24 GB vs 10 GB
• Memory Bus Width: 384 bit vs 320 bit
• Memory Bandwidth: 1.01 TB/s vs 760.3 GB/s
• Pixel Rate: 208.3 GPixel/s vs 136.8 GPixel/s
• Texture Rate: 625.0 GTexel/s vs 342.0 GTexel/s
• FP32 Compute: 40.00 TFLOPS vs 21.89 TFLOPS
• FP16 Compute: 40.00 TFLOPS vs 21.89 TFLOPS
• TDP: 450 W vs 320 W
• Slot Width: Triple-slot vs Dual-slot
• Power Connectors: 1x 16-pin vs 2x 8-pin
• Suggested PSU: 850 W vs 700 W
• Bus Interface: PCIe 4.0 x16 vs PCIe 1.0 x4
• Display Outputs: 1x HDMI 2.1, 3x DisplayPort 1.4a vs No outputs
• Dimensions (Length): 336 mm (13.2 inches) vs 285 mm (11.2 inches)
• Dimensions (Height): 140 mm (5.5 inches) vs 112 mm (4.4 inches)
• Dimensions (Width): 61 mm (2.4 inches) vs not listed
• Release Date: 2022-01-26 vs 2021-07-27
• Generation: GeForce 30 vs Mining GPUs
• Launch MSRP: 1,999 USD vs not listed

Fields that are identical include the chip (GA102), architecture (Ampere), process node (8 nm), foundry (Samsung), transistor count (28,300 million), die size (628 mm²), transistor density (45.1M per mm²), API support, and production status (End-of-life). The CMP 90HX also lacks a listed predecessor, successor, and launch MSRP, while the RTX 3090 Ti has all three.