AMD Radeon Instinct MI300 vs Intel Arc Pro B370 Comparison

Head-to-Head Benchmarks

The database contains no recorded benchmark scores for either the AMD Radeon Instinct MI300 or the Intel Arc Pro B370. Both entries show an average benchmark score of 0 and a percentile ranking of 50 against all GPUs. With no head-to-head benchmark results, no wins are registered for either product (winsA: 0, winsB: 0). The absence of measured performance data means a direct numerical comparison of frame rates, compute throughput, or rendering performance cannot be established from the recorded information. What the data does provide are the raw specification sheets, which reveal two fundamentally different hardware types. The MI300 is a data center accelerator with 14,080 shading units, 880 texture mapping units, and a peak FP32 throughput of 47.87 TFLOPS. The Arc Pro B370 is an integrated graphics processor with 1,280 shading units, 40 TMUs, 20 ROPs, and an FP32 rating of 6.144 TFLOPS. The MI300 holds an 8.8x advantage in shading unit count and a 7.8x advantage in FP32 compute. Those gaps define the separation between the two parts more clearly than any single benchmark run could.

FAQ

Q: Which GPU has the higher memory bandwidth?

A: The AMD Radeon Instinct MI300 uses 128 GB of HBM3 memory across an 8192-bit bus, delivering 6.55 TB/s bandwidth. The Intel Arc Pro B370 uses system shared memory, so its bandwidth is listed as system dependent and cannot be compared directly.

Q: What is the power consumption difference between the two?

A: The MI300 carries a TDP of 600 W and requires 2x 8-pin power connectors with a suggested PSU of 1000 W. The Arc Pro B370 has a TDP of 25 W, uses no external power connectors, and has no suggested PSU rating. The MI300 consumes 24x the power of the integrated part.

Q: How do the process nodes compare?

A: The MI300 is fabricated on a 5 nm process at TSMC. The Arc Pro B370 uses a 3 nm process at Intel. No transistor count or die size is recorded for the Intel part; the MI300 integrates 153,000 million transistors on a 1017 mm² die.

Q: Do both GPUs support ray tracing?

A: The Intel Arc Pro B370 includes 10 ray tracing cores. The AMD Radeon Instinct MI300 has no recorded ray tracing core count; its ROP count is listed as 0, and its pixel rate is 0 MPixel/s, indicating it is not designed for conventional rasterization workloads.

Q: Which GPU has a higher boost clock?

A: The Intel Arc Pro B370 boosts to 2400 MHz from a 300 MHz base. The AMD Radeon Instinct MI300 boosts to 1700 MHz from a 1000 MHz base. The Intel part has a 700 MHz higher boost clock, though its shading unit count is far lower.

Q: What display outputs does each GPU support?

A: The AMD Radeon Instinct MI300 has no display outputs. The Intel Arc Pro B370 lists its display outputs as portable device dependent, meaning output capability varies with the host device.

The Verdict

The recorded data positions these two products for completely separate use cases. The AMD Radeon Instinct MI300, with 47.87 TFLOPS FP32, 6.55 TB/s memory bandwidth, and 128 GB of HBM3, is a server accelerator aimed at compute-heavy data center tasks. The Intel Arc Pro B370, with 6.144 TFLOPS FP32, 10 ray tracing cores, and a 25 W TDP, is an integrated processor intended for portable devices with modest graphics needs. Neither part targets the same buyer. A builder assembling a rack-scale compute node selects the MI300 for its massive memory pool and raw throughput. A system integrator designing a low-power portable device chooses the Arc Pro B370 for its minimal power draw, integrated bus interface, and complete API support including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4. The MI300 has no recorded API support, no display outputs, and no ray tracing cores, reinforcing its role as a compute-only accelerator. The verdict is straightforward: pick the MI300 for dense compute and memory capacity, pick the Arc Pro B370 for integrated graphics with modern API coverage in a power-constrained environment.

Specification Differences

The two GPUs differ on nearly every measurable specification. The AMD Radeon Instinct MI300 uses a 5 nm process at TSMC, while the Intel Arc Pro B370 uses a 3 nm process at Intel. Transistor count for the MI300 is 153,000 million; for the Intel part, it is unknown. Die size for the MI300 is 1017 mm²; the Intel part has no recorded die size. Base clocks are 1000 MHz for the MI300 and 300 MHz for the Arc Pro B370. Boost clocks are 1700 MHz and 2400 MHz respectively. Memory capacity is 128 GB of HBM3 on the MI300 versus system shared memory on the Intel part. Memory bus width is 8192 bit versus system shared. Memory bandwidth is 6.55 TB/s versus system dependent. Shading units are 14,080 versus 1,280. TMUs are 880 versus 40. ROPs are 0 versus 20. The Intel part has 10 ray tracing cores; the MI300 has none recorded. Pixel rate is 0 MPixel/s on the MI300 and 48.00 GPixel/s on the Arc Pro B370. Texture rate is 1,496.0 GTexel/s versus 96.00 GTexel/s. FP32 compute is 47.87 TFLOPS versus 6.144 TFLOPS. FP16 compute is 383.0 TFLOPS (8:1) versus 12.29 TFLOPS (2:1). TDP is 600 W versus 25 W. Power connectors are 2x 8-pin versus none. Suggested PSU is 1000 W versus none. Bus interface is PCIe 5.0 x16 versus IGP. Display outputs are none versus portable device dependent. The MI300 measures 267 mm in length and 111 mm in height; the Intel part has no recorded dimensions. Release dates differ as well: the MI300 launched on 2023-01-03, the Arc Pro B370 on 2026-01-26.

Architecture Differences

The AMD Radeon Instinct MI300 is built on CDNA 3.0 architecture with the Aqua Vanjaram chip, part of the Radeon Instinct (MIx) generation. Its predecessor is FirePro Data Center. The Intel Arc Pro B370 uses Xe3-LPG architecture with the Panther Lake chip, belonging to the Arc Graphics-WM (Panther Lake) generation, with HD Graphics-WM as its predecessor. The MI300 has no recorded display outputs, no API support entries, and no ray tracing cores, aligning with its compute-focused CDNA design. The Arc Pro B370 carries full API support: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, plus 10 ray tracing cores, consistent with a graphics-oriented Xe3-LPG architecture. The MI300's memory subsystem is a dedicated 128 GB HBM3 pool with an 8192-bit bus, while the Arc Pro B370 shares system memory with no dedicated VRAM. The MI300's FP16 rate of 383.0 TFLOPS is delivered at an 8:1 ratio relative to FP32, indicating a heavy bias toward reduced-precision compute workloads. The Arc Pro B370's FP16 rate of 12.29 TFLOPS comes at a 2:1 ratio, showing a more balanced but far lower precision throughput. The MI300's transistor density is recorded at 150.4M per mm²; the Intel part has no density figure. The MI300 is a discrete accelerator with 2x 8-pin power connectors and a 1000 W suggested PSU. The Arc Pro B370 is an integrated processor with no power connectors and no suggested PSU, drawing 25 W. The MI300 has no ROPs and a 0 MPixel/s pixel rate, meaning it cannot rasterize output for display. The Arc Pro B370 has 20 ROPs and a 48.00 GPixel/s pixel rate, confirming its role in rendering graphics to a screen.

Where Each One Wins

The AMD Radeon Instinct MI300 wins decisively in raw compute density. Its 47.87 TFLOPS FP32 output is 7.8x higher than the Arc Pro B370's 6.144 TFLOPS. FP16 compute is even more lopsided: 383.0 TFLOPS versus 12.29 TFLOPS, a 31.2x advantage. Texture rate favors the MI300 at 1,496.0 GTexel/s versus 96.00 GTexel/s. Memory capacity and bandwidth are entirely different classes: 128 GB at 6.55 TB/s versus shared memory with system-dependent bandwidth. For workloads that scale with memory capacity, such as large model inference or high-performance computing, the MI300 has no recorded competition in this comparison. The Intel Arc Pro B370 wins in power efficiency and integration. Its 25 W TDP is 24x lower than the MI300's 600 W. It requires no power connectors and no suggested PSU, while the MI300 needs 2x 8-pin and a 1000 W PSU. The Arc Pro B370 also wins on boost clock, reaching 2400 MHz versus the MI300's 1700 MHz. It has 20 ROPs and a 48.00 GPixel/s pixel rate, while the MI300 has 0 ROPs and 0 MPixel/s, so the Intel part is the only one of the two capable of rasterized display output. The Arc Pro B370 supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4; the MI300 lists no API support. Ray tracing is available on the Arc Pro B370 with 10 dedicated cores; the MI300 has none recorded. The Intel part is also the only one with an active production status, while the MI300 has no recorded production status. For portable devices, the Arc Pro B370 fits because its bus interface is IGP and its dimensions are not recorded, meaning it integrates directly into a host processor. The MI300, at 267 mm long and 111 mm tall, is a standalone accelerator card. The data supports a clear split: the MI300 for compute-heavy data center tasks with massive memory and throughput, the Arc Pro B370 for integrated graphics in low-power portable systems with modern API and ray tracing support.