Intel Arc A570M vs Intel Arc B390 Comparison

Head-to-Head Benchmarks

The recorded data places the Intel Arc A570M and Intel Arc B390 in entirely different performance strata. The Arc A570M, based on the DG2-256 chip, posts a Geekbench OpenCL score of 58,239. This result places it in the 88th percentile of all GPUs in the database. The Arc B390, built on the Panther Lake chip, records a 3DMark Steel Nomad DX12 score of 1,482. That score puts it in the 9th percentile of all GPUs. Direct cross-referencing between the two is complicated by the different benchmark suites used, but the percentile gap is decisive: 88th versus 9th indicates a fundamental difference in compute capability.

The Arc A570M's nearest rivals in the database include the AMD Radeon RX 6950 XT, which scores 58,392, a delta of -0.3%. The NVIDIA P102-100 scores 58,528, a delta of -0.5%. The AMD Radeon PRO V710 scores 58,657, a delta of -0.7%. On the other side, the AMD Radeon RX 5600 OEM scores 58,085, a delta of +0.3%. These margins are tight, all within 1%. The data shows the Arc A570M sits in a dense cluster of desktop-class GPUs, trading places with each rival by less than a single percentage point.

The Arc B390's nearest rivals are much lower in absolute score. The NVIDIA GeForce GT 520MX scores 1,463, a delta of +1.3%. The NVIDIA GeForce 800M scores 1,460, a delta of +1.5%. The NVIDIA GeForce GT 625 OEM scores 1,446, a delta of +2.5%. The NVIDIA GeForce GT 710 scores 1,443, a delta of +2.7%. The Arc B390 leads each of these by a small margin, but the comparison set itself reveals its position: these are entry-level integrated or very old discrete parts. The Arc B390 is ahead of them, yet the absolute scores are orders of magnitude below the Arc A570M's result.

The FP32 throughput figures reinforce this split. The Arc A570M delivers 5.325 TFLOPS, while the Arc B390 delivers 7.680 TFLOPS. Interestingly, the Arc B390 shows a higher raw FP32 number despite its much lower benchmark score. This discrepancy suggests that the benchmark results reflect more than raw shader math; memory bandwidth, driver state, and workload characteristics all play a role. The Arc A570M has 224.0 GB/s of dedicated bandwidth on an 8 GB GDDR6 pool, while the Arc B390 relies on system shared memory with bandwidth described as system dependent. The Arc B390's FP32 advantage does not translate into a benchmark advantage in the recorded tests.

Pixel and texture rates also differ. The Arc A570M lists 83.20 GPixel/s and 166.4 GTexel/s. The Arc B390 lists 60.00 GPixel/s and 120.0 GTexel/s. The Arc A570M leads in both rasterization throughput metrics. The Arc B390 counters with a higher boost clock, 2500 MHz versus 1300 MHz, and a higher FP32 figure, but the recorded benchmark data does not reflect that theoretical advantage.

The wins are not shared. The head-to-head benchmark array in the database is empty, and the win counters show 0 for both parts. The available evidence points to the Arc A570M as the stronger performer in the recorded OpenCL test, while the Arc B390's only recorded result is a low-end 3DMark run. No single benchmark pits both parts against each other directly, so the comparison relies on percentile placement and nearest-rival deltas.

The Verdict

The data indicates that the Intel Arc A570M is the higher-performing part in absolute terms. Its 88th percentile ranking and OpenCL score of 58,239 place it alongside desktop-class GPUs such as the AMD Radeon RX 6950 XT and NVIDIA P102-100. The Arc B390's 9th percentile ranking and 1,482 Steel Nomad score place it near the NVIDIA GeForce GT 520MX and GT 710. Any user selecting between these two based on recorded performance data should choose the Arc A570M without hesitation.

The Arc B390 does have advantages in the specification sheet. It uses a newer 3 nm process node from Intel, while the Arc A570M uses TSMC's 6 nm node. The Arc B390 has a higher boost clock of 2500 MHz, higher FP32 throughput at 7.680 TFLOPS, and higher FP16 throughput at 15.36 TFLOPS. It also draws a slightly higher TDP of 80 W versus 75 W. These are real differences, but the benchmark data does not show them translating into winning results. The Arc B390's shared memory architecture and low percentile score suggest that its compute resources are not being utilized effectively in the recorded test.

The Arc A570M also leads in memory configuration. It has 8 GB of dedicated GDDR6 on a 128-bit bus with 224.0 GB/s bandwidth. The Arc B390 uses system shared memory with system dependent bandwidth. For workloads that depend on memory bandwidth, the Arc A570M has a clear structural advantage. The Arc B390's higher FP32 rate cannot compensate for the lack of dedicated memory bandwidth in the recorded benchmark.

The production status for both is listed as Active. The Arc A570M released on 2023-07-31, while the Arc B390 released on 2026-01-26. The newer part is not the faster one in this comparison. The Arc B390 belongs to the Arc Graphics-M (Panther Lake) generation with Xe3-LPG architecture, while the Arc A570M belongs to the Alchemist (Arc 5 Mobile) generation with Xe-HPG architecture. Architecture generation alone does not determine performance.

Architecture Differences

The Intel Arc A570M uses the DG2-256 chip built on the Xe-HPG architecture. This is the Alchemist generation, specifically Arc 5 Mobile. The process node is 6 nm at TSMC. The die size is 269 mm², and the transistor count is 11,500 million. Transistor density calculates to 42.8 million transistors per square millimeter.

The Intel Arc B390 uses the Panther Lake chip built on the Xe3-LPG architecture. This is the Arc Graphics-M generation for Panther Lake. The process node is 3 nm at Intel. The die size and transistor count are both listed as unknown, so no density figure is available.

The Arc A570M has 2048 shading units, 128 texture mapping units, and 64 raster output units. It includes 16 ray tracing cores. The Arc B390 has 1536 shading units, 48 texture mapping units, and 24 raster output units. It includes 12 ray tracing cores. The Arc A570M has more of each resource type.

The Arc A570M uses dedicated GDDR6 memory: 8 GB on a 128-bit bus with 224.0 GB/s bandwidth. The Arc B390 uses system shared memory with no dedicated pool, no fixed bus width, and bandwidth that depends on the host system. This is a fundamental architectural difference. The Arc A570M's memory interface is designed for consistent bandwidth, while the Arc B390's performance will vary with the platform it is installed in.

The Arc A570M connects via PCIe 4.0 x8. The Arc B390 uses IGP as its bus interface, meaning it is integrated into the processor package. Both are listed as IGP slot width, but the Arc A570M's PCIe interface suggests it can also operate as a discrete-style mobile part. The Arc B390's IGP interface ties it directly to the Panther Lake processor.

Both parts support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The API feature sets are identical in the database. The difference lies in the underlying hardware resources and memory architecture.

Specification Differences

The clock speeds differ significantly. The Arc A570M has a base clock of 900 MHz and a boost clock of 1300 MHz. The Arc B390 has a base clock of 300 MHz and a boost clock of 2500 MHz. The Arc B390's boost clock is nearly double the Arc A570M's boost clock, but its base clock is three times lower.

Memory differs completely. The Arc A570M has 8 GB GDDR6 on a 128-bit bus with 224.0 GB/s bandwidth. The Arc B390 has system shared memory, system shared type, system shared bus width, and system dependent bandwidth. The memory clock for the Arc A570M is 1750 MHz with 14 Gbps effective. The Arc B390 has no dedicated memory clock listed.

Compute unit counts differ. The Arc A570M has 2048 shading units, 128 TMUs, and 64 ROPs. The Arc B390 has 1536 shading units, 48 TMUs, and 24 ROPs. The Arc A570M leads in every count. Ray tracing cores are 16 versus 12.

Throughput numbers go in opposite directions. The Arc A570M has 83.20 GPixel/s and 166.4 GTexel/s. The Arc B390 has 60.00 GPixel/s and 120.0 GTexel/s. The Arc B390 wins on FP32: 7.680 TFLOPS versus 5.325 TFLOPS. It also wins on FP16: 15.36 TFLOPS versus 10.65 TFLOPS. Both FP16 figures are listed at a 2:1 ratio.

Power draw is close. The Arc A570M is 75 W, the Arc B390 is 80 W. The Arc A570M has no power connector listed, while the Arc B390 lists None. The Arc A570M is PCIe 4.0 x8, the Arc B390 is IGP. Both have portable device dependent display outputs. Both are Active production status.

FAQ

Q: Which GPU has the higher benchmark percentile?

A: The Intel Arc A570M sits in the 88th percentile of all GPUs. The Intel Arc B390 sits in the 9th percentile.

Q: How does the Arc A570M compare to its nearest rival?

A: The Arc A570M scores 58,239 in Geekbench OpenCL. The AMD Radeon RX 6950 XT scores 58,392, a delta of -0.3%. The AMD Radeon RX 5600 OEM scores 58,085, a delta of +0.3%. The Arc A570M is within 1% of all four listed rivals.

Q: Does the Arc B390 have any performance advantage over the Arc A570M?

A: The Arc B390 has a higher FP32 throughput of 7.680 TFLOPS versus 5.325 TFLOPS, and a higher FP16 throughput of 15.36 TFLOPS versus 10.65 TFLOPS. Its boost clock is 2500 MHz versus 1300 MHz. However, its recorded benchmark score of 1,482 in 3DMark Steel Nomad DX12 places it in the 9th percentile.

Q: What memory configurations do the two GPUs use?

A: The Arc A570M uses 8 GB of GDDR6 on a 128-bit bus with 224.0 GB/s bandwidth. The Arc B390 uses system shared memory with system dependent bandwidth.

Q: Which GPU has more shading units?

A: The Arc A570M has 2048 shading units. The Arc B390 has 1536 shading units. The Arc A570M also has more TMUs (128 versus 48), more ROPs (64 versus 24), and more ray tracing cores (16 versus 12).

Q: What are the process nodes for each GPU?

A: The Arc A570M uses a 6 nm process at TSMC. The Arc B390 uses a 3 nm process at Intel.

Where Each One Wins

The Intel Arc A570M wins in rasterization throughput. Its 83.20 GPixel/s pixel rate and 166.4 GTexel/s texture rate exceed the Arc B390's 60.00 GPixel/s and 120.0 GTexel/s. The Arc A570M also wins in memory bandwidth with 224.0 GB/s of dedicated GDDR6 bandwidth, versus the Arc B390's system dependent shared memory.

The Arc A570M wins in the recorded benchmark. Its OpenCL score of 58,239 and 88th percentile placement are far above the Arc B390's 1,482 Steel Nomad score and 9th percentile. The nearest rival comparison confirms this: the Arc A570M trades places with desktop-class GPUs like the AMD Radeon RX 6950 XT, while the Arc B390 trades places with the NVIDIA GeForce GT 520MX.

The Arc B390 wins on paper in raw compute density. Its FP32 output of 7.680 TFLOPS is 44% higher than the Arc A570M's 5.325 TFLOPS. Its FP16 output of 15.36 TFLOPS is also higher. Its boost clock of 2500 MHz is substantially higher. The 3 nm Intel process node is newer than the 6 nm TSMC node. These specification advantages do not appear in the benchmark results.

The Arc B390 also wins on integration. It uses an IGP bus interface with no power connector, suited for a processor-integrated design. The Arc A570M uses PCIe 4.0 x8, which requires a dedicated connection. The Arc B390's power draw of 80 W is only 5 W higher than the Arc A570M's 75 W, so the integration advantage does not come with a large power penalty.

For workloads that depend on raw shader math and are not memory bound, the Arc B390's higher FP32 and FP16 figures could be relevant. The recorded data does not include such a workload, so this remains a specification-level observation. For workloads that depend on memory bandwidth, pixel fill, or texture fill, the Arc A570M has the clear advantage in the recorded specifications.

The database shows no head-to-head benchmark wins for either part, and the win counters are zero for both. The choice between these two GPUs, based strictly on recorded data, favors the Arc A570M for benchmark performance and memory configuration. The Arc B390 offers a newer process node, higher clocks, and higher theoretical compute throughput, but its only recorded benchmark result lands it in the bottom 9th percentile of all GPUs.