AMD Instinct MI350P vs Intel Arc B580 Comparison

Head-to-Head Benchmarks

The AMD Instinct MI350P and Intel Arc B580 occupy entirely different corners of the GPU market, and the recorded data reflects that divide clearly. The MI350P is a compute-oriented accelerator with no display outputs and no benchmark scores listed in the database. Its average benchmark score is recorded as zero, and it holds a 50th percentile position among all GPUs. The Arc B580, by contrast, is a consumer graphics card with a 68th percentile ranking and an average benchmark score of 23,021 across ten recorded tests. This is not a head-to-head comparison in the traditional sense; the MI350P has no gaming or graphics benchmarks to set against the B580's suite of DirectX and compute tests.

The Intel Arc B580 delivers its strongest results in synthetic compute workloads. Its Geekbench Vulkan score of 109,672 stands as the highest single benchmark in its recorded set, followed by a Geekbench OpenCL score of 92,821. These figures indicate respectable asynchronous compute performance for a consumer card. The PassMark G3D score of 15,748 and PassMark GPU Compute score of 7,729 further confirm that the B580 handles both graphics and compute workloads with measurable competence. In the DirectX suite, the B580 scores 128 in PassMark DirectX 11, 76 in DirectX 12, 76 in DirectX 10, and 183 in DirectX 9. The PassMark G2D score of 709 rounds out the picture, showing modest 2D performance compared to its 3D results.

When placed against its nearest rivals in the database, the Arc B580's average score of 23,021 is essentially tied with the AMD Radeon RX 580 2048SP, which posts an average score of 23,061, a delta of negative 0.2 percent. The NVIDIA GeForce RTX 2080 comes in at 22,895, putting the B580 0.6 percent ahead. The NVIDIA GeForce RTX 3080 scores 23,172, meaning the B580 trails by 0.7 percent. The NVIDIA P106-100 leads the group with 23,249, leaving the B580 1 percent behind. These deltas are small, all within a single percentage point, indicating that the B580 slots into a performance band alongside older flagship cards from the previous generation.

The MI350P, with no benchmark entries, cannot be compared on the same scale. Its 50th percentile ranking versus the B580's 68th percentile suggests that the database places the MI350P below the B580 in overall standing, but this is a consequence of missing test data rather than a direct performance verdict. The MI350P's raw specifications tell a different story. Its FP32 throughput of 36.04 TFLOPS is 2.6 times the B580's 13.67 TFLOPS. Its FP16 output of 36.04 TFLOPS (1:1 ratio) exceeds the B580's 27.34 TFLOPS (2:1 ratio) by roughly 32 percent. Texture rate favors the MI350P at 1,126.4 GTexel/s versus 427.2 GTexel/s, a 2.6 times advantage. The MI350P's memory bandwidth of 8.19 TB/s dwarfs the B580's 456.0 GB/s, an 18-fold difference.

Architecture Differences

The two cards are built on fundamentally different architectures. The AMD Instinct MI350P uses CDNA 4.0, a compute-focused design with no graphics API support. The database lists its DirectX, OpenGL, and Vulkan support as N/A, and its pixel rate is recorded as 0 MPixel/s. It has no display outputs. The Intel Arc B580 uses Xe2-HPG, the Battlemage generation architecture, with full DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4 support. It provides 1x HDMI 2.1a and 3x DisplayPort 2.1 outputs.

Process technology separates the two as well. The MI350P is built on a 3 nm node at TSMC, while the B580 uses a 5 nm node, also at TSMC. The MI350P packs 73,000 million transistors onto a 1,190 mm² die, yielding a transistor density of 61.3 million per square millimeter. The B580 contains 19,600 million transistors on a 272 mm² die, with a higher density of 72.1 million per square millimeter. The B580's smaller die achieves greater packing efficiency, though the MI350P's massive die size reflects its HBM3e memory stack integration.

The MI350P's chip is designated MI350 128CU, indicating 128 compute units. It carries 8,192 shading units and 512 texture mapping units, with no ROPs recorded. The B580's BMG-G21 chip includes 2,560 shading units, 160 TMUs, 80 ROPs, and 20 ray tracing cores. The MI350P has no RT cores listed. The B580's pixel rate of 213.6 GPixel/s demonstrates its rasterization capability, a feature the MI350P lacks entirely by design.

Memory architecture differs sharply. The MI350P uses 144 GB of HBM3e on an 8,192-bit bus, delivering 8.19 TB/s of bandwidth. The B580 uses 12 GB of GDDR6 on a 192-bit bus, delivering 456.0 GB/s. The MI350P's memory clock is listed as 2000 MHz (8 Gbps effective), while the B580 runs at 2375 MHz (19 Gbps effective). The MI350P's base clock is 1000 MHz with a boost of 2200 MHz. The B580's base and boost clocks are both 2670 MHz, a single fixed clock rate.

Power delivery and interface requirements diverge as well. The MI350P has a TDP of 600 W, uses a single 16-pin power connector, and requires a 1000 W suggested PSU. The B580 has a TDP of 190 W, uses a single 8-pin connector, and calls for a 450 W suggested PSU. The MI350P connects via PCIe 5.0 x16, while the B580 uses PCIe 4.0 x8.

The Verdict

The data points to two different purchasing decisions. The Intel Arc B580 is a complete consumer graphics card. It has display outputs, full API support, a 68th percentile ranking, and an average benchmark score of 23,021. It competes within one percentage point of the RTX 2080, RTX 3080, RX 580 2048SP, and P106-100 in the database's nearest rival group. Its 12 GB of GDDR6 memory and 190 W TDP make it suitable for systems with modest power budgets. The B580's launch MSRP is 249 USD.

The AMD Instinct MI350P is not a consumer card. It has no display outputs, no graphics API support, and no benchmark scores in the database. Its 600 W TDP, 1000 W suggested PSU, and single 16-pin connector target server and datacenter environments. Its 144 GB of HBM3e memory and 8.19 TB/s bandwidth are aimed at large-scale compute workloads, not desktop gaming or workstation graphics. The MI350P's 50th percentile ranking reflects the absence of recorded benchmarks rather than measured performance.

The B580 is the choice for anyone building a system that needs to render frames, run DirectX or Vulkan applications, or output video through HDMI or DisplayPort. The MI350P is the choice for compute tasks that can leverage its FP32 and FP16 throughput, massive memory pool, and extreme bandwidth, provided the host system can supply 600 W and support a PCIe 5.0 x16 card with no video output. The two cards do not overlap in function, and the database records no head-to-head benchmarks between them.

Specification Differences

The MI350P and B580 differ in nearly every measurable specification. The MI350P uses a 3 nm process, while the B580 uses 5 nm. Transistor count is 73,000 million for the MI350P versus 19,600 million for the B580. Die size is 1,190 mm² versus 272 mm². Transistor density is 61.3M per mm² versus 72.1M per mm². Base clock is 1000 MHz versus 2670 MHz. Boost clock is 2200 MHz versus 2670 MHz. Memory size is 144 GB versus 12 GB. Memory type is HBM3e versus GDDR6. Bus width is 8,192 bit versus 192 bit. Bandwidth is 8.19 TB/s versus 456.0 GB/s. Shading units are 8,192 versus 2,560. TMUs are 512 versus 160. ROPs are 0 versus 80. RT cores are absent versus 20. Pixel rate is 0 MPixel/s versus 213.6 GPixel/s. Texture rate is 1,126.4 GTexel/s versus 427.2 GTexel/s. FP32 is 36.04 TFLOPS versus 13.67 TFLOPS. FP16 is 36.04 TFLOPS (1:1) versus 27.34 TFLOPS (2:1). TDP is 600 W versus 190 W. Power connectors are 1x 16-pin versus 1x 8-pin. Suggested PSU is 1000 W versus 450 W. Bus interface is PCIe 5.0 x16 versus PCIe 4.0 x8. Display outputs are none versus 1x HDMI 2.1a and 3x DisplayPort 2.1. API support is N/A versus DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4. Dimensions are 267 mm by 111 mm by 40 mm versus 272 mm by 115 mm by 45 mm. Release dates are 2026-05-06 for the MI350P and 2024-12-12 for the B580.

FAQ

Q: Does the AMD Instinct MI350P have any display outputs?

A: No. The database lists its display outputs as "No outputs," and its pixel rate is recorded as 0 MPixel/s, with DirectX, OpenGL, and Vulkan support listed as N/A.

Q: How does the Intel Arc B580 compare to the NVIDIA GeForce RTX 2080 in average benchmark score?

A: The B580 has an average benchmark score of 23,021, while the RTX 2080 scores 22,895, putting the B580 0.6 percent ahead.

Q: What is the memory bandwidth difference between the two cards?

A: The MI350P delivers 8.19 TB/s over an 8,192-bit HBM3e interface, while the B580 delivers 456.0 GB/s over a 192-bit GDDR6 interface.

Q: Which card has a higher FP32 throughput?

A: The MI350P leads with 36.04 TFLOPS, compared to the B580's 13.67 TFLOPS, a 2.6 times advantage.

Q: What power supply is suggested for each card?

A: The MI350P requires a 1000 W suggested PSU with a 600 W TDP and a single 16-pin connector. The B580 requires a 450 W suggested PSU with a 190 W TDP and a single 8-pin connector.

Q: What is the release date and production status of the Intel Arc B580?

A: The B580 was released on 2024-12-12 and its production status is listed as Active. The MI350P has a release date of 2026-05-06 and no production status recorded.