Notebookcheck Logo

Intel Arc A380

Intel Arc A380

The Intel Arc A380 is a low-cost single-seed desktop GPU and is manufactured at TSMC using the 6 nm process. The graphics processor has the internal designation DG2-128 and combines 1,024 shader units. In terms of VRAM, you still get 6 GB GDDR6, which are connected via a 96-bit interface (15.5 Gbps, 186 GB/s). This makes the Intel Arc A380 the fourth fastest graphics card model from Intel.

Features

As the Intel Arc A380 is also based on the Alchemist architecture, the small model also has the same features as the faster models of the current Arc models. Nevertheless, there are significant performance limitations compared to the faster models. The most useful feature is certainly AV1 support, which enables encoding and decoding on the hardware side.

Performance

The performance of the Intel Arc A380 cannot keep up with current graphics cards from AMD and Nvidia. In view of this, this graphics card is not recommended exclusively for gaming use.

Power consumption

The power consumption of the Intel Arc A380 is estimated by the manufacturer at 75 watts (TGP). In terms of efficiency, the Intel Arc A380 is not convincing due to its low 3D performance. In the internal duel, the Intel Arc A380 is slightly more efficient than an Intel Arc A580 or an Intel Arc A750.

Arc Series

Arc B580 compare 2560 @ 2.67 - 2.67 GHz192 Bit @ 19000 MHz
Arc A770 compare 4096 @ 2.1 - 2.4 GHz256 Bit @ 16000 MHz
Arc A750 compare 3584 @ 2.05 - 2.4 GHz256 Bit @ 16000 MHz
Arc A580 compare 3072 @ 1.7 - 2 GHz256 Bit @ 16000 MHz
A770M compare 4096 @ 1.65 GHz256 Bit @ 16000 MHz
A730M compare 3072 @ 1.1 GHz192 Bit @ 14000 MHz
A570M compare 16 @ 0.9 - 1.3 GHz128 Bit @ 14000 MHz
A550M compare 16 @ 0.9 - 2.01 GHz128 Bit @ 14000 MHz
A530M compare 12 @ 0.9 - 1.3 GHz128 Bit @ 14000 MHz
A370M compare 8 @ 1.55 GHz64 Bit @ 14000 MHz
Arc A380 1024 @ 2 GHz96 Bit @ 15500 MHz
A310 compare 6 @ 2 GHz64 Bit @ 15500 MHz
A350M compare 6 @ 1.15 GHz64 Bit @ 14000 MHz
CodenameDG2-128
ArchitectureXe HPG
Pipelines1024 - unified
TMUs64
ROPs32
Tensor / AI Cores128
Core Speed2000 MHz
Shader Speed2050 MHz
Theoretical Performance8397 TFLOPS FP16, 4194 TFLOPS FP32
CacheL2: 4 MB
Memory Speed15500 effective = 1937 MHz
Memory Bus Width96 Bit
Memory TypeGDDR6
Max. Amount of Memory6 GB
Shared Memoryno
Memory Bandwidth186 GB/s
APIDirectX 12_2, Shader 6.6, OpenGL 4.6, OpenCL 3.0, Vulkan 1.3
Power Consumption75 Watt
Transistor Count7.2 Billion
Die Size 157 mm²
technology6 nm
PCIe4 8x
Displays4 Displays (max.), HDMI 2.1, DisplayPort 2.0
Date of Announcement14.07.2022

Benchmarks

Performance Rating - 3DMark 11 + Fire Strike + Time Spy
3.8 pt (11%)
3DMark Vantage
3DM Vant. Perf. total +
3DM Vant. Perf. GPU no PhysX +
3DMark 06 3DMark 06 - Standard 1280x1024 +
SPECviewperf 12
specvp12 sw-03 +
specvp12 snx-02 +
specvp12 showcase-01 +
specvp12 mediacal-01 +
specvp12 maya-04 +
specvp12 energy-01 +
specvp12 creo-01 +
specvp12 catia-04 +
specvp12 3dsmax-05 +