facebook rss twitter

AMD Vega 20 benchmark results spotted

by Mark Tyson on 30 April 2018, 11:30

Tags: AMD (NYSE:AMD)

Quick Link: HEXUS.net/qads6f

Add to My Vault: x

Some very early benchmarks for an AMD Radeon Vega 20 based graphics card have bobbed to the surface of the 3DMark online result browser. VideoCardz shared the story but hasn’t provided a link to the online results, just some screenshots.

The source supposes that an AMD employee was testing an engineering sample Vega 20 based card and forgot to turn off online result validation – presenting us with this info-treat. Earlier in the year the device ID of 66A0 was confirmed as belonging to Vega 20 thanks to some Linux driver submissions.

As well as the benchmark result it is revealed, in supplementary sys-info pages, that the Vega 20 is fitted with 32GB of HBM2. Clock speeds are not reported correctly, asserts VideoCardz – so it isn’t really running at 1.0GHz. In brief, Vega 20 is expected to be a die shrink of Vega 10.

Remember Radeon Instinct Vega 7nm was announced back at CES 2018 in January. It is thought that it could be an engineering sample of this card that has been put through its paces in 3DMark11 to provide the above results. The Vega 20 powered Instinct is aimed at tasks with a focus on AI and machine learning and neither it, nor the early drivers, are going to be optimised for things like gaming benchmarks.

Please add a pinch of salt to the above.



HEXUS Forums :: 6 Comments

Login with Forum Account

Don't have an account? Register today!
Will be looking to buy a card this year as fury struggling in some games at 1600p (16:10 screen) could be a choice time will tell
….so they have developed their own version of TENSOR core?
32GB? I thought there was a shortage of that stuff.
WildW_UK
32GB? I thought there was a shortage of that stuff.

It's aiming to replace the FE vega, which had 16GB. Consumer GPUs will have less RAM
If it's an early sample then they are often clocked far slower than the final plus uses unoptimised drivers that contain more debug code turned on in the background. I very much doubt the above numbers represent the performance of the final product.