3 ms·
Why go so extreme when you can have fp12? Perhaps have 4 high bit exponent and low signed int8 mantissa. Or vice versa, 7 bit exponent, sign and 4 bit mantissa
by thriftwy 4y ago
Why go so extreme when you can have fp12? Perhaps have 4 high bit exponent and low signed int8 mantissa.
Or vice versa, 7 bit exponent, sign and 4 bit mantissa.
- jasonjmcghee 4y agoI think the general idea is to make use of SIMD, and generally the max size is 2^x. So if you're trying to multiply as many numbers as possible in, say, 64 bits, FP8 would get you 8x8 and FP16 would get you 4x4. FP12 would get you 5x5 with some unused space which would be a huge amount of extra work to implement for a 20% gain in efficiency.
- moonchild 4y agoYou can pack arrays of NPOT elements by breaking each element into POT components and transposing. So e.g. you could have one array of 8-bit mantissas and one array of 4-bit exponents.