AMD Instinct MI350P vs Intel Arc 130T Mobile Comparison

Head-to-Head Benchmarks

The recorded data contains no direct head-to-head benchmark results for the AMD Instinct MI350P and Intel Arc 130T Mobile. Both products show zero benchmark scores, zero recorded wins in their head-to-head comparison, and an identical 50th percentile ranking against all GPUs in the database. This absence of measured performance data means the comparison must rely entirely on the architectural and specification differences captured in the database, which are substantial.

The AMD Instinct MI350P delivers 36.04 TFLOPS of FP32 compute, while the Intel Arc 130T Mobile delivers 3.942 TFLOPS. This places the AMD part at roughly 9.1 times the single-precision throughput of the Intel part, a direct arithmetic comparison derived solely from the listed figures. In FP16, the AMD part again lists 36.04 TFLOPS with a 1:1 ratio, while the Intel part lists 7.885 TFLOPS with a 2:1 ratio, indicating the Intel architecture doubles its FP16 rate relative to FP32 while the AMD architecture maintains parity. The AMD part achieves a texture rate of 1,126.4 GTexel/s against the Intel part's 123.2 GTexel/s, a 9.1x advantage that mirrors the FP32 gap. The pixel rate comparison is inverted: the AMD part records 0 MPixel/s because it has 0 ROPs, while the Intel part achieves 61.60 GPixel/s from its 28 ROPs. This is a fundamental design divergence rather than a performance weakness in either product; the AMD accelerator is not built for rasterized pixel output, while the Intel mobile GPU is.

Memory capacity and bandwidth also diverge sharply. The AMD Instinct MI350P uses 144 GB of HBM3e across an 8192-bit bus, delivering 8.19 TB/s of bandwidth. The Intel Arc 130T Mobile uses system shared memory with system dependent bandwidth, meaning its memory performance is tied to the host platform rather than a fixed specification. The AMD memory clock is listed at 2000 MHz with 8 Gbps effective, while the Intel part has no dedicated memory clock because it shares system memory. Clock behavior shows both parts boosting to 2200 MHz, but the AMD base clock is 1000 MHz versus the Intel base clock of 300 MHz, indicating the AMD part sustains a higher operating frequency floor.

The Verdict

The database shows two products with no overlapping use cases. The AMD Instinct MI350P is a 600 W dual-slot accelerator with a 267 mm length, designed for compute workloads that require massive FP32 throughput, 144 GB of HBM3e capacity, and 8.19 TB/s bandwidth. It has no display outputs, no DirectX, OpenGL, or Vulkan API support, and a 0 MPixel/s pixel rate, confirming it is not a graphics card in any conventional sense. The Intel Arc 130T Mobile is a 35 W integrated graphics processor (IGP) with 896 shading units, 7 ray tracing cores, full API support including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, and a 61.60 GPixel/s pixel rate, confirming its role as a mobile graphics solution.

The data indicates the AMD part is for datacenter or server compute where rasterization is irrelevant. The Intel part is for portable devices where power draw of 35 W and integrated packaging are mandatory. The 600 W TDP of the AMD part versus the 35 W TDP of the Intel part is a 17.1x difference in power consumption. The AMD part requires a 1000 W suggested PSU and a 1x 16-pin power connector, while the Intel part has no power connector requirements because it is an IGP. There is no scenario in the recorded data where these two products compete for the same socket, chassis, or workload. The AMD part wins everything related to raw compute, memory capacity, and memory bandwidth. The Intel part wins everything related to graphics output, API compatibility, ray tracing, and power efficiency.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The AMD Instinct MI350P lists 36.04 TFLOPS of FP32, while the Intel Arc 130T Mobile lists 3.942 TFLOPS. The AMD part is approximately 9.1 times higher in this metric.

Q: Does the AMD Instinct MI350P support DirectX?

A: No. The database records DirectX as "N/A" for the AMD Instinct MI350P. It also records OpenGL and Vulkan as "N/A". The Intel Arc 130T Mobile supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What memory does each product use?

A: The AMD Instinct MI350P uses 144 GB of HBM3e with an 8192-bit bus and 8.19 TB/s bandwidth. The Intel Arc 130T Mobile uses system shared memory with system dependent bandwidth.

Q: What is the transistor count of each chip?

A: The AMD Instinct MI350P lists 73,000 million transistors on a 1190 mm² die. The Intel Arc 130T Mobile lists its transistor count as "unknown" and its die size as "unknown".

Q: What are the boost clocks of these two parts?

A: Both the AMD Instinct MI350P and the Intel Arc 130T Mobile list a boost clock of 2200 MHz. The base clocks differ: 1000 MHz for AMD and 300 MHz for Intel.

Q: Does the AMD Instinct MI350P have ray tracing cores?

A: The database records ray tracing cores as null for the AMD Instinct MI350P. The Intel Arc 130T Mobile lists 7 ray tracing cores.

Specification Differences

The two products differ across every major specification field in the database. The AMD Instinct MI350P uses a 3 nm process node from TSMC, while the Intel Arc 130T Mobile uses a 5 nm process node from TSMC. The AMD chip contains 73,000 million transistors on a 1190 mm² die, with a transistor density of 61.3M per mm². The Intel chip lists unknown transistor count and die size. The shading unit count is 8192 for AMD and 896 for Intel. Texture mapping units are 512 for AMD and 56 for Intel. Raster output units are 0 for AMD and 28 for Intel. The AMD part has no ray tracing cores listed, while the Intel part has 7. The AMD part has no display outputs, while the Intel part's display outputs are listed as "Portable Device Dependent". The AMD slot width is dual-slot, while the Intel slot width is IGP. The AMD power connector is 1x 16-pin with a suggested PSU of 1000 W, while the Intel part has no power connector and no suggested PSU. The AMD bus interface is PCIe 5.0 x16, while the Intel bus interface is IGP. The AMD dimensions are 267 mm length, 111 mm height, and 40 mm width, while the Intel dimensions are all null. The AMD TDP is 600 W, while the Intel TDP is 35 W. The AMD release date is recorded as May 6, 2026, while the Intel release date is recorded as January 12, 2025. The AMD production status is null, while the Intel production status is "Active". The AMD predecessor is "Radeon Instinct", while the Intel predecessor is "HD Graphics-M".

Architecture Differences

The AMD Instinct MI350P uses the CDNA 4.0 architecture with the MI350 128CU chip, while the Intel Arc 130T Mobile uses the Xe-LPG+ architecture with the Arrow Lake-H chip. The AMD generation is "Instinct (MIx)", while the Intel generation is "Arc Graphics-M (Arrow Lake)". The AMD memory type is HBM3e with a fixed 8192-bit bus and 8.19 TB/s bandwidth. The Intel memory type is system shared, with a system shared bus width and system dependent bandwidth. The AMD FP16 performance is 36.04 TFLOPS at a 1:1 ratio to FP32, indicating equal throughput for both precisions. The Intel FP16 performance is 7.885 TFLOPS at a 2:1 ratio to FP32, indicating twice the FP32 throughput. The AMD pixel rate is 0 MPixel/s because the part has no ROPs, while the Intel pixel rate is 61.60 GPixel/s. The AMD texture rate is 1,126.4 GTexel/s, while the Intel texture rate is 123.2 GTexel/s. The AMD part has no API support recorded for DirectX, OpenGL, or Vulkan. The Intel part supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The AMD part has no ray tracing cores, while the Intel part has 7. The AMD part is a dual-slot accelerator with a 16-pin power connector and a 1000 W suggested PSU. The Intel part is an IGP with no power connector, no suggested PSU, and no discrete dimensions. The AMD process node is 3 nm, the Intel process node is 5 nm, both from TSMC.

Where Each One Wins

The AMD Instinct MI350P wins decisively in compute density, memory capacity, and memory bandwidth. Its 36.04 TFLOPS FP32 and FP16 throughput, 144 GB of HBM3e, and 8.19 TB/s bandwidth make it suitable for workloads that process large datasets in memory, such as large-scale matrix operations or inference tasks that exceed the memory capacity of smaller accelerators. Its 8192 shading units and 512 texture mapping units provide the parallel execution resources for high-throughput compute. The 600 W TDP and 1000 W suggested PSU indicate a stationary, power-rich environment such as a server chassis. The absence of display outputs and graphics APIs confirms that its winning domain is compute-only acceleration, not graphics rendering.

The Intel Arc 130T Mobile wins in graphics output, API compatibility, ray tracing, and power efficiency. Its 61.60 GPixel/s pixel rate, 28 ROPs, and 7 ray tracing cores provide actual rasterization and ray-traced rendering capability. Its support for DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 means it can run modern graphics workloads that the AMD part cannot. Its 35 W TDP and IGP slot width make it suitable for portable devices where space and power are constrained. Its system shared memory approach means it scales with the host platform rather than carrying dedicated memory, which is typical for integrated graphics. The Intel part also has a recorded production status of "Active", while the AMD part has no production status recorded.

The data shows a complete separation of function. The AMD part is a compute accelerator with no graphics capabilities. The Intel part is a graphics processor with no discrete compute aspirations. Neither product has recorded benchmark scores, so performance conclusions are based strictly on the specification deltas. The AMD part is 9.1 times higher in FP32 throughput and texture rate, but the Intel part is the only one of the two with any pixel output capability. The AMD part is the only one with fixed memory capacity and bandwidth, while the Intel part depends entirely on system memory. The release dates differ by roughly 16 months, with Intel shipping first. The transistor count for the AMD part is 73,000 million, while the Intel part's transistor count is unknown, which limits direct comparison of manufacturing complexity beyond the node difference.