AMD Instinct MI350X vs Intel Arc Pro A60M Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark comparisons between the AMD Instinct MI350X and the Intel Arc Pro A60M. Both entries show an average benchmark score of zero and a percentile ranking of 50 against all GPUs, which places them at the median of the database's distribution but provides no comparative performance data.

The absence of benchmark scores means the relative performance must be derived from the architectural specifications and compute characteristics recorded in the database. The AMD Instinct MI350X delivers 72.09 TFLOPS of FP32 compute, while the Intel Arc Pro A60M delivers 5.325 TFLOPS of FP32 compute. This represents a 13.5x advantage for the AMD part in single-precision floating-point throughput, a gap that reflects their fundamentally different positioning within the database.

For FP16 workloads, the AMD Instinct MI350X again outputs 72.09 TFLOPS with a 1:1 ratio, meaning there is no throughput advantage over its FP32 rate. The Intel Arc Pro A60M outputs 10.65 TFLOPS of FP16 with a 2:1 ratio, doubling its FP32 rate. Even with this ratio advantage, the Intel part still trails by a factor of 6.8 in half-precision compute.

Texture throughput shows a similar pattern. The AMD Instinct MI350X reaches 2,252.8 GTexel/s, while the Intel Arc Pro A60M reaches 166.4 GTexel/s, a difference of roughly 13.5x. The pixel rate comparison is inverted in an unusual way: the AMD part records 0 MPixel/s, while the Intel part records 83.20 GPixel/s. This reflects the AMD accelerator's lack of a conventional raster output stage, as its ROP count is recorded as zero.

Memory bandwidth heavily favors the AMD part. The Instinct MI350X provides 8.19 TB/s of bandwidth across an 8192-bit bus, while the Arc Pro A60M provides 256.0 GB/s across a 128-bit bus. That is a 32x difference in memory bandwidth. The AMD part also carries 288 GB of HBM3e memory versus 8 GB of GDDR6 on the Intel part, a 36x capacity advantage.

Clock speeds are closer than the compute rates might suggest. The AMD Instinct MI350X runs at a 1000 MHz base and 2200 MHz boost, while the Intel Arc Pro A60M runs at 900 MHz base and 1300 MHz boost. The AMD part's higher boost clock partially explains its compute advantage, but the scale of the gap comes primarily from the massive difference in shading units: 16,384 on the AMD part versus 2,048 on the Intel part. TMUs follow the same 8x ratio, with 1,024 on the AMD part and 128 on the Intel part.

The process node and die size differences are substantial. The AMD Instinct MI350X is built on a 3 nm process at TSMC with 185,000 million transistors on a 2380 mm² die, yielding a transistor density of 77.7M per mm². The Intel Arc Pro A60M uses a 6 nm process at TSMC with 11,500 million transistors on a 269 mm² die, yielding 42.8M per mm². The AMD part packs 16x more transistors onto a die that is 8.8x larger, while achieving 1.8x higher transistor density.

Power envelopes differ enormously. The AMD Instinct MI350X carries a TDP of 1000 W and a suggested PSU of 1400 W, while the Intel Arc Pro A60M carries a TDP of 95 W. The AMD part consumes roughly 10.5x the power of the Intel part, which is consistent with its far larger silicon area and compute throughput.

Architecture Differences

The two accelerators belong to entirely different product families. The AMD Instinct MI350X sits in the Instinct (MIx) generation under the CDNA 4.0 architecture, while the Intel Arc Pro A60M belongs to the Alchemist generation under the Xe-HPG architecture and is classified as a Pro-Series Mobile part.

The AMD Instinct MI350X uses the MI350 256CU chip, which indicates a 256 compute unit configuration. The Intel Arc Pro A60M uses the DG2-256 chip. The compute unit organization differs fundamentally: the AMD part exposes 16,384 shading units, 1,024 TMUs, and zero ROPs, while the Intel part exposes 2,048 shading units, 128 TMUs, 64 ROPs, and 16 ray tracing cores. The AMD part records no ray tracing core count, matching its compute-accelerator orientation.

Memory architecture diverges sharply. The AMD Instinct MI350X uses 288 GB of HBM3e on an 8192-bit bus with a memory clock of 2000 MHz at 8 Gbps effective. The Intel Arc Pro A60M uses 8 GB of GDDR6 on a 128-bit bus with a memory clock of 2000 MHz at 16 Gbps effective. Despite identical memory clock base values, the effective data rates differ because of the memory type and bus width. The AMD part's bandwidth advantage comes from the massive bus width, not from a higher memory clock.

Process technology shows a two-generation gap. The AMD part is fabricated on a 3 nm process, while the Intel part uses a 6 nm process, both at TSMC. Transistor counts reflect this: 185,000 million on the AMD part versus 11,500 million on the Intel part. Die size is 2380 mm² for the AMD part versus 269 mm² for the Intel part. Transistor density is 77.7M per mm² for the AMD part versus 42.8M per mm² for the Intel part, indicating that the smaller process node delivers higher packing density in addition to the raw die size advantage.

Interface capabilities differ. The AMD Instinct MI350X uses PCIe 5.0 x16, while the Intel Arc Pro A60M uses PCIe 4.0 x16. The AMD part is an OAM Module with no power connectors and no display outputs. The Intel part is an IGP with portable-device-dependent display outputs. The AMD part supports no DirectX, OpenGL, or Vulkan APIs, while the Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. This confirms the AMD part is a pure compute accelerator, while the Intel part retains graphics API support.

Physical dimensions are only recorded for the AMD part: 102 mm in length and 165 mm in width, described as 4 inches by 6.5 inches. The database records no dimensions for the Intel part.

FAQ

Q: Which GPU has higher FP32 compute throughput?

A: The AMD Instinct MI350X delivers 72.09 TFLOPS of FP32 compute, which is 13.5x the 5.325 TFLOPS delivered by the Intel Arc Pro A60M.

Q: How do the memory subsystems compare?

A: The AMD Instinct MI350X uses 288 GB of HBM3e on an 8192-bit bus with 8.19 TB/s bandwidth. The Intel Arc Pro A60M uses 8 GB of GDDR6 on a 128-bit bus with 256.0 GB/s bandwidth. The AMD part has 32x the bandwidth and 36x the capacity.

Q: Does the Intel Arc Pro A60M support graphics APIs?

A: Yes. The Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The AMD Instinct MI350X records N/A for DirectX, OpenGL, and Vulkan, and has no display outputs.

Q: What are the power requirements of each part?

A: The AMD Instinct MI350X has a TDP of 1000 W and a suggested PSU of 1400 W. The Intel Arc Pro A60M has a TDP of 95 W and no recorded suggested PSU.

Q: What process nodes are used?

A: The AMD Instinct MI350X uses a 3 nm process at TSMC with 185,000 million transistors on a 2380 mm² die. The Intel Arc Pro A60M uses a 6 nm process at TSMC with 11,500 million transistors on a 269 mm² die.

Q: Which part has ray tracing hardware?

A: The Intel Arc Pro A60M has 16 ray tracing cores. The AMD Instinct MI350X records no ray tracing core count, consistent with its compute-focused CDNA 4.0 architecture.

The Verdict

The database presents two products with no overlapping use cases. The AMD Instinct MI350X is a 1000 W OAM module with 288 GB of HBM3e, 16,384 shading units, and no display or graphics API support. The Intel Arc Pro A60M is a 95 W mobile IGP with 8 GB of GDDR6, 2,048 shading units, 16 ray tracing cores, and full graphics API support including DirectX 12 Ultimate and Vulkan 1.4.

For compute throughput, the AMD Instinct MI350X is the clear choice. It leads by 13.5x in FP32, 6.8x in FP16, and 13.5x in texture rate. Its 8.19 TB/s memory bandwidth and 288 GB capacity support workloads that require large model residency and high-bandwidth data movement. The 1000 W TDP and 1400 W suggested PSU indicate the power delivery infrastructure required to sustain this performance.

For graphics-oriented workloads or mobile deployment, the Intel Arc Pro A60M is the only option with the required features. It carries 64 ROPs and 16 ray tracing cores, supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, and fits within a 95 W envelope. Its 83.20 GPixel/s pixel rate, while modest, is real, whereas the AMD part records 0 MPixel/s and cannot drive a display.

The release dates place the AMD part in June 2025 and the Intel part in June 2023. The AMD part succeeds the Radeon Instinct line, while the Intel part has no recorded predecessor. Both parts sit at the 50th percentile in the database, but this ranking reflects the absence of benchmark scores rather than equivalent measured performance.

Specification Differences

The following fields differ between the AMD Instinct MI350X and the Intel Arc Pro A60M:

• Chip: MI350 256CU vs DG2-256
• Architecture: CDNA 4.0 vs Xe-HPG
• Generation: Instinct (MIx) vs Alchemist (Pro-Series Mobile)
• Process Node: 3 nm vs 6 nm
• Transistors: 185,000 million vs 11,500 million
• Die Size: 2380 mm² vs 269 mm²
• Transistor Density: 77.7M / mm² vs 42.8M / mm²
• Base Clock: 1000 MHz vs 900 MHz
• Boost Clock: 2200 MHz vs 1300 MHz
• Memory Clock: 2000 MHz 8 Gbps effective vs 2000 MHz 16 Gbps effective
• Memory Size: 288 GB vs 8 GB
• Memory Type: HBM3e vs GDDR6
• Memory Bus Width: 8192 bit vs 128 bit
• Memory Bandwidth: 8.19 TB/s vs 256.0 GB/s
• Shading Units: 16384 vs 2048
• TMUs: 1024 vs 128
• ROPs: 0 vs 64
• Ray Tracing Cores: None recorded vs 16
• Pixel Rate: 0 MPixel/s vs 83.20 GPixel/s
• Texture Rate: 2,252.8 GTexel/s vs 166.4 GTexel/s
• FP32: 72.09 TFLOPS vs 5.325 TFLOPS
• FP16: 72.09 TFLOPS (1:1) vs 10.65 TFLOPS (2:1)
• TDP: 1000 W vs 95 W
• Slot Width: OAM Module vs IGP
• Power Connectors: None vs not recorded
• Suggested PSU: 1400 W vs not recorded
• Bus Interface: PCIe 5.0 x16 vs PCIe 4.0 x16
• Display Outputs: No outputs vs Portable Device Dependent
• DirectX: N/A vs 12 Ultimate (12_2)
• OpenGL: N/A vs 4.6
• Vulkan: N/A vs 1.4
• Dimensions: 102 mm x 165 mm vs not recorded
• Release Date: 2025-06-11 vs 2023-06-05
• Production Status: Not recorded vs Active