Intel Arc Pro B60 Dual vs NVIDIA GeForce RTX 5090 SE Comparison

Head-to-Head Benchmarks

The recorded data shows no direct benchmark scores for either the Intel Arc Pro B60 Dual or the NVIDIA GeForce RTX 5090 SE. Both entries carry an average benchmark score of zero, and the head-to-head benchmark table is empty. Consequently, the wins tally reads zero for each part, and no percentile differentiation exists beyond a shared 50th percentile placement among all GPUs in the database.

What the database does provide is a clear set of theoretical peak figures that define the performance envelope of each card. In raw FP32 throughput, the RTX 5090 SE delivers 66.94 TFLOPS against the Arc Pro B60 Dual's 12.29 TFLOPS. That is a 5.45x gap in favor of NVIDIA. The FP16 comparison follows a different pattern: the Intel card reaches 24.58 TFLOPS using a 2:1 ratio, while the RTX 5090 SE sustains 66.94 TFLOPS at 1:1. NVIDIA's advantage remains substantial but narrows to 2.72x in this metric.

Texture and pixel rates reinforce the same hierarchy. The RTX 5090 SE posts 1,045.9 GTexel/s and 380.3 GPixel/s, while the Arc Pro B60 Dual manages 384.0 GTexel/s and 192.0 GPixel/s. The NVIDIA card leads by 2.72x in texturing and 1.98x in pixel fill. Memory bandwidth tells a similar story: 1.34 TB/s versus 456.0 GB/s, a 2.94x difference. The Intel card's GDDR6 memory runs at 19 Gbps effective across a 192-bit bus, while the RTX 5090 SE uses GDDR7 at 28 Gbps effective over a 384-bit bus.

The one area where the Arc Pro B60 Dual shows a nominal clock advantage is base frequency. Intel lists 2000 MHz base versus NVIDIA's 1740 MHz, a 260 MHz difference. Boost clocks are closer: 2400 MHz for Intel, 2377 MHz for NVIDIA. These clock figures do not compensate for the massive difference in shading units, 2560 versus 14080, nor the TMU and ROP counts of 160 and 80 against 440 and 160. The RTX 5090 SE also fields 110 RT cores and 440 tensor cores, while the Intel part lists 20 RT cores and no tensor core entry.

Architecture Differences

The two GPUs come from different architectural lineages. Intel uses the Xe2-HPG architecture on the BMG-G21 chip, part of the Battlemage Pro Series. NVIDIA counters with Blackwell 2.0 on the GB202 die, belonging to the GeForce 50 generation. Both are fabricated by TSMC on a 5 nm process node, so the manufacturing process is identical. The die sizes, however, diverge sharply: Intel's chip measures 272 mm², NVIDIA's 750 mm². Transistor counts follow suit, with 19,600 million on the Intel die and 92,200 million on the NVIDIA die. Transistor density per square millimeter is 72.1M for Intel and 122.9M for NVIDIA, indicating the Blackwell die packs logic more tightly.

Memory architecture differs in type and bus width. The Arc Pro B60 Dual uses 24 GB of GDDR6 on a 192-bit interface. The RTX 5090 SE also carries 24 GB, but in GDDR7 form across a 384-bit bus. The resulting bandwidth gap, 456.0 GB/s versus 1.34 TB/s, stems from both the wider bus and the faster memory standard. Effective memory clocks are 19 Gbps for Intel and 28 Gbps for NVIDIA.

The interface and power delivery also separate the two. Intel's card connects via PCIe 5.0 x8, while NVIDIA uses PCIe 5.0 x16. Both are dual-slot designs with a single 16-pin power connector. The Intel card has a 400 W TDP and an 800 W suggested power supply, whereas the RTX 5090 SE lists a 500 W TDP and 900 W suggested PSU. Physical dimensions favor NVIDIA on length: 267 mm versus 300 mm. Height and width are nearly identical, 110 mm versus 111 mm and 40 mm for both. Display outputs differ as well: Intel provides four mini-DisplayPort 2.1 connectors, while NVIDIA offers one HDMI 2.1b and three DisplayPort 2.1b ports.

API support is identical on paper. Both cards claim DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The feature sets for ray tracing and compute are not directly comparable from the database, but the RT core counts, 20 versus 110, and tensor core presence on the NVIDIA side indicate different design priorities. Intel lists no tensor core specification, while NVIDIA's 440 tensor cores align with its Blackwell compute architecture.

The Verdict

The database presents a straightforward conclusion: the NVIDIA GeForce RTX 5090 SE dominates the Intel Arc Pro B60 Dual across every measured computational metric. FP32 performance is 5.45x higher, FP16 is 2.72x higher, texture rate is 2.72x higher, pixel rate is 1.98x higher, and memory bandwidth is 2.94x higher. The NVIDIA card also carries 5.5x more shading units, 2.75x more TMUs, 2x more ROPs, and 5.5x more RT cores.

The Intel card's advantages are limited to a higher base clock, 2000 MHz versus 1740 MHz, a shorter release-to-market timeline, and a lower launch MSRP of 1,199 USD against NVIDIA's 1,499 USD. The 300 USD difference, however, does not offset the performance gap in any benchmark-relevant sense, as no actual benchmark scores exist to weigh price against performance.

For users selecting between these two based on the recorded data, the RTX 5090 SE is the clear choice for any workload that stresses FP32 compute, memory bandwidth, or ray tracing. The Arc Pro B60 Dual, with its 24 GB memory and 400 W TDP, may suffice for tasks where the lower power draw and smaller physical footprint matter, but the performance data does not support a case for parity.

FAQ

Q: Which GPU has higher FP32 performance?

A: The NVIDIA GeForce RTX 5090 SE delivers 66.94 TFLOPS, which is 5.45x the Intel Arc Pro B60 Dual's 12.29 TFLOPS.

Q: Do both cards have the same memory capacity?

A: Yes, both have 24 GB. The Intel card uses GDDR6 on a 192-bit bus with 456.0 GB/s bandwidth, while the NVIDIA card uses GDDR7 on a 384-bit bus with 1.34 TB/s bandwidth.

Q: What is the difference in RT core counts?

A: The RTX 5090 SE has 110 RT cores, while the Arc Pro B60 Dual has 20 RT cores. NVIDIA also has 440 tensor cores, while Intel lists none.

Q: Are the power requirements different?

A: The Intel card has a 400 W TDP and suggests an 800 W PSU. The NVIDIA card has a 500 W TDP and suggests a 900 W PSU. Both use a single 16-pin power connector.

Q: What are the launch prices?

A: The Intel Arc Pro B60 Dual has a launch MSRP of 1,199 USD. The NVIDIA GeForce RTX 5090 SE has a launch MSRP of 1,499 USD.

Q: Which GPU has more display outputs?

A: The Intel card has four mini-DisplayPort 2.1 outputs. The NVIDIA card has one HDMI 2.1b and three DisplayPort 2.1b outputs.

Where Each One Wins

The NVIDIA GeForce RTX 5090 SE wins decisively in compute-heavy workloads. Its 66.94 TFLOPS FP32 and 66.94 TFLOPS FP16 (1:1) figures make it suitable for general-purpose GPU compute, AI inference via its 440 tensor cores, and high-resolution rendering. The 1.34 TB/s memory bandwidth supports large data sets and high-resolution textures. The 110 RT cores provide a substantial ray tracing advantage over Intel's 20 RT cores. The 440 TMUs and 160 ROPs drive high fill rates, and the PCIe 5.0 x16 interface offers double the bandwidth of Intel's x8 connection.

The Intel Arc Pro B60 Dual wins in a narrow set of criteria. It has a higher base clock, 2000 MHz versus 1740 MHz, which may benefit lightly threaded workloads that scale with clock frequency. Its 400 W TDP is 100 W lower than NVIDIA's 500 W, and its 800 W suggested PSU is 100 W lower, potentially easing system integration requirements. The card is longer at 300 mm versus 267 mm, but that is not an advantage. The four mini-DisplayPort 2.1 outputs exceed NVIDIA's three DisplayPort ports, which could matter for multi-display setups.

For FP16 workloads, the Intel card's 2:1 ratio means it halves its FP32 throughput to reach 24.58 TFLOPS, whereas NVIDIA maintains full FP32 rates at 1:1. That distinction favors NVIDIA for any mixed-precision task. The transistor density figures, 122.9M per mm² for NVIDIA versus 72.1M for Intel, indicate a more complex design on the Blackwell die, though both share the same 5 nm TSMC process.

In summary, the RTX 5090 SE is the performance leader across all substantive metrics. The Arc Pro B60 Dual offers a lower power envelope, a higher base clock, and more mini-DisplayPort outputs, but the recorded data shows no benchmark scenario where it closes the gap in raw throughput.