AMD Radeon 890M vs NVIDIA GeForce MX350 Comparison

# NVIDIA GeForce MX350 vs AMD Radeon 890M

The mobile GPU space has two very different philosophies on display here. The NVIDIA MX350 is a small, power-efficient discrete part from the Pascal generation, aimed at thin-and-light laptops where a little extra graphics muscle is needed. The AMD Radeon 890M is an integrated graphics processor (IGP) built into the Strix Point mobile APU, using the newer RDNA 3.5 architecture. The data in the database shows a clear split: the AMD part is dramatically faster in raw compute, while the NVIDIA part counters with a lower power draw and a simpler feature set. The numbers tell the story, and the architecture explains why.

Head-to-Head Benchmarks

The most direct comparison comes from the recorded benchmark scores. In the Geekbench OpenCL test, the NVIDIA scores 8,689 points while the AMD scores 37,254, a delta of -76.7%. That is a massive gap, with the AMD part almost 4.3 times faster in this compute-oriented workload.

The Vulkan test shows a similar, though slightly less extreme, result. The MX350 scores 13,077 while the Radeon 890M scores 40,808, a -68.0% difference. The AMD integrated graphics part is the clear winner here, and the margin is not small.

Looking at the broader picture, the average benchmark score across all recorded tests is 10,883 for the MX350 and 9,210 for the Radeon 890M. The MX350 is the faster part in the aggregate, though the specific compute tests show the 890M in a dominant position. These aggregate numbers, however, include a wider range of workloads, where the NVIDIA's software optimization may play a larger role.

Where Each One Wins

The data is unambiguous on the split. The AMD Radeon 890M wins both head-to-head compute benchmarks listed in the database: Geekbench OpenCL and Geekbench Vulkan. The OpenCL test is a pure compute test, and the RDNA 3.5 architecture of the 890M simply has far more shading units, 1,024 vs 640, and a much higher FP32 throughput of 5.939 TFLOPS vs 1.879 TFLOPS. The Geekbench Vulkan test is a more graphics-oriented workload, but the AMD part still wins by a wide margin, 40,808 to 13,077.

The NVIDIA MX350, on the other hand, has no recorded wins in the head-to-head comparison. Its strongest result relative to the competition is in the OpenCL test, where it is only 76.7% behind. It does, however, draw far less power (20W vs 15W TDP, though the AMD's power budget is for the entire APU, not just the GPU) and is a separate chip, which can be a consideration in some system designs.

Architecture Differences

The silicon is fundamentally different. The NVIDIA GeForce MX350 is built on the GP107S chip, a Pascal architecture part fabricated on a 14nm process at Samsung. It packs 3,300 million transistors on a 132 mm² die, which is a transistor density of 25.0M per mm². Its 640 shading units, 32 texture mapping units, and 16 ROPs are clocked at a base of 1354 MHz with a boost of 1468 MHz. The memory is 2GB of GDDR5 on a 64-bit bus, delivering 56.06 GB/s of bandwidth.

The AMD Radeon 890M is the integrated graphics processor inside Strix Point, using the RDNA 3.5 architecture on a 4nm TSMC process. It packs an enormous 34,000 million transistors onto a 233 mm² die, reaching a density of 145.9M per mm². Its 1024 shading units are laid out across 64 texture mapping units and 32 ROPs. Clocks are substantially different: a 400 MHz base with a 2900 MHz boost. The memory is system shared, so its bandwidth is system dependent, not a dedicated pool.

The core count and clock advantage are why the AMD part dominates the compute benchmarks. The MX350 cannot compete on raw shading unit count or on raw FP32 throughput. The NVIDIA part also has a much lower transistor density, which is a function of its older (Pascal) and its simpler, smaller design. The MX350 does have a 64-bit memory bus and discrete memory, which can be a benefit in some system designs, as it does not drain the main system memory or its bandwidth.

FAQ

Q: Which GPU has a newer architecture?

A: The AMD Radeon 890M is built on RDNA 3.5, while the NVIDIA MX350 uses the older Pascal architecture. The AMD is by a much more recent, larger design with a more modern feature set.

Q: Does the Radeon 890M have dedicated video memory?

A: No. Its memory is shared with the system, and marked as "System Shared" in the database. This is typical for an integrated graphics processor. The MX350, conversely, has 2GB of dedicated GDDR5 memory.

Q: How is the FP32 performance compared?

A: The AMD's FP32 is 5.939 TFLOPS, while the NVIDIA's is 1.879 TFLOPS. The AMD has more than 3 times the FP32 throughput, which shows in its dominant Geekbench OpenCL score.

Q: Which GPU has a higher pixel fill rate?

A: The AMD Radeon 890M is rated at 92.80 GPixel/s, while the NVIDIA MX350 is rated at 23.49 GPixel/s. The AMD has about 4 times the pixel throughput of the NVIDIA.

Q: What is the power draw for each part?

A: The AMD Radeon 890M has a TDP of 15W, which is the whole Strix Point APU. The NVIDIA MX350 is a discrete GPU with a TDP of 20W. The AMD's is lower, but it's important to note that the Intel TDP covers the entire APU, not just the graphics.

Q: What is the most important architectural difference?

A: The Radeon 890M includes 16 ray accelerators (RT cores) and has a native 1:1 FP16 ratio, meaning its FP16 and FP32 performance are identical at 5.939 TFLOPS. The MX350 has no RT cores, and its FP16 is a 1:64 ratio, giving a much lower FP16 throughput of 29.36 GFLOPS.

Specification Differences

The two parts differ in nearly every field recorded in the database. The process node is a clear one: NVIDIA is on 14nm, AMD on 4nm. The transistor count has a similar gap: 3,300 million vs 34,000 million, in favor of AMD. Die size also follows, 132 mm² vs 233 mm². The die density is 25.0M/mm² for NVIDIA and 145.9M/mm² for AMD.

Memory is a major split. The MX350 has 2GB GDDR5 on a 64-bit bus, with a dedicated 56.06 GB/s. The 890M has shared system memory, so its capacity and bandwidth are not part of the GPU but dependent on the host system's memory configuration.

In the API support department, the NVIDIA lists DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The AMD also lists DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both have current API support, but the AMD has a more recent DirectX feature level.

The Verdict

From the data alone, the AMD Radeon 890M is the better choice for pure graphics performance. In both head-to-head tests, it leads by a wide margin, and its architectural resources, with 1024 shading units, 64 TMUs, 32 ROPs, and 16 ray accelerators, are far beyond the NVIDIA. The MX350’s only advantages are its dedicated memory pool and its lower system-level power draw, which can matter for certain mobile design constraints.

For a user whose primary concern is raw GPU compute or graphics throughput, the Radeon 890M is the obvious choice based on the database. The NVIDIA MX350 uses less power and is a compact discrete piece, but there is no benchmark that shows it ahead. The recorded data points in one clear direction: for the same price class, AMD's integrated solution is by far the stronger graphics part.