Lines Matching defs:VLEN
1606 // VLEN=32 support is incomplete.
2231 // If we're extracting only data from the first VLEN bits of the source
2234 // TODO: We can generalize this when the exact VLEN is known.
2241 // TODO: For sizes which aren't multiples of VLEN sizes, this may not be
2692 // We prefer to use LMUL=1 for VLEN sized types. Use fractional lmuls for
4152 // If we're compiling for an exact VLEN value, we can split our work per
5117 // If compiling with an exactly known VLEN, see if we can split a
5478 // For exact VLEN m2 or greater, try to split to m1 operations if we
6860 report_fatal_error("Support for VLEN==32 is incomplete.");
9141 // If we're compiling for an exact VLEN value, we can always perform
9145 if (auto VLEN = Subtarget.getRealVLen();
9146 VLEN && ContainerVT.bitsGT(M1VT)) {
9148 unsigned ElemsPerVReg = *VLEN / ElemVT.getFixedSizeInBits();
9357 // If we're compiling for an exact VLEN value and we have a known
9399 // O(n^2) / (VLEN/ETYPE) work. On the surface, going through the stack
9608 // (vscale * VF). Where vscale is defined as VLEN/RVVBitsPerBlock. The
9727 // LMUL * VLEN should be greater than or equal to EGS * SEW
10575 // If the subvector vector is a fixed-length type and we don't know VLEN
10815 // If the subvector vector is a fixed-length type and we don't know VLEN
11358 // If we know the exact VLEN and our fixed length vector completely fills
11420 // If we know the exact VLEN and our fixed length vector completely fills
18980 // We can use the minimum and maximum VLEN values to bound VLENB. We
18981 // know VLEN must be a power of two.
21752 // We define vscale to be VLEN/RVVBitsPerBlock. VLEN is always a power
22034 // If the minimum VLEN is less than RISCV::RVVBitsPerBlock we don't support