1
Tom's Hardware: "Intel's next-gen Arrow Lake GPU will have new Xe-LPG Plus Architecture with XMX"
(www.tomshardware.com)
this post was submitted on 24 Nov 2023
1 points (100.0% liked)
Hardware
47 readers
1 users here now
A place for quality hardware news, reviews, and intelligent discussion.
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I have no idea why this is important or really what it even is but Apple had a pretty video about it in their WWDC catalog so I guess it's a trend now.
Standard float16 uses 1bit sign + 5bit exponent + 10bit fraction.
bfloat16 uses 1bit sign + 8bit exponent + 7bit fraction.
bfloat16 basically gives the same exponent precision as a standard float32. But most neural networks don't require a huge fraction range. So bfloat16 gives you the possibility of executing 2x 8bit NP FLOPs vs using a float32 to do the same 1x8bit NP FLOP.
Having the ALU support this format allows for the scheduler to pack 4xbfloat16 that can be executed in parallel in a standard 64bit ALU. So basically you double or quadruple the 8bit NP FLOPs that you would get from using traditional float16/32 representations.
Makes sense. Thanks for that!