AMD Radeon 400 series leaks and rumors

Today, we are going to take a look at some AMD Radeon 400 series leaks and rumors that have surfaced in the past few days.

AMD Polaris coming mid-year?

Starting with the most spicy information, Robert Hallock, the head of Global Technical Marketing at AMD, confirmed what we’ve been assuming for the past few weeks. Polaris architecture will debut ‘mid-year’, which could mean either June or July. AMD is possibly aiming at June announcement to keep year-to-year cycle with new series. However we are probably looking at paper launch only, as actual availability of Polaris cards will depend on more factors like HBM2/GDDR5X availability or 14nm FinFET production. Despite that, the schedule for Radeon 400 series would look like this:

• AMD Radeon M400 series : April 2016
• AMD Radeon 400 series : ‘Mid-year’

AMD Polaris 11 to feature 8.6 billion transistors?

This is coming from 3DCenter where people are trying to figure out what the tweets from AMD marketing team actually mean. This is a follow up to their previous conversation on Twitter about ‘spicing up’ GDC 2016, thus references to peppers.

Of course assuming this is anyhow related to transistor count is in my opinion an exaggeration. By reading in-between the lines we could get an impression that the Capsaicin (chilli peppers active component) chemical notation mentioned in the tweet could somehow refer to 8.6 billion transistors.

This would be almost on par with Fiji’s 8.9 billion. However ‘Project F’ leak suggested that Polaris 11 would have 232mm² die area, which therefore be 2,48 denser than Fiji. Is it possible with 14nm FinFET? Hard to say.

• Fiji : 14,93 million transistors per mm²
• Polaris 11 : 37,06 million transistors per mm²

 AMD to give away R9 Fury X2 powered computers at GDC?

Yet another tweet, but this time from the brightest star (aka Polaris) in AMD marketing team, Roy Taylor himself. The Radeon-themed systems in this photo are likely the same systems that were announced by Taylor at VRLA Winter Expo. Back then he said:

We promised you we would take two of our highest end GPUs and put it inside that tiny box and if you go downstairs we actually have a demonstration of a dual GPU, 12 TeraFlops, fastest GPU solution in the world, inside of Tiki.

which basically confirms that these computers are ‘dual-GPU’ systems with 12 TFLOPs computing power. It also means that Fury X2 is not running at the same clock speed as Fury X2, but something close to 732 MHz per GPU, as long as these are Fiji chips with 4096 Stream Processors.