Radeon R9 380

The Radeon R9 380 is based on the Antigua GPU, which is actually an update of the familiar Tonga chip (Radeon R9 285). It has not undergone any architectural changes. AMD has only optimized power consumption GPU – most likely due to a manufacturing process that allows for more crystals to be produced that can operate at a lower voltage. The resulting power reserve was spent on increasing the operating frequency from 918 to 970 MHz.

The Antigua chip, which is equipped with the Radeon R9 380, is similar in configuration of computing units (1792 shader ALUs, 112 texture units, 32 ROPs) to the “cut” Tahiti chip included in the Radeon R9 280, and before it the Radeon HD 7950, but differs in a narrower 256 -bit memory bus.

However, this version of Antigua/Tonga is not fully functional either. It is a larger ASIC than Tahiti (5 billion and 4,313 billion transistors, respectively), physically contains the same number of compute units (2048 shader ALUs, 128 texture units) and has the same memory bus configuration (6 controllers, total bit depth - 384 bits). Thus, AMD left room for a potential Radeon R9 380X built on a fully unlocked version GPU Antigua/Tonga.

Radeon R9 380 Specifications

Another change in the specifications of the Radeon R9 380 compared to the R9 285 is that the RAM bandwidth has been increased from 5500 to 5700 Gbps, and the capacity can reach up to 4 GB.