AMD Instinct MI350X vs Intel Arc B370 Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark results between the AMD Instinct MI350X and the Intel Arc B370. The headToHeadBenchmarks field is empty, and the MI350X has no benchmark entries or average score, while the Arc B370 has a single recorded result. This makes a direct comparative analysis of measured performance impossible from the data available.

The Intel Arc B370 posts an average benchmark score of 1184 points in the 3DMark Steel Nomad DX12 test. Its percentile ranking against all GPUs in the database is 5, meaning the recorded data places it in the lower portion of the performance distribution. The nearest rivals in the database for the Arc B370 are all older or lower-tier parts. The ATI Mobility Radeon HD 5570 scores 1186, a delta of -0.2% relative to the Arc B370, effectively matching it. The ATI Radeon HD 5770 scores 1190, a delta of -0.5%, also essentially identical. The AMD Radeon HD 7650M scores 1192, a delta of -0.7%, again within a rounding error. The AMD FirePro M2000 scores 1168, a delta of 1.4%, meaning the Arc B370 is 1.4% ahead of that part. These deltas are all within 2 percentage points, so the data indicates the Arc B370 occupies a performance band very close to these legacy parts.

The AMD Instinct MI350X has no benchmark scores, no average score, and no nearest rivals in the database. Its percentile ranking is 50, but with an average score of 0, this percentile cannot be interpreted as a measured performance result. The absence of data means no wins can be attributed to either product in a head-to-head sense. The MI350X cannot be positioned against the Arc B370 using any recorded benchmark output.

Architecture Differences

The two products diverge sharply in almost every architectural dimension recorded. The AMD Instinct MI350X uses the CDNA 4.0 architecture, built on a 3 nm process at TSMC, with the MI350 256CU chip. The Intel Arc B370 uses the Xe3-LPG architecture, built on a 3 nm process at Intel, with the Panther Lake chip. Both are 3 nm parts, but the foundries differ.

The MI350X integrates 185,000 million transistors on a die size of 2380 mm², yielding a transistor density of 77.7M per mm². The Arc B370 has unknown transistor count, unknown die size, and no recorded density. The MI350X is a massive accelerator die, while the Arc B370 is an integrated graphics processor with no discrete dimensions listed.

Compute resources differ by an order of magnitude. The MI350X has 16384 shading units, 1024 TMUs, and no ROPs. The Arc B370 has 1280 shading units, 40 TMUs, and 20 ROPs. The MI350X reports a texture rate of 2,252.8 GTexel/s and a pixel rate of 0 MPixel/s, consistent with a compute-focused accelerator with no display output. The Arc B370 reports a texture rate of 96.00 GTexel/s and a pixel rate of 48.00 GPixel/s, consistent with a graphics-oriented integrated part.

The MI350X delivers 72.09 TFLOPS FP32 and 72.09 TFLOPS FP16 at a 1:1 ratio. The Arc B370 delivers 6.144 TFLOPS FP32 and 12.29 TFLOPS FP16 at a 2:1 ratio. The MI350X has no recorded ray tracing cores, while the Arc B370 includes 10 RT cores. Neither part lists tensor cores in the database.

Memory systems are fundamentally different. The MI350X uses 288 GB of HBM3e on an 8192-bit bus, with 8.19 TB/s of bandwidth, at a memory clock of 2000 MHz (8 Gbps effective). The Arc B370 uses system shared memory, with system dependent bandwidth and no dedicated memory clock. Clock speeds also differ: the MI350X runs at a base of 1000 MHz and a boost of 2200 MHz, while the Arc B370 runs at a base of 300 MHz and a boost of 2400 MHz.

Power and physical configuration differ completely. The MI350X has a TDP of 1000 W, is an OAM Module, has no power connectors, and requires a suggested PSU of 1400 W. It measures 102 mm in length and 165 mm in width. The Arc B370 has a TDP of 25 W, is an IGP, has no power connectors, no suggested PSU, and no dimensions recorded. The MI350X uses a PCIe 5.0 x16 bus interface and has no display outputs. The Arc B370 uses an IGP bus interface and its display outputs are portable device dependent.

The MI350X has no API support recorded (DirectX, OpenGL, and Vulkan all listed as N/A). The Arc B370 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. This reflects the MI350X's role as a compute accelerator without a graphics driver stack in the database, versus the Arc B370's role as a fully featured integrated GPU.

Where Each One Wins

Based strictly on the recorded data, the AMD Instinct MI350X wins on raw compute throughput, memory capacity, memory bandwidth, and power envelope scale. Its 72.09 TFLOPS FP32 and FP16 figures dwarf the Arc B370's 6.144 TFLOPS FP32 and 12.29 TFLOPS FP16. Its 288 GB HBM3e pool with 8.19 TB/s bandwidth has no equivalent in the Arc B370, which relies on system shared memory with system dependent bandwidth. The MI350X also leads in shading units, TMUs, texture rate, transistor count, and die size. These figures place it in a category for data center scale acceleration, high-performance computing, and large model inference workloads where massive memory and FP32 throughput are primary requirements.

The Intel Arc B370 wins on graphics-specific features and integration. It has 20 ROPs to the MI350X's zero, a pixel rate of 48.00 GPixel/s versus 0 MPixel/s, and 10 ray tracing cores to the MI350X's none. It supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, while the MI350X records no API support. The Arc B370's 25 W TDP fits into power constrained, portable device contexts, whereas the MI350X demands 1000 W and an OAM module form factor. The Arc B370 also has a higher boost clock at 2400 MHz versus 2200 MHz, though its base clock is far lower at 300 MHz versus 1000 MHz.

The Arc B370 has the only observed benchmark result in this comparison, scoring 1184 in 3DMark Steel Nomad DX12. That score, however, places it near legacy parts like the ATI Mobility Radeon HD 5570, ATI Radeon HD 5770, and AMD Radeon HD 7650M, all within 1% of its score. No equivalent benchmark exists for the MI350X in the database, so no measured graphics or compute win can be assigned to it.

FAQ

Q: Which GPU has a higher FP32 compute throughput?

A: The AMD Instinct MI350X delivers 72.09 TFLOPS FP32, compared to the Intel Arc B370's 6.144 TFLOPS FP32, a roughly 11.7 times difference based on the recorded figures.

Q: What memory configuration does each product use?

A: The MI350X uses 288 GB of HBM3e on an 8192-bit bus with 8.19 TB/s bandwidth. The Arc B370 uses system shared memory with system dependent bandwidth and no dedicated memory size or bus width recorded.

Q: Does the Intel Arc B370 support ray tracing?

A: Yes, the Arc B370 includes 10 ray tracing cores. The MI350X has no ray tracing cores recorded in the database.

Q: What is the power consumption of each part?

A: The MI350X has a TDP of 1000 W and a suggested PSU of 1400 W. The Arc B370 has a TDP of 25 W and no suggested PSU listed.

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

A: The Arc B370 scores 1184, which is 0.2% below the ATI Mobility Radeon HD 5570 (1186), 0.5% below the ATI Radeon HD 5770 (1190), 0.7% below the AMD Radeon HD 7650M (1192), and 1.4% above the AMD FirePro M2000 (1168).

Q: Does the MI350X have any benchmark results in the database?

A: No, the MI350X has no benchmark entries and an average benchmark score of 0. Its percentile ranking of 50 cannot be tied to any measured result.

The Verdict

The recorded data supports a clear split by use case. The AMD Instinct MI350X is a 1000 W OAM module with 288 GB of HBM3e, 8.19 TB/s of bandwidth, and 72.09 TFLOPS of FP32 compute. It has no display outputs and no graphics API support in the database, which positions it as a compute-only accelerator for workloads that require massive memory capacity and sustained FP32 throughput. The Intel Arc B370 is a 25 W integrated GPU with 1280 shading units, 10 ray tracing cores, DirectX 12 Ultimate support, and a measured 3DMark Steel Nomad score of 1184, placing it near the performance level of the ATI Mobility Radeon HD 5570 and ATI Radeon HD 5770.

A user or system integrator selecting between these two parts would choose based on the physical and workload constraints. The MI350X requires a 1400 W PSU, an OAM module slot, and a PCIe 5.0 x16 connection, and it offers no video output. The Arc B370 is an IGP with no discrete power connector, no suggested PSU, and portable device dependent display outputs. The MI350X is designed for environments where 1000 W of power and a large accelerator board are acceptable. The Arc B370 is designed for embedded or mobile contexts where 25 W is the limit and graphics output is required.

Benchmark results do not exist for the MI350X in this database, so no measured performance comparison can be made. The Arc B370's only recorded score, 1184, is near the bottom of the database's percentile distribution at rank 5. The MI350X's percentile of 50 is not supported by any benchmark data and should not be read as a performance figure. The data therefore cannot support a claim that one product outperforms the other in any shared test. What the data does support is a functional distinction: the MI350X is a high-power, high-memory compute accelerator with no graphics stack, and the Arc B370 is a low-power integrated graphics processor with modern API support and a measured, if modest, 3DMark result.