The following code section works for VREG_W = 64 to 1024 but not for any bigger configurations:
"addi t1, %[vl], -1 \n\t"
"vrgather.vx v16, v0, t1 \n\t"
"vslide1up.vx v12, v0, %[ptr] \n\t"
"vmv.x.s %[ptr_out], v16 \n\t"
"vsuxei32.v v8, (x0), v12 \n\t"
"vadd.vi v12, v12, 4 \n\t"
The following code section works for VREG_W = 64 to 1024 but not for any bigger configurations:
"addi t1, %[vl], -1 \n\t"
"vrgather.vx v16, v0, t1 \n\t"
"vslide1up.vx v12, v0, %[ptr] \n\t"
"vmv.x.s %[ptr_out], v16 \n\t"
"vsuxei32.v v8, (x0), v12 \n\t"
"vadd.vi v12, v12, 4 \n\t"