A new report has surfaced suggesting that AMD has managed to create a new revision of the Polaris GPU that is up to 50 percent more efficient, which should allow AMD to create graphics card with both lower TDP or higher GPU clock speeds.

According to details from a new report coming from Wccftech.com, the gain comes from improvements in the 14nm metal masking layers used on Polaris architecture as well as enhanced binning process, which allowed AMD to decrease the TDP of existing Polaris 10 GPU from 150W down to 95W while Polaris 11 GPU dropped from 75W to less than 50W.

This allowed AMD to also increase compute performance of the Polaris 11 GPU to 2.5TFLOPs by increasing the GPU clock speeds which is a decent improvement compared to the original 2.1TFLOPs.

The report suggests that these new revisions are already shipping in the embedded market and AMD will most likely use the same GPUs in the mobile market in order to compete with Nvidia's new mobile Pascal lineup, while these same revised GPUs could also come to the desktop market with a previously detailed Radeon RX 4X5 series.

Earlier slide showed that AMD planned for a Polaris GPU refresh for the new RX 4X5 series which should have higher clock speeds and improved GPU efficiency.

Unfortunately, AMD has not officially confirmed these details but this will give AMD even more room to compete with Nvidia in both desktop and mobile mainstream market.



Source: Overclock3D.net.