AMD Instinct MI350P vs Intel Arc B370 Comparison

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark results between the AMD Instinct MI350P and the Intel Arc B370. The Intel Arc B370 has a single recorded benchmark score in the 3DMark Steel Nomad DX12 test, achieving 1184 points. The AMD Instinct MI350P has no recorded benchmark scores in the database.

Looking at the Intel Arc B370's nearest rivals in the database, its performance profile is tightly clustered. The ATI Mobility Radeon HD 5570 scores 1186 points, placing the Arc B370 just 0.2% behind that part. The ATI Radeon HD 5770 scores 1190 points, a 0.5% gap. The AMD Radeon HD 7650M leads with 1192 points, putting the Arc B370 0.7% behind. The AMD FirePro M2000 trails at 1168 points, with the Arc B370 ahead by 1.4%.

These margins are exceptionally narrow. The entire spread between the fastest rival (1192) and the slowest rival (1168) is only 24 points, or roughly 2%. The Arc B370 sits almost exactly in the middle of this cluster. The percentile data confirms this positioning: the Arc B370 ranks in the 5th percentile of all GPUs in the database, indicating that most recorded graphics processors score substantially higher.

The AMD Instinct MI350P holds a 50th percentile ranking with an average benchmark score of 0, meaning it has no performance data recorded. Without a Steel Nomad score for the MI350P, no direct percentage comparison can be calculated. The MI350P's compute specifications, however, suggest a fundamentally different performance class. Its FP32 throughput of 36.04 TFLOPS is roughly 5.9 times the Arc B370's 6.144 TFLOPS. Texture rate shows an even larger gap: 1,126.4 GTexel/s versus 96.00 GTexel/s, a factor of about 11.7.

The pixel rate comparison is unusual. The MI350P reports 0 MPixel/s, indicating it lacks traditional raster output pipelines. The Arc B370 delivers 48.00 GPixel/s from its 20 ROPs. This reflects their different design targets: the MI350P is an accelerator without display outputs, while the Arc B370 is an integrated graphics solution with display functionality.

The FP16 comparison reveals different precision strategies. The MI350P delivers 36.04 TFLOPS FP16 at a 1:1 ratio with FP32, meaning it does not double-rate half-precision work. The Arc B370 delivers 12.29 TFLOPS FP16 at a 2:1 ratio, doubling its FP32 throughput. Even with this advantage, the MI350P still leads in raw FP16 compute by roughly 2.9 times.

The Verdict

The data presents two products with almost no overlap in purpose or performance class. The AMD Instinct MI350P is a 600 W dual-slot accelerator with 144 GB of HBM3e memory on an 8192-bit bus, delivering 8.19 TB/s of bandwidth. It uses a 3 nm TSMC process with 73,000 million transistors on a 1190 mm² die. The Intel Arc B370 is a 25 W integrated graphics processor with system-shared memory and no dedicated VRAM.

Benchmark results indicate the Arc B370 belongs to a performance tier near decade-old discrete GPUs. Its 1184 Steel Nomad score places it within 1% of the ATI Mobility Radeon HD 5570, ATI Radeon HD 5770, and AMD Radeon HD 7650M. The 5th percentile ranking confirms this is a low-tier result in the overall database distribution.

The MI350P has no recorded benchmarks, so its real-world game or workload performance cannot be statistically verified. Its compute specifications, however, place it in an entirely different category. The shading unit count of 8192 versus 1280, the memory bandwidth of 8.19 TB/s versus system-dependent shared memory, and the texture rate of 1,126.4 GTexel/s versus 96.00 GTexel/s all indicate a multi-hundred-TFLOPS-class accelerator.

Users selecting between these two parts should base the decision on the workload type. The MI350P targets data center compute tasks requiring massive memory capacity and bandwidth, with no display outputs and no graphics API support (DirectX, OpenGL, and Vulkan all listed as N/A). The Arc B370 targets portable devices needing basic graphics acceleration, supporting DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, with display outputs described as portable device dependent.

The release dates show the MI350P launching later, on 2026-05-06, while the Arc B370 arrived on 2026-01-26. The Arc B370 carries an active production status, while the MI350P's production status is unlisted.

FAQ

Q: How does the Intel Arc B370 compare to its nearest rivals in the database?

A: The Arc B370 scores 1184 points in 3DMark Steel Nomad DX12. It sits 0.2% behind the ATI Mobility Radeon HD 5570 (1186 points), 0.5% behind the ATI Radeon HD 5770 (1190 points), 0.7% behind the AMD Radeon HD 7650M (1192 points), and 1.4% ahead of the AMD FirePro M2000 (1168 points).

Q: What is the FP32 compute difference between the two GPUs?

A: The AMD Instinct MI350P delivers 36.04 TFLOPS FP32, while the Intel Arc B370 delivers 6.144 TFLOPS. This gives the MI350P approximately 5.9 times the single-precision throughput.

Q: Which GPU supports graphics APIs?

A: The Intel Arc B370 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The AMD Instinct MI350P lists N/A for DirectX, OpenGL, and Vulkan, indicating no graphics API support.

Q: What memory configurations do the two cards use?

A: The MI350P has 144 GB of HBM3e memory with an 8192-bit bus and 8.19 TB/s bandwidth. The Arc B370 uses system shared memory with system-dependent bandwidth.

Q: What are the power requirements?

A: The MI350P has a 600 W TDP with a 1x 16-pin power connector and a suggested 1000 W PSU. The Arc B370 has a 25 W TDP, uses no power connectors, and is an integrated graphics processor (IGP).

Q: How do the pixel rates compare?

A: The MI350P reports 0 MPixel/s due to having 0 ROPs. The Arc B370 delivers 48.00 GPixel/s from its 20 ROPs.

Specification Differences

The two GPUs differ across nearly every recorded specification. The MI350P uses the MI350 128CU chip with CDNA 4.0 architecture, while the Arc B370 uses the Panther Lake chip with Xe3-LPG architecture. Both use a 3 nm process, but the MI350P comes from TSMC while the Arc B370 comes from Intel.

The MI350P lists 73,000 million transistors on a 1190 mm² die with a density of 61.3M per mm². The Arc B370's transistor count, die size, and density are unknown. Clock speeds differ substantially: the MI350P has a 1000 MHz base and 2200 MHz boost, while the Arc B370 has a 300 MHz base and 2400 MHz boost.

Memory configurations are completely different. The MI350P has 144 GB HBM3e on an 8192-bit bus with 8.19 TB/s bandwidth and 2000 MHz memory clock (8 Gbps effective). The Arc B370 uses system shared memory with no dedicated VRAM. The MI350P has 8192 shading units, 512 TMUs, and 0 ROPs. The Arc B370 has 1280 shading units, 40 TMUs, and 20 ROPs. The Arc B370 also has 10 ray tracing cores, while the MI350P lists none.

Power and physical specifications diverge sharply. The MI350P draws 600 W, uses a 1x 16-pin connector, requires a 1000 W PSU, and is dual-slot with dimensions of 267 mm length, 111 mm height, and 40 mm width. The Arc B370 draws 25 W, has no power connectors, no PSU suggestion, and is an IGP with no listed dimensions.

The MI350P has no display outputs, while the Arc B370's outputs are portable device dependent. The bus interface is PCIe 5.0 x16 for the MI350P and IGP for the Arc B370. The MI350P lists no API support (DirectX, OpenGL, Vulkan all N/A), while the Arc B370 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Architecture Differences

The AMD Instinct MI350P uses CDNA 4.0 architecture, designed specifically for compute acceleration. It belongs to the Instinct (MIx) generation and lists Radeon Instinct as its predecessor. The Intel Arc B370 uses Xe3-LPG architecture, part of the Arc Graphics-M (Panther Lake) generation, and has no predecessor listed.

The MI350P's CDNA 4.0 design prioritizes memory bandwidth and compute density. Its 8192 shading units process FP32 and FP16 at the same 36.04 TFLOPS rate (1:1 ratio), indicating a compute-focused design without consumer graphics features. The 0 ROPs and 0 MPixel/s pixel rate confirm this focus. The 144 GB HBM3e pool with 8.19 TB/s bandwidth targets large model training and inference workloads.

The Arc B370's Xe3-LPG architecture is a low-power graphics solution for mobile and portable devices. Its FP16 rate of 12.29 TFLOPS is double its FP32 rate of 6.144 TFLOPS (2:1 ratio), a common consumer GPU optimization. The 10 ray tracing cores enable hardware-accelerated ray tracing in supported games. The 48.00 GPixel/s pixel rate from 20 ROPs supports display output and rasterization.

The MI350P's 3 nm TSMC process with 73,000 million transistors on a 1190 mm² die represents a large accelerator-class chip. The Arc B370's transistor count and die size are not recorded, but its 3 nm Intel process and 25 W TDP indicate a small integrated design. The MI350P's transistor density of 61.3M per mm² reflects the trade-off between compute resources and interconnects on the large die.

The MI350P's 2000 MHz memory clock with 8 Gbps effective speed drives its 8192-bit bus to 8.19 TB/s. The Arc B370's system-shared memory has no dedicated bandwidth figure, making its memory performance system dependent. The MI350P's PCIe 5.0 x16 interface supports host communication for accelerator workloads, while the Arc B370's IGP bus connects directly to the processor.

The MI350P's CDNA 4.0 architecture omits display outputs and graphics APIs entirely, confirming its role as a compute accelerator. The Arc B370's Xe3-LPG architecture includes full graphics API support and portable display outputs, confirming its role as an integrated graphics solution. The MI350P's 600 W TDP and dual-slot cooling require a 1000 W PSU, while the Arc B370's 25 W TDP requires no additional power infrastructure.