In this post I’ll go through the disassembly (commented inline) of the following C++ snippet: #include <concepts> #include <coroutine> class task { public: class promise_type { public: class always_suspend {Continue reading
FP32 can be converted to BFloat16 simply by truncating the lower 16 bits. So we can do this with some bit operations. Internally two sets of FP32s (i.e., elements inContinue reading
Sort of an excersise.. The four LUTs for converting numeric values to ASCII chars is ugly. Besides, since both _mm256_shuffle_epi8 and _mm256_blendv_epi8 suffers from low throughput (1 CPI / 2Continue reading