AMD is set to launch its next generation Radeon RX Vega GPU at SIGGRAPH 2017, which kicks off on the 30th of July. According to a new leak, however, the RX Vega custom models from AIB partners will not be available until a few weeks after the official launch.
The Sunnyvale, Calif.-based chipmaker will release reference cards initially, including both air- and water-cooled versions as it did with the Frontier Edition. This comes via the German technology site 3DCenter which has leaked the BIOS release schedule for Radeon RX Vega.
The Vega release schedule suggests that the bill of materials, that is the required parts to fab a Vega RX, was sent to board partners early June. The partners then started the engineering validation tests the same month until the beginning of July which means the working samples should be ready as we speak.
|BOM Release||Bill Of Materials Release||Beginning to end of June|
|EVT||Engineering Validation Test||End of June to beginning of July|
|DVT||Design Validation Test||Beginning until mid-July|
|WS||Working sample (all with engineering hardware and software)||In the middle of July|
|EMI Test||Electromagnetic Interference Test (EMC)||In the middle of July|
|PVT||Production Validation Test||End of July to beginning of August|
|PVT Sorting||?||Beginning of August|
|PPBIOS||Input of the final source BIOS from AMD||Beginning of August (final source BIOS to come on August 2 )|
|Ramp & MP||Start of mass production||Will be defined later, AMD date is still missing|
The product validation test (PVT) will be done before AMD releases the final BIOS on 2nd of August. The delay is to ensure that AMD could achieve the best possible efficiency at the BIOS level.
Once the Adaptive Voltage & Frequency Scaling (AVFS) is working properly, the source BIOS will be sent to board partners on which the RX Vega custom models will be finalized.
At this point, it’s unknown when the mass production of RX Vega is scheduled to happen, but you can expect it to begin by mid-September.
In terms of specs, the flagship RX Vega will be using HBM2 and running at a core clock of 1630MHz – a modest improvement of 30MHz over the Frontier Edition.
The HBM2 frequency of the card is set at 945MHz, delivering a total bandwidth of 484GB/s. It will also feature 12.5 TFLOPS of FP32 and ~25 TFLOPS of FP16 compute.
Furthermore, the RX Vega will come with fully optimised gaming drivers. If a previous 3DMark benchmark is to be believed, the card should outperform Nvidia’s GeForce GTX 1080 by around 15% but it will still be some ways off the pace of a GTX 1080 Ti.