Lines Matching defs:VLEN

1550   // VLEN=32 support is incomplete.
2315   // If we're extracting only data from the first VLEN bits of the source
2318   // TODO: We can generalize this when the exact VLEN is known.
2323   // TODO: For sizes which aren't multiples of VLEN sizes, this may not be
2705     // We prefer to use LMUL=1 for VLEN sized types. Use fractional lmuls for
2770   // If we know the exact VLEN, and our VL is exactly equal to VLMAX,
4070   // If we're compiling for an exact VLEN value, we can split our work per
4993 // If compiling with an exactly known VLEN, see if we can split a
5194   // For exact VLEN m2 or greater, try to split to m1 operations if we
6445       report_fatal_error("Support for VLEN==32 is incomplete.");
8484     // If we're compiling for an exact VLEN value, we can always perform
8488     if (auto VLEN = Subtarget.getRealVLen();
8489         VLEN && ContainerVT.bitsGT(M1VT)) {
8491       unsigned ElemsPerVReg = *VLEN / ElemVT.getFixedSizeInBits();
8692   // If we're compiling for an exact VLEN value and we have a known
8734   // O(n^2) / (VLEN/ETYPE) work.  On the surface, going through the stack
8950 // (vscale * VF). Where vscale is defined as VLEN/RVVBitsPerBlock. The
9071 // LMUL * VLEN should be greater than or equal to EGS * SEW
10060   // If the subvector vector is a fixed-length type and we don't know VLEN
10303   // If the subvector vector is a fixed-length type and we don't know VLEN
10770   // If we know the exact VLEN and our fixed length vector completely fills
10833   // If we know the exact VLEN and our fixed length vector completely fills
17901     // We can use the minimum and maximum VLEN values to bound VLENB.  We
17902     // know VLEN must be a power of two.
21177   // We define vscale to be VLEN/RVVBitsPerBlock.  VLEN is always a power