AMD Instinct MI350P vs Intel Arc Pro B60 Dual Comparison

Head-to-Head Benchmarks

The recorded database contains no head-to-head benchmark results for the AMD Instinct MI350P against the Intel Arc Pro B60 Dual. Both entries return zero wins in direct comparison, and the average benchmark score for each is zero. The percentile versus all GPUs is identical at the 50th mark for both parts, which places them at the median of the tracked GPU population despite their entirely different design goals. The absence of measured results means the performance relationship between these two accelerators must be derived from their architectural and specification data, not from direct testing.

The AMD Instinct MI350P delivers 36.04 TFLOPS of FP32 compute, which is roughly three times the 12.29 TFLOPS of the Intel Arc Pro B60 Dual. In FP16 workloads, the AMD part also outputs 36.04 TFLOPS with a 1:1 ratio, while the Intel part reaches 24.58 TFLOPS using a 2:1 ratio. The AMD accelerator leads in texture throughput at 1,126.4 GTexel/s versus 384.0 GTexel/s for Intel. However, the Intel Arc Pro B60 Dual counters with a pixel rate of 192.0 GPixel/s, while the AMD part records 0 MPixel/s, reflecting its lack of any display or raster output capability.

Memory bandwidth is another decisive split. The AMD Instinct MI350P accesses 144 GB of HBM3e across an 8192-bit bus, producing 8.19 TB/s of bandwidth. The Intel Arc Pro B60 Dual uses 24 GB of GDDR6 on a 192-bit bus, delivering 456.0 GB/s. The AMD bandwidth advantage is roughly 18 times, which is consistent with its role as a compute accelerator designed for memory-heavy workloads. The Intel part, by contrast, provides a conventional memory subsystem suited to graphics and professional visualization tasks.

FAQ

Q: Which GPU has more shading units?

A: The AMD Instinct MI350P has 8192 shading units, compared to 2560 shading units on the Intel Arc Pro B60 Dual.

Q: What is the memory size difference?

A: The AMD Instinct MI350P has 144 GB of HBM3e memory, while the Intel Arc Pro B60 Dual has 24 GB of GDDR6 memory.

Q: Which card supports display outputs?

A: Only the Intel Arc Pro B60 Dual has display outputs, offering 4x mini-DisplayPort 2.1. The AMD Instinct MI350P has no display outputs.

Q: How do the boost clocks compare?

A: The Intel Arc Pro B60 Dual boosts to 2400 MHz, while the AMD Instinct MI350P boosts to 2200 MHz. The Intel part also has a higher base clock at 2000 MHz versus 1000 MHz.

Q: What is the transistor count for each?

A: The AMD Instinct MI350P contains 73,000 million transistors on a 1190 mm² die, and the Intel Arc Pro B60 Dual contains 19,600 million transistors on a 272 mm² die.

Q: Which card has ray tracing cores?

A: The Intel Arc Pro B60 Dual has 20 ray tracing cores. The AMD Instinct MI350P does not list ray tracing cores.

Architecture Differences

The AMD Instinct MI350P is built on CDNA 4.0 architecture using a 3 nm process at TSMC, while the Intel Arc Pro B60 Dual uses Xe2-HPG architecture, also fabricated by TSMC but on a 5 nm node. The AMD chip, labeled MI350 128CU, packs 73,000 million transistors into a 1190 mm² die, producing a transistor density of 61.3M per mm². The Intel chip, BMG-G21, contains 19,600 million transistors on a 272 mm² die, yielding a higher density of 72.1M per mm². The smaller Intel die achieves better packing efficiency, but the AMD die is far larger in absolute terms.

The AMD part uses HBM3e memory with a massive 8192-bit bus, while the Intel part uses GDDR6 with a 192-bit bus. This difference reflects the intended use: the Instinct MI350P targets data-center compute with extreme bandwidth demands, whereas the Arc Pro B60 targets professional graphics with moderate memory requirements. The AMD accelerator has no display outputs, no raster operation pipelines, and no API support listed for DirectX, OpenGL, or Vulkan. The Intel accelerator supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, along with 80 ROPs and 20 ray tracing cores.

The FP16 ratio also differs. The AMD part runs FP16 at the same rate as FP32 (1:1), meaning it does not double throughput for half precision. The Intel part runs FP16 at twice the FP32 rate (2:1), delivering 24.58 TFLOPS from a 12.29 TFLOPS FP32 base. This indicates different precision strategies, with the AMD part prioritizing consistent throughput and the Intel part offering accelerated half-precision math.

Specification Differences

The two cards differ across nearly every major specification field. The AMD Instinct MI350P uses a 3 nm process, while the Intel Arc Pro B60 Dual uses 5 nm. Transistor counts are 73,000 million versus 19,600 million, and die sizes are 1190 mm² versus 272 mm². The AMD part has 8192 shading units, 512 TMUs, and 0 ROPs; the Intel part has 2560 shading units, 160 TMUs, and 80 ROPs. Only the Intel part lists ray tracing cores, with 20 present.

Clock speeds favor Intel: base clock of 2000 MHz versus 1000 MHz, and boost clock of 2400 MHz versus 2200 MHz. Memory clocks also differ, with the Intel part at 2375 MHz (19 Gbps effective) and the AMD part at 2000 MHz (8 Gbps effective). Memory size, type, bus width, and bandwidth all favor AMD: 144 GB of HBM3e on an 8192-bit bus with 8.19 TB/s, versus 24 GB of GDDR6 on a 192-bit bus with 456.0 GB/s.

The power profile diverges as well. The AMD card has a 600 W TDP and suggests a 1000 W power supply, while the Intel card has a 400 W TDP and suggests an 800 W power supply. Both use a single 16-pin power connector and a dual-slot cooler. The bus interface differs: PCIe 5.0 x16 for AMD versus PCIe 5.0 x8 for Intel. Dimensions are similar in height and width, but the Intel card is longer at 300 mm (11.8 inches) versus 267 mm (10.5 inches) for AMD. The AMD part has no display outputs; the Intel part has 4x mini-DisplayPort 2.1. The Intel card is listed as active in production, while the AMD card has no production status recorded. The Intel part has a launch MSRP of 1,199 USD; the AMD part has no launch MSRP listed.

Where Each One Wins

The AMD Instinct MI350P wins in raw compute throughput, memory capacity, memory bandwidth, and texture processing. Its 36.04 TFLOPS FP32 output, 144 GB HBM3e memory, 8.19 TB/s bandwidth, and 1,126.4 GTexel/s texture rate make it suited for data-center compute tasks such as large-scale matrix operations, scientific simulation, and workloads that require enormous memory pools. The 8192-bit bus and 73,000 million transistors indicate a design built for sustained compute density rather than interactive graphics.

The Intel Arc Pro B60 Dual wins in graphics-related features and practical deployment. It has 80 ROPs, 20 ray tracing cores, 192.0 GPixel/s pixel throughput, and 4x mini-DisplayPort 2.1 outputs. It supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, making it usable in professional visualization, content creation, and any application requiring a physical display connection. Its higher base and boost clocks, along with a smaller 272 mm² die, suggest better frequency headroom and easier integration into systems with lower power budgets, given its 400 W TDP and 800 W suggested power supply.

The AMD part has no display outputs and no graphics API support, so it cannot drive monitors or run conventional graphics workloads. The Intel part, while lower in raw compute, offers a complete graphics stack. The AMD card is longer in compute capability but absent in visual output; the Intel card is the reverse. The choice between them depends entirely on whether the workload demands compute and memory scale or graphics and display capability.

The Verdict

The recorded data shows two accelerators with almost no overlap in intended function. The AMD Instinct MI350P is a compute-first device: no display outputs, no graphics APIs, 600 W TDP, and a memory subsystem built around 144 GB of HBM3e with 8.19 TB/s bandwidth. Its FP32 performance of 36.04 TFLOPS is three times that of the Intel part, and its FP16 output matches its FP32 rate at 36.04 TFLOPS. Any user needing maximum memory capacity, bandwidth, or raw FP32 throughput should select the AMD Instinct MI350P based on these specifications.

The Intel Arc Pro B60 Dual is the only one of the two that can output to a display, with 4x mini-DisplayPort 2.1, and the only one with graphics API support, including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4. It also provides ray tracing cores and a functional pixel pipeline at 192.0 GPixel/s. Its 24 GB GDDR6 memory and 456.0 GB/s bandwidth are far below the AMD part, but its 400 W TDP, 800 W suggested power supply, and compact 272 mm² die make it practical for workstation graphics. The Intel card is the appropriate pick for rasterization, ray tracing, display-driven workflows, and any environment that requires a standard graphics driver stack.

The launch MSRP for the Intel Arc Pro B60 Dual is 1,199 USD, while the AMD card has no listed price. The Intel card is also marked as active in production, whereas the AMD card has no production status. These factors, combined with the specification split, indicate that the two cards serve separate markets rather than competing directly. The data does not show a single winner; it shows two specialized tools with distinct strengths.