AMD Radeon Instinct MI300X vs Intel Arc Pro B390 Comparison

Head-to-Head Benchmarks

The recorded data shows no direct head-to-head benchmark results between the AMD Radeon Instinct MI300X and the Intel Arc Pro B390. The database contains zero benchmark entries for either part, and the wins column for each product is empty. Without measured scores, the comparison must rely entirely on the architectural and specification data captured in the database.

The MI300X delivers 81.72 TFLOPS of FP32 compute, which is approximately 10.6 times the 7.680 TFLOPS of the Arc Pro B390. In FP16, the MI300X reaches 653.7 TFLOPS using an 8:1 ratio, while the Arc Pro B390 achieves 15.36 TFLOPS with a 2:1 ratio. The MI300X FP16 figure is roughly 42.6 times higher. Texture rate follows the same pattern: the MI300X produces 2,553.6 GTexel/s versus 120.0 GTexel/s for the Arc Pro B390, a 21.3 times advantage.

The Intel part wins on pixel throughput. The Arc Pro B390 records 60.00 GPixel/s, while the MI300X shows 0 MPixel/s. The MI300X has no ROPs, which explains the zero pixel rate. The Arc Pro B390 includes 24 ROPs and 12 ray tracing cores, features entirely absent from the MI300X entry.

Memory capacity and bandwidth strongly favor the AMD accelerator. The MI300X carries 192 GB of HBM3 memory across an 8192-bit bus, delivering 10.3 TB/s of bandwidth. The Arc Pro B390 uses system shared memory, with the bus width, type, and bandwidth all listed as system dependent. The MI300X memory clock is 2525 MHz, effective 10.1 Gbps, whereas the Arc Pro B390 memory clock is listed as system shared.

Clock speeds show a mixed picture. The Arc Pro B390 has a higher boost clock at 2500 MHz versus 2100 MHz for the MI300X. The MI300X has a higher base clock at 1000 MHz versus 300 MHz for the Intel part. The Intel part is built on a 3 nm process from Intel, while the MI300X uses a 5 nm process from TSMC.

The Verdict

The data indicates two fundamentally different products serving different purposes. The AMD Radeon Instinct MI300X is a data center accelerator module with 19456 shading units, 1216 TMUs, and no display outputs. The Intel Arc Pro B390 is an integrated graphics processor with 1536 shading units, 48 TMUs, and 24 ROPs, designed for portable devices.

For compute workloads, the MI300X is the clear choice. Its FP32 throughput of 81.72 TFLOPS dwarfs the 7.680 TFLOPS of the Arc Pro B390. The 192 GB HBM3 pool with 10.3 TB/s bandwidth provides massive data access that system shared memory cannot match. The 750 W TDP and OAM module form factor indicate a server install base, not a client device.

For integrated graphics duty, the Arc Pro B390 is the only viable option in this pairing. It supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, while the MI300X lists no APIs at all. The 80 W TDP and IGP slot width fit mobile and compact systems. The 2500 MHz boost clock is higher than the MI300X boost of 2100 MHz, and the 60.00 GPixel/s pixel rate confirms it can drive displays.

The MI300X has no display outputs, so it cannot function as a primary graphics solution. The Arc Pro B390 has display outputs that are portable device dependent. Users needing graphics output must choose the Intel part. Users needing raw compute density must choose the AMD part.

Architecture Differences

The AMD Radeon Instinct MI300X uses the CDNA 3.0 architecture implemented on the Aqua Vanjaram chip. The Intel Arc Pro B390 uses the Xe3-LPG architecture on the Panther Lake chip. These are separate design philosophies: CDNA 3.0 targets accelerated computing, while Xe3-LPG targets integrated graphics in processors.

The MI300X is fabricated on a 5 nm process at TSMC with 153,000 million transistors on a 1017 mm² die. The transistor density works out to 150.4 million transistors per square millimeter. The Intel part uses a 3 nm process at Intel, but the transistor count and die size are recorded as unknown in the database.

The MI300X includes 19456 shading units and 1216 TMUs but zero ROPs. The Arc Pro B390 includes 1536 shading units, 48 TMUs, and 24 ROPs. The Intel part also has 12 ray tracing cores, a feature the MI300X does not list. Neither product lists tensor core counts.

The MI300X has no API support recorded for DirectX, OpenGL, or Vulkan. The Arc Pro B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. This confirms the MI300X is not intended for graphics API workloads, while the Arc Pro B390 carries full modern graphics API coverage.

The MI300X memory subsystem uses HBM3 with an 8192-bit bus and 10.3 TB/s bandwidth. The Arc Pro B390 uses system shared memory with system dependent bandwidth. The MI300X memory clock is 2525 MHz with 10.1 Gbps effective data rate, while the Intel part has no dedicated memory clock.

Specification Differences

The MI300X has a base clock of 1000 MHz and a boost clock of 2100 MHz. The Arc Pro B390 has a base clock of 300 MHz and a boost clock of 2500 MHz. The AMD part runs a higher base clock, the Intel part runs a higher boost clock.

Memory configuration is completely different. The MI300X has 192 GB of HBM3 on an 8192-bit bus with 10.3 TB/s bandwidth. The Arc Pro B390 has system shared memory, system shared type, system shared bus width, and system dependent bandwidth.

Shading resources differ by an order of magnitude. The MI300X has 19456 shading units, 1216 TMUs, and 0 ROPs. The Arc Pro B390 has 1536 shading units, 48 TMUs, and 24 ROPs. The MI300X has no ray tracing cores listed, while the Arc Pro B390 has 12.

Pixel and texture rates reflect these differences. The MI300X texture rate is 2,553.6 GTexel/s versus 120.0 GTexel/s for the Intel part. Pixel rate is 0 MPixel/s for the MI300X and 60.00 GPixel/s for the Arc Pro B390.

Power and form factor diverge sharply. The MI300X has a 750 W TDP with an OAM Module slot width and no power connectors listed. The suggested PSU is 1150 W. The Arc Pro B390 has an 80 W TDP with an IGP slot width, no power connectors, and no suggested PSU listed.

The MI300X uses a PCIe 5.0 x16 bus interface. The Arc Pro B390 uses an IGP bus interface. Display outputs are absent on the MI300X, while the Arc Pro B390 has portable device dependent outputs.

Release dates differ by roughly two years. The MI300X released on 2023-12-05, and the Arc Pro B390 released on 2026-01-26. The MI300X predecessor is FirePro Data Center, and the Arc Pro B390 predecessor is HD Graphics-WM. The Intel part has an active production status, while the AMD part has no production status recorded.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The AMD Radeon Instinct MI300X delivers 81.72 TFLOPS of FP32, compared to 7.680 TFLOPS for the Intel Arc Pro B390.

Q: Do these GPUs use the same memory type?

A: No. The MI300X uses 192 GB of HBM3 memory on an 8192-bit bus with 10.3 TB/s bandwidth. The Arc Pro B390 uses system shared memory with system dependent bandwidth.

Q: Which GPU supports ray tracing?

A: Only the Intel Arc Pro B390 lists ray tracing support, with 12 ray tracing cores. The AMD MI300X entry does not list any ray tracing cores.

Q: Can the AMD Radeon Instinct MI300X output video to a display?

A: No. The MI300X lists no display outputs, while the Arc Pro B390 has display outputs that are portable device dependent.

Q: What is the power consumption difference?

A: The MI300X has a 750 W TDP with a suggested PSU of 1150 W. The Arc Pro B390 has an 80 W TDP and no suggested PSU listed.

Q: Which GPU supports modern graphics APIs?

A: The Intel Arc Pro B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI300X has no API support recorded in the database.

Where Each One Wins

The AMD Radeon Instinct MI300X wins in every compute density metric. Its FP32 throughput of 81.72 TFLOPS, FP16 throughput of 653.7 TFLOPS, and texture rate of 2,553.6 GTexel/s place it in a performance class far above the Arc Pro B390. The 192 GB HBM3 memory with 10.3 TB/s bandwidth supports large models and datasets that system shared memory cannot accommodate. The 8192-bit bus provides exceptional memory parallelism.

The Intel Arc Pro B390 wins in graphics output and API compatibility. It is the only one of the two with display outputs, pixel rate, ROPs, and ray tracing cores. The 60.00 GPixel/s pixel rate and 24 ROPs allow actual rendering. The 2500 MHz boost clock is 400 MHz higher than the MI300X boost clock, which helps latency-sensitive graphics work. The 80 W TDP makes it suitable for integration into portable devices, while the 750 W MI300X requires a data center power envelope.

The MI300X is suited for compute-heavy server workloads such as accelerated analytics, large matrix operations, and high-bandwidth data processing. The Arc Pro B390 is suited for integrated graphics in mobile and compact systems where display output, API support, and low power draw are required.

The data shows no overlap in intended use cases. The MI300X is a compute accelerator with no graphics path. The Arc Pro B390 is an integrated GPU with no dedicated memory. Each product wins in its own domain, and the recorded specifications indicate no scenario where one could substitute for the other.