AMD and Nvidia let the cheap graphics card segment “deteriorate” so much that Intel took advantage of it quite elegantly. The latest Arc A580 in the Sparkle (Orc) design has a lot of imperfections that you might not get over, but at the same time it has settled into the top spot when it comes to price/performance ratio in gaming. With this primacy, it profiles itself as one of the leading adepts into low-budget builds.
| Parameters | Sparkle Arc A580 Orc | |
| Asus Dual RTX 4070 12G | ||
| Architecture | Xe HPG Alchemist | |
| Die | ACM-G10 | |
| Manufacturing node | 6 nm TSMC | |
| Die size | 406 mm² | |
| Transistor count | 21.7 bn. | |
| Compute units | 24 | |
| Shaders/CUDA cores | 3072 | |
| Base Clock | 1700 MHz | |
| Game Clock (AMD) | – | |
| Boost Clock | 2000 MHz | |
| RT units | 24 | |
| AI/tensor cores | 384 | |
| ROPs | 96 | |
| TMUs | 195 | |
| L2 Cache | 8 MB | |
| Infinity Cache | – | |
| Interface | PCIe 4.0 ×16 | |
| Multi-GPU interconnect | – | |
| Memory | 8 GB GDDR6 | |
| Memory clock (effective) | 16.0 GHz | |
| Memory bus | 256 bit | |
| Memory bandwidth | 512.0 GB/s | |
| Pixel fillrate | 192.0 Gpx/s | |
| Texture fillrate | 384.0 Gtx/s | |
| FLOPS (FP32) | 12.29 TFLOPS | |
| FLOPS (FP64) | emulated only | |
| FLOPS (FP16) | 24.58 TFLOPS | |
| AI/tensor TOPS (INT8) | – | |
| AI/tensor FLOPS (FP16) | – | |
| TDP | N/A | |
| Power connectors | 1× 8-pin | |
| Card lenght | 222 mm | |
| Card slots used | 41 mm | |
| Shader Model | 6.6 | |
| DirectX/Feature Level | DX 12 Ultimate (12_2) | |
| OpenGL | 4.6 | |
| Vulkan | 1.3 | |
| OpenCL | 3.0 | |
| CUDA | – | |
| Video encoder engine | Xe Media Engine ×2 | |
| Encoding formats | HEVC (10bit 4:4:4), H.264, VP9 (10bit, 4:4:4), AV1 (10bit) | |
| Encoding resolution | 8K | |
| Video decoder engine | Xe Media Engine ×2 | |
| Decoding formats | HEVC, H.264, VP9, AV1 | |
| Decoding resolution | 8K | |
| Max. Monitor resolution | 7680 × 4320 px | |
| HDMI | 1× 2.0b | |
| DisplayPort | 3× 2.0 (UHBR 10) | |
| USB-C | – | |
| MSRP | 205 EUR |
