
Graphics Core Next
AMD graphics processor architecture introduced in 2012
Graphics Core Next is a microarchitecture for graphics processors developed by AMD. Launched in 2012, it replaced the TeraScale design with a non-VLIW structure to improve parallel processing efficiency.
- Introduced in 2012 by AMD
- Uses 64-thread wavefronts for parallel processing
- Replaced the VLIW-based TeraScale architecture
- Powers PlayStation 4 and Xbox One consoles
- Supports asynchronous compute engines
What it is
Graphics Core Next, often abbreviated as GCN, is a series of graphics processor architectures created by AMD. It debuted in 2012 as a replacement for the earlier TeraScale design. The architecture focuses on handling parallel data tasks efficiently. Unlike previous designs that relied on complex compiler instructions, GCN uses a simpler model where hardware manages thread execution. This approach allows the processor to handle graphics and general computing workloads more effectively. It supports both 32-bit and 64-bit computing environments.
How it came about
AMD developed GCN to address limitations in its TeraScale architecture. TeraScale used Very Long Instruction Word technology, which required compilers to find independent operations statically. This often led to inefficient use of hardware resources. GCN removed these instruction slots entirely. Instead, each instruction encodes a single operation. The goal was to make performance depend on wavefront parallelism rather than compiler optimization. This shift aimed to improve occupancy and overall throughput for both graphics and compute tasks.
Inside the chip
The core building block of GCN is the Compute Unit. Each unit contains 64 shader processors and four texture mapping units. Threads are grouped into wavefronts of 64 threads. These groups execute in lock-step, sharing a single program counter. This method reduces power and area requirements. The architecture includes scalar and vector units. Vector units process 16 elements per cycle, while scalar units handle one. A scheduler assigns wavefronts to hardware units to hide memory latency. Asynchronous Compute Engines allow graphics and compute tasks to run concurrently.
How it was made
It details the internal logical structure, such as register files and cache sizes, but lacks physical production data. The focus remains on the architectural design rather than the silicon fabrication specifics.
Where it went
GCN powers a wide range of AMD products. It is found in discrete graphics cards like the Radeon HD 7000 and 400 series. The architecture also integrates into Accelerated Processing Units for desktops and laptops. Notably, customized GCN variants serve as the graphics heart of major gaming consoles. The PlayStation 4 uses the Liverpool APU, while the Xbox One family uses Durango. These chips handle both rendering and general compute tasks in consumer electronics.
Versions and family
GCN evolved through several generations. The first generation appeared in Southern Islands GPUs. The second generation added FreeSync support and appeared in Sea Islands chips and console APUs. The third generation, known as Volcanic and Pirate Islands, improved video encoding and memory compression. The fourth generation launched with the Radeon 400 series. Each iteration refined the compute units and added features like hardware schedulers and primitive discard accelerators.
Why it matters
GCN marked a significant shift in GPU design by moving away from VLIW dependencies. It enabled better support for heterogeneous computing, allowing CPUs and GPUs to share memory spaces. This architecture laid the groundwork for modern APIs like Vulkan and Mantle. Its adoption in major gaming consoles ensured widespread developer familiarity. The design principles influenced subsequent AMD architectures, including RDNA and CDNA, which split graphics and compute paths further.




Questions people ask about Graphics Core Next
What is a wavefront in GCN?
A wavefront is a group of 64 threads that execute together in lock-step. It is the basic unit of scheduling in GCN. All threads in a wavefront share a single program counter and process instructions simultaneously. This helps hide memory latency by switching to other wavefronts when one waits for data.
How does GCN differ from TeraScale?
TeraScale used VLIW instructions requiring static compiler optimization. GCN removes instruction slots, encoding single operations per instruction. This shifts performance dependence from compiler skill to hardware-managed wavefront parallelism, improving efficiency and occupancy for diverse workloads.
Which consoles use GCN architecture?
The PlayStation 4 uses a custom GCN-based APU called Liverpool. The Xbox One and Xbox One S use a similar custom APU named Durango. Both integrate GCN compute units to handle graphics and general processing tasks within the console hardware.
Good reading and sources
- Official websiteOfficial site
- Checking your browserAnandTech, Article
- AMD Radeon RX 480 8GB ReviewTom's Hardware, Article
The facts on this page come from the Wikipedia article "Graphics Core Next", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.
Like Graphics Core Next
All graphics processorsMore from AMD
All of them
🇺🇸 United StatesRadeon HD 7000 series
AMD graphics processor family using 28 nm process and GCN architectureGPU · GPU architecture · 2012