Intel Arc A580 vs Intel Arc Pro B390 Comparison

Head-to-Head Benchmarks

The Intel Arc A580 and Intel Arc Pro B390 occupy very different positions in the database, and the recorded measurements reflect that divide. The Arc A580 holds an average benchmark score of 57,756, placing it in the 87th percentile of all GPUs tracked. The Arc Pro B390, by contrast, has no recorded benchmark scores, an average score of zero, and sits at the 50th percentile. The database currently contains no head-to-head benchmark entries between these two parts, so direct comparisons must be drawn from their individual performance profiles and architectural parameters.

The Arc A580 delivers 12.29 TFLOPS of FP32 compute, while the Arc Pro B390 delivers 7.680 TFLOPS. That represents a 60% advantage for the A580 in raw single-precision throughput. In FP16 workloads, the A580 produces 24.58 TFLOPS versus 15.36 TFLOPS for the B390, again a 60% lead. The A580 also far outstrips the B390 in texture and pixel throughput: 384.0 GTexel/s versus 120.0 GTexel/s, and 192.0 GPixel/s versus 60.00 GPixel/s, respectively.

The A580's recorded benchmarks show a 3DMark Steel Nomad DX12 score of 2,229, a Geekbench OpenCL score of 91,657, and a Geekbench Vulkan score of 79,381. These are substantial numbers for a desktop discrete GPU. The B390 has no such entries in the database, meaning its compute capabilities cannot be validated through the same test suite at this time.

Nearest rival data for the A580 places it within a tight cluster. The AMD Radeon RX 5600 OEM averages 58,085, which is 0.6% higher than the A580's average. The AMD Radeon RX 9070 GRE averages 57,367, 0.7% lower. The Intel Arc A570M averages 58,239, 0.8% higher, and the AMD Radeon RX 6950 XT averages 58,392, 1.1% higher. The A580 essentially trades blows with these parts, landing within roughly 1% of each. The B390 has no nearest rivals listed, and with no benchmark scores, it cannot be placed in any comparable grouping.

The clock behavior also differs sharply. The A580 runs at a 1700 MHz base clock and boosts to 2000 MHz, with memory running at 2000 MHz or 16 Gbps effective. The B390 has a very low 300 MHz base clock but boosts to 2500 MHz, which is 25% higher than the A580's boost. That higher boost ceiling does not compensate for the B390's much smaller execution resources, however.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel Arc A580 has an average benchmark score of 57,756 and sits in the 87th percentile of all GPUs. The Intel Arc Pro B390 has no recorded benchmark scores, an average score of zero, and sits in the 50th percentile.

Q: How do the two compare in raw FP32 compute performance?

A: The Arc A580 delivers 12.29 TFLOPS of FP32 compute, while the Arc Pro B390 delivers 7.680 TFLOPS. The A580 is approximately 60% faster in this metric.

Q: What are the memory configurations?

A: The Arc A580 uses 8 GB of GDDR6 memory on a 256-bit bus with 512.0 GB/s of bandwidth. The Arc Pro B390 uses system shared memory, with a system dependent bandwidth and a shared bus width.

Q: Do both GPUs support the same graphics APIs?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The API feature sets are identical in the database records.

Q: Which GPU has a higher boost clock?

A: The Arc Pro B390 boosts to 2500 MHz, which is higher than the Arc A580's 2000 MHz boost clock. However, the A580 has a much higher base clock at 1700 MHz versus 300 MHz for the B390.

Q: How does the Arc A580 compare to its nearest rivals?

A: The A580's average score of 57,756 is 0.6% below the AMD Radeon RX 5600 OEM, 0.7% above the AMD Radeon RX 9070 GRE, 0.8% below the Intel Arc A570M, and 1.1% below the AMD Radeon RX 6950 XT. All four rivals fall within roughly 1% of the A580's average.

Where Each One Wins

The Arc A580 wins in every category where the database records measurable performance. It has more shading units (3072 versus 1536), more texture mapping units (192 versus 48), more raster operation units (96 versus 24), and more ray tracing cores (24 versus 12). Its pixel rate of 192.0 GPixel/s is more than three times the B390's 60.00 GPixel/s, and its texture rate of 384.0 GTexel/s is exactly three times the B390's 120.0 GTexel/s.

For compute-heavy workloads, the A580's FP32 output of 12.29 TFLOPS and FP16 output of 24.58 TFLOPS give it a decisive edge over the B390's 7.680 and 15.36 TFLOPS, respectively. The A580 also has dedicated GDDR6 memory with 512.0 GB/s of bandwidth, while the B390 depends on system shared memory with bandwidth that varies by platform. This makes the A580 the appropriate choice for tasks that stress memory throughput, such as high-resolution rendering or large dataset processing.

The Arc Pro B390 wins in power efficiency and form factor. Its TDP is 80 W, less than half the A580's 175 W. The B390 is an integrated graphics processor (IGP) with no power connectors and no suggested PSU rating, while the A580 is a dual-slot card requiring two 8-pin power connectors and a 450 W suggested PSU. The B390 also has a higher boost clock at 2500 MHz versus 2000 MHz, which may help in short bursts of lightly threaded work where its smaller shader count is not the bottleneck.

The B390's integrated nature means it occupies no expansion slot and draws power from the host platform, making it suitable for compact or portable systems where a discrete card cannot fit. The A580's PCIe 4.0 x16 interface and dedicated memory make it the stronger option for sustained, memory-intensive workloads. The database shows no benchmark entries for the B390, so its real-world performance in these scenarios remains unmeasured.

Specification Differences

The two GPUs differ across nearly every recorded specification. The A580 is built on the DG2-512 chip using the Xe-HPG architecture, part of the Alchemist generation (Arc 5). The B390 uses the Panther Lake chip with the Xe3-LPG architecture, part of the Arc Graphics-WM (Panther Lake) generation. The A580 uses a 6 nm process from TSMC, while the B390 uses a 3 nm process from Intel. The A580 has 21,700 million transistors on a 406 mm² die with a transistor density of 53.4M per mm²; the B390's transistor count and die size are listed as unknown.

Clock speeds differ substantially. The A580 has a 1700 MHz base and 2000 MHz boost, with memory at 2000 MHz (16 Gbps effective). The B390 has a 300 MHz base and 2500 MHz boost, with system shared memory. Memory configurations are entirely different: 8 GB GDDR6 on a 256-bit bus with 512.0 GB/s bandwidth for the A580, versus system shared memory with system dependent bandwidth for the B390.

Execution resources are halved or more on the B390: 1536 shading units versus 3072, 48 TMUs versus 192, 24 ROPs versus 96, and 12 ray tracing cores versus 24. The pixel rate is 60.00 GPixel/s versus 192.0 GPixel/s, and the texture rate is 120.0 GTexel/s versus 384.0 GTexel/s. FP32 is 7.680 TFLOPS versus 12.29 TFLOPS, and FP16 is 15.36 TFLOPS versus 24.58 TFLOPS, both at a 2:1 ratio.

Power and physical specifications are also divergent. The A580 has a 175 W TDP, dual-slot width, two 8-pin power connectors, and a 450 W suggested PSU. The B390 has an 80 W TDP, IGP slot width, no power connectors, and no suggested PSU. The A580 uses a PCIe 4.0 x16 bus interface and outputs 1x HDMI 2.1 and 3x DisplayPort 2.0. The B390 uses an IGP bus interface and its display outputs are portable device dependent. The A580 released on 2023-10-09, while the B390 has a release date of 2026-01-26. The A580's predecessor is Xe Graphics and its successor is Battlemage; the B390's predecessor is HD Graphics-WM and it has no listed successor.

Architecture Differences

The architectural gap between these two parts is wide, reflecting different design goals and process technologies. The A580 uses the Xe-HPG architecture on a discrete DG2-512 chip, fabricated on TSMC's 6 nm process. It contains 21,700 million transistors across a 406 mm² die. The B390 uses the newer Xe3-LPG architecture on Intel's 3 nm process, built around the Panther Lake chip. Its transistor count and die size are not recorded in the database, but the 3 nm process indicates a much denser, more modern fabrication approach.

The A580 is a fully discrete GPU with its own 8 GB of GDDR6 memory, a 256-bit memory bus, and 512.0 GB/s of dedicated bandwidth. This memory subsystem is designed for sustained throughput in graphics and compute workloads. The B390 is an integrated graphics processor, sharing system memory with the host CPU. Its memory bandwidth is system dependent, meaning performance scales with the platform's memory configuration rather than a fixed specification.

The shader and fixed-function hardware is significantly scaled down on the B390. It has half the shading units, a quarter of the texture mapping units, a quarter of the raster operation units, and half the ray tracing cores compared to the A580. The B390's boost clock of 2500 MHz is higher, but with far fewer execution units, its peak throughput is much lower, as reflected in the FP32, texture, and pixel rate figures.

Both architectures support the same API levels: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The A580's Xe-HPG architecture is the older Alchemist design, while the B390's Xe3-LPG is a third-generation Xe architecture built for low-power integrated use. The B390's 80 W TDP and IGP form factor indicate a design optimized for power efficiency within a mobile or compact platform, whereas the A580's 175 W TDP and dual-slot cooler target desktop performance.

The database records no benchmarks for the B390, so its architectural advantages, such as the newer 3 nm process and higher boost clock, cannot be quantified against the A580's measured scores. The A580's performance data places it in the 87th percentile, within 1% of several AMD rivals, confirming its position as a competitive mid-range discrete GPU. The B390's 50th percentile ranking with no scores suggests it is not yet validated in the database, leaving its architectural potential unverified.