diff options
author | Bill Traynor <wmat@riscv.org> | 2023-08-01 13:15:58 -0400 |
---|---|---|
committer | Bill Traynor <wmat@riscv.org> | 2023-08-01 13:15:58 -0400 |
commit | bebbad41087bbfb713c15db173cc96daf2bd1a81 (patch) | |
tree | 1f3305bac68564041004111f1d34e893e0c9642f /src/v-st-ext.adoc | |
parent | 9cd24d514bc52cb3863b8a547cd1ed40d07727ea (diff) | |
download | riscv-isa-manual-bebbad41087bbfb713c15db173cc96daf2bd1a81.zip riscv-isa-manual-bebbad41087bbfb713c15db173cc96daf2bd1a81.tar.gz riscv-isa-manual-bebbad41087bbfb713c15db173cc96daf2bd1a81.tar.bz2 |
Setting up the inclusion of Vector.
Added Vector and all supporting files.
Diffstat (limited to 'src/v-st-ext.adoc')
-rw-r--r-- | src/v-st-ext.adoc | 22 |
1 files changed, 11 insertions, 11 deletions
diff --git a/src/v-st-ext.adoc b/src/v-st-ext.adoc index e52bc59..619492b 100644 --- a/src/v-st-ext.adoc +++ b/src/v-st-ext.adoc @@ -165,7 +165,7 @@ The `vtype` register has five fields, `vill`, `vma`, `vta`, `vsew[2:0]`, and `vlmul[2:0]`. Bits `vtype[XLEN-2:8]` should be written with zero, and non-zero values in this field are reserved. -include::vtype-format.adoc[] +include::images/wavedrom/vtype-format.adoc[] NOTE: A small implementation supporting ELEN=32 requires only seven bits of state in `vtype`: two bits for `ma` and `ta`, two bits for @@ -887,11 +887,11 @@ floating-point load/store 12-bit immediate field to provide further vector instruction encoding, with bit 25 holding the standard vector mask bit (see <<sec-vector-mask-encoding>>). -include::vmem-format.adoc[] +include::images/wavedrom/vmem-format.adoc[] -include::valu-format.adoc[] +include::images/wavedrom/valu-format.adoc[] -include::vcfg-format.adoc[] +include::images/wavedrom/vcfg-format.adoc[] Vector instructions can have scalar or vector source operands and produce scalar or vector results, and most vector instructions can be @@ -1148,11 +1148,11 @@ their arguments, and write the new value of `vl` into `rd`. vsetvl rd, rs1, rs2 # rd = new vl, rs1 = AVL, rs2 = new vtype value ---- -include::vcfg-format.adoc[] +include::images/wavedrom/vcfg-format.adoc[] ==== `vtype` encoding -include::vtype-format.adoc[] +include::images/wavedrom/vtype-format.adoc[] The new `vtype` value is encoded in the immediate fields of `vsetvli` and `vsetivli`, and in the `rs2` register for `vsetvl`. @@ -1348,7 +1348,7 @@ floating-point load/store 12-bit immediate field to provide further vector instruction encoding, with bit 25 holding the standard vector mask bit (see <<sec-vector-mask-encoding>>). -include::vmem-format.adoc[] +include::images/wavedrom/vmem-format.adoc[] [cols="4,12"] |=== @@ -2173,7 +2173,7 @@ The vector arithmetic instructions use a new major opcode (OP-V = 1010111~2~) which neighbors OP-FP. The three-bit `funct3` field is used to define sub-categories of vector instructions. -include::valu-format.adoc[] +include::images/wavedrom/valu-format.adoc[] [[sec-arithmetic-encoding]] ==== Vector Arithmetic Instruction encoding @@ -3461,7 +3461,7 @@ The following table gives the seven MSBs of the output significand as a function of the LSB of the normalized input exponent and the six MSBs of the normalized input significand; the other bits of the output significand are zero. -include::vfrsqrt7.adoc[] +include::images/wavedrom/vfrsqrt7.adoc[] NOTE: For example, when SEW=32, vfrsqrt7(0x00718abc ({approx} 1.043e-38)) = 0x5f080000 ({approx} 9.800e18), and vfrsqrt7(0x7f765432 ({approx} 3.274e38)) = 0x1f820000 ({approx} 5.506e-20). @@ -3548,7 +3548,7 @@ The following table gives the seven MSBs of the normalized output significand as a function of the seven MSBs of the normalized input significand; the other bits of the normalized output significand are zero. -include::vfrec7.adoc[] +include::images/wavedrom/vfrec7.adoc[] If the normalized output exponent is 0 or -1, the result is subnormal: the output exponent is 0, and the output significand is given by concatenating @@ -5173,5 +5173,5 @@ computation in single-precision will suffice. === Vector Instruction Listing -include::inst-table.adoc[] +include::images/wavedrom/inst-table.adoc[] |