AMD Instinct MI350P vs AMD Radeon 890M Comparison

Head-to-Head Benchmarks

The AMD Instinct MI350P and AMD Radeon 890M occupy entirely different performance classes, and the recorded data reflects that gap. The MI350P delivers 36.04 TFLOPS of FP32 compute, a figure 6.1 times higher than the 890M's 5.939 TFLOPS. That difference in raw floating-point throughput translates directly into the compute-heavy workloads the MI350P targets.

Texture processing shows a similar hierarchy. The MI350P achieves a texture rate of 1,126.4 GTexel/s, while the 890M manages 185.6 GTexel/s. That is a 6.1x advantage in favor of the accelerator card. Pixel throughput, however, tells a different story: the MI350P lists a pixel rate of 0 MPixel/s because it has no ROPs and no display outputs, while the 890M reaches 92.80 GPixel/s. The integrated GPU is designed to drive displays; the accelerator is not.

The 890M has actual benchmark scores in the database. Its highest recorded result is 40,808 in Geekbench Vulkan, followed by 37,254 in Geekbench OpenCL. Passmark G3D shows 8,076, and Passmark GPU Compute shows 4,157. The 890M's average benchmark score across all recorded tests is 9,210, placing it in the 45th percentile of all GPUs. Its nearest rival, the AMD Radeon Vega 8, scores 9,221, which is only 0.1% higher. The NVIDIA GeForce GTX 960 sits 0.7% ahead at 9,273, and the GTX 465 is 0.9% ahead at 9,294. The GTX 850M leads by 1% with a score of 9,302.

The MI350P has no benchmark entries, no average score, and no nearest rivals in the database. Its percentile rank of 50 is a placeholder rather than a measured performance position. That means the data cannot confirm actual workload performance for the MI350P, only its theoretical specifications.

DirectX and Vulkan API support further separates the two. The 890M supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI350P lists N/A for DirectX, OpenGL, and Vulkan. It is not a graphics card in the traditional sense; it is a compute accelerator with no graphics API exposure.

FAQ

Q: How does the FP32 compute performance compare between the two?

A: The MI350P delivers 36.04 TFLOPS of FP32, which is 6.1 times the 890M's 5.939 TFLOPS. Both also offer FP16 at a 1:1 ratio with their respective FP32 figures.

Q: What is the memory configuration for each product?

A: The MI350P uses 144 GB of HBM3e memory on an 8192-bit bus with 8.19 TB/s bandwidth. The 890M uses system shared memory with system dependent bandwidth, meaning its memory performance varies based on the host platform.

Q: Does the MI350P support graphics APIs?

A: No. The database lists DirectX, OpenGL, and Vulkan as N/A for the MI350P. The 890M supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4.

Q: How does the 890M compare to its nearest rivals in benchmark scores?

A: The 890M's average score of 9,210 is 0.1% below the Radeon Vega 8's 9,221, 0.7% below the GTX 960's 9,273, 0.9% below the GTX 465's 9,294, and 1% below the GTX 850M's 9,302.

Q: What process nodes do the two chips use?

A: The MI350P uses a 3 nm process, while the 890M uses a 4 nm process. Both are fabricated by TSMC.

Q: What are the power requirements?

A: The MI350P has a TDP of 600 W and requires a 1000 W suggested power supply with a single 16-pin connector. The 890M has a 15 W TDP and uses no power connectors.

Architecture Differences

The MI350P is built on CDNA 4.0 architecture, AMD's dedicated compute platform. Its chip, labeled MI350 128CU, contains 73,000 million transistors on a 1190 mm² die. The transistor density is 61.3 million per mm². The 890M uses RDNA 3.5 architecture, specifically the Strix Point chip, with 34,000 million transistors on a 233 mm² die. Its transistor density is significantly higher at 145.9 million per mm², reflecting the more compact integrated design.

The compute resources differ by an order of magnitude. The MI350P has 8,192 shading units, 512 texture mapping units, and zero ROPs. The 890M has 1,024 shading units, 64 TMUs, and 32 ROPs. The 890M also includes 16 ray tracing cores, while the MI350P lists none. Neither product lists tensor cores in the database.

Clock behavior also diverges. The MI350P runs at a 1000 MHz base clock and boosts to 2200 MHz. The 890M has a 400 MHz base clock but boosts to 2900 MHz, a higher absolute boost frequency despite the lower base. Memory clocks show a similar split: the MI350P's HBM3e runs at 2000 MHz with 8 Gbps effective transfer, while the 890M relies entirely on system shared memory.

The MI350P measures 267 mm in length, 111 mm in height, and 40 mm in width, occupying a dual-slot form factor. The 890M is an IGP with no physical dimensions recorded, designed to be integrated into a mobile processor package.

Specification Differences

The two products differ across nearly every specification field in the database. Process node: 3 nm for the MI350P versus 4 nm for the 890M. Transistor count: 73,000 million versus 34,000 million. Die size: 1190 mm² versus 233 mm². Transistor density: 61.3M per mm² versus 145.9M per mm².

Base clock: 1000 MHz versus 400 MHz. Boost clock: 2200 MHz versus 2900 MHz. Memory size: 144 GB versus system shared. Memory type: HBM3e versus system shared. Bus width: 8192 bit versus system shared. Bandwidth: 8.19 TB/s versus system dependent.

Shading units: 8192 versus 1024. TMUs: 512 versus 64. ROPs: 0 versus 32. Ray tracing cores: none versus 16. Pixel rate: 0 MPixel/s versus 92.80 GPixel/s. Texture rate: 1,126.4 GTexel/s versus 185.6 GTexel/s. FP32: 36.04 TFLOPS versus 5.939 TFLOPS. FP16: 36.04 TFLOPS versus 5.939 TFLOPS.

TDP: 600 W versus 15 W. Slot width: dual-slot versus IGP. Power connectors: 1x 16-pin versus none. Suggested PSU: 1000 W versus not listed. Bus interface: PCIe 5.0 x16 versus PCIe 4.0 x8. Display outputs: none versus portable device dependent.

API support: N/A for DirectX, OpenGL, and Vulkan on the MI350P, versus DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 on the 890M. Production status: not listed for the MI350P, active for the 890M. Release date: the MI350P is dated 2026-05-06, the 890M is dated 2024-07-14. Predecessor: Radeon Instinct for the MI350P, Navi II IGP for the 890M. The 890M has an average benchmark score of 9,210 and a 45th percentile rank; the MI350P has no recorded benchmarks and a 50th percentile placeholder.

The Verdict

The data defines two products with minimal overlap. The AMD Instinct MI350P is a dual-slot, 600 W accelerator with 144 GB of HBM3e memory, 8.19 TB/s of bandwidth, and 36.04 TFLOPS of FP32 compute. It has no display outputs, no graphics API support, and no benchmark scores in the database. It targets server-class compute workloads where raw throughput and memory capacity are the primary concerns.

The AMD Radeon 890M is a 15 W integrated GPU with 5.939 TFLOPS of FP32 compute, 16 ray tracing cores, and full support for DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4. Its recorded benchmark scores place it in the 45th percentile, within 1% of several discrete GPUs from previous generations, including the GTX 960 and GTX 465. It relies on system shared memory, so its bandwidth depends on the host platform.

For users selecting based on the database, the choice hinges on workload. The MI350P offers 6.1x the FP32 throughput and 6.1x the texture rate of the 890M, along with 144 GB of dedicated HBM3e memory. The 890M offers graphics output, ray tracing, modern API support, and actual measured benchmark results. The MI350P's 50th percentile rank is not based on tested performance, while the 890M's 45th percentile reflects real benchmark data. The MI350P suits compute environments without display requirements; the 890M suits mobile platforms that need integrated graphics with broad API compatibility.