今回は趣向を変えて、プログラミングテクニックに近い事を書きます。 最近のCPUには、SSEとかAVXとかNEONなどと呼ばれてる、要は幅の大きなレジスターを使って複数の計算を一回でやる機能…SIMDと呼ばれてる…があります。 これ、普通にソースコードを書い ...

It is important to note that DEBUG/RELEASE mode only affects the C kernel implementation, and thus the repeated runs for the assembly implementations can simply be looked at as more data to draw ...

この1つの命令で複数の演算器を動かすといいうやり方は「SIMD(Single Instruction stream Multiple Data stream)」と呼ばれる。 図2.6のようにレジスタと演算器のペアを4組並べ、1つの命令ユニットからの命令をすべての組に供給すれば、同じ命令で4つのデータを同時に ...

Is low-level programming a sin or a virtue? It depends. When programming for using vector processing on a modern processor, ideally I’d write some code in my favorite language and it would run as fast ...

To do so, the same kernel was programmed in C, non-SIMD x86-64, XMM-based SIMD AVX2, and YMM-based SIMD AVX2, all the while keeping track of their exact execution times to compare and contrast their ...

One of the fun parts of the ESP32-S3 microcontroller is that it got upgraded to the newer Cadence Xtensa LX7 processor core, which turns out to have a range of SIMD instructions that can help to ...