Single Cycle Throughput Math (SCM)

Customized subsets for cutting-edge applications.
1 math function = 1 instruction (no interim memory access)
Each function has its own fixed latency.
Pipelined to deliver GFFPS (Giga Floating-point Functions Per Second) throughput per core Silicon proven with RISC-V dual core CPU: 1 GHz & 65 mW / SCM core (verified by MPW silicon from TSMC’s 28nm process)

SCM performs OpenCL math functions

10x, 2x, acos, acosh, acospi, asin, asinh, asinpi, atan, atan2, atan2pi, atanh, atanpi, cbrt, ceil, copysign, cos, cosh, cospi, erf, erfc, ex, exm1, fabs, fadd,

log, log10, log1p, log2, logb, maxmag, minmag, modf, nan, nextafter, nfmsub, nop, pow, pown, powr, rcp, remainder, remquo, rint, rootn, round, rsqrt, sigmoid, sin, sinh, sinpi, sqrt, tan, tanh, tanpi, tgamma, and trunc,

FCLASS_S, FCVT_S_W, FCVT_S_WU, FCVT_W_S, FCVT_WU_S, fdim, fdiv, FEQ_S, FLE_S, floor, FLT_S, fmadd, fmax, fmin, fmod, fmsub, fmul, FMV_S_W, FMV_W_S, fnmadd, fract, frexp, FSGNJ_S, FSGNJN_S, FSGNJX_S, fsub, hypot, ilogb, ldexp, lgamma, lgammar