From: Luke Kenneth Casson Leighton Date: Mon, 20 Mar 2023 13:43:30 +0000 (+0000) Subject: add wording for binlut crbinlut to ls007 X-Git-Tag: opf_rfc_ls001_v3~135 X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=2672abf1129c6470d519ac82c4e91a8552855690;p=libreriscv.git add wording for binlut crbinlut to ls007 https://bugs.libre-soc.org/show_bug.cgi?id=1017 --- diff --git a/openpower/sv/rfc/ls007.mdwn b/openpower/sv/rfc/ls007.mdwn index ea08f6faa..3f6a4129c 100644 --- a/openpower/sv/rfc/ls007.mdwn +++ b/openpower/sv/rfc/ls007.mdwn @@ -220,18 +220,14 @@ CRB-form | 0.5|6.8 |9.10|11.13|14.15|16.18|19.25|26.30| 31| Form | |----|----|----|-----|-----|-----|-----|-----|---|----------| -| NN | BF | msk|BFA | msk | BFB | TLI | XO |TLI| CRB-Form | - -| 0.5| 6-8 | 9-11 | 12-14 | 15-17 | 18-20 | 21-28 | 29-30 | 31 | -|----|-----|------|-------|-------|-------|-------|-------|-----| -| PO | BF | BFA | BFB | BFC | msk | TLI | XO | msk | +| PO | BF | msk|BFA | msk | BFB | TLI | XO |TLI| CRB-Form | Pseudocode: ``` -a <- CR[4*BF+32:4*BFA+35] -b <- CR[4*BFA+32:4*BFB+35] -c <- CR[4*BFB+32:4*BFC+35] +a <- CR[4*BF+32:4*BF+35] +b <- CR[4*BFA+32:4*BFA+35] +c <- CR[4*BFB+32:4*BFB+35] ternary <- (~a & ~b & ~c & TLI[0]*4) | # 4 copies of TLI[0] (~a & ~b & c & TLI[1]*4) | # 4 copies of TLI[1] (~a & b & ~c & TLI[2]*4) | # ... @@ -286,26 +282,37 @@ Add this section to Book I 3.3.13 Pseudocode: ``` -if nh = 1 then - lut <- (RC)[56:59] -else - lut <- (RC)[60:63] -do i = 0 to 63 - idx <- (RB)[i] || (RA)[i] # compute index from current bits - result[i] <- lut[3 - idx] # subtract from 3 to index in LSB0 order +if nh = 1 then lut <- (RC)[56:59] +else lut <- (RC)[60:63] +result <- (~RA & ~RB & lut[0]*64) | + (~RA & RB & lut[1]*64) | + ( RA & ~RB & lut[2]*64) | + ( RA & RB & lut[3]*64)) RT <- result ``` +For each integer value i, 0 to 63, do the following. + + If nh contains a 0, let lut be the four LSBs of RC + (bits 60 to 63). Otherwise let lut be the next + four LSBs of RC (bits 56 to 59). + + Let j be the value of the concatenation of the + contents of bit i of RT with bit i of RB. + + The value of bit j of lut is placed into bit i of RT. + Special registers altered: ``` None ``` -**Programming Note**: +**Programmer's Note**: -Dynamic Ternary Logic may be emulated by appropriate combination of `binlog` and `ternlogi`, -using the `nh` (next half) operand to select first and second nibble: +Dynamic Ternary Logic may be emulated by appropriate combination of +`binlog` and `ternlogi`, using the `nh` (next half) operand to select +first and second nibble: ``` # compute r3 = ternlog(r4, r5, r6, table=r7) @@ -326,18 +333,33 @@ ternlogi r3, r4, r6, 0b11011000 With ternary (LUT3) dynamic instructions being very costly, and CR Fields being only 4 bit, a binary (LUT2) variant is better -| 0.5|6.8 | 9.11|12.14|15.17|18.21|22...30 |31| -| -- | -- | --- | --- | --- |-----| -------- |--| -| NN | BT | BA | BB | BC |m0-m3|000101110 |0 | +| 0.5|6.8 |9.10|11.13|14.15|16.18|19.25|26.30| 31| Form | +|----|----|----|-----|-----|-----|-----|-----|---|----------| +| PO | BF | msk|BFA | msk | BFB | // | XO |// | CRB-Form | + +``` +a <- CR[4*BF+32:4*BFA+35] +b <- CR[4*BFA+32:4*BFA+35] +lut <- CR[4*BFB+32:4*BFB+35] +binary <- (~a & ~b & lut[0]*4) | + (~a & b & lut[1]*4) | + ( a & ~b & lut[2]*4) | + ( a & b & lut[3]*4)) +do i = 0 to 3 + if msk[i] = 1 then + CR[4*BF+32+i] <- binary[i] +``` - mask = m0..m3 - for i in range(4): - a,b = CRs[BA][i], CRs[BB][i]) - if mask[i] CRs[BT][i] = lut2(CRs[BC], a, b) +For each integer value i, 0 to 3, do the following. -When SVP64 Vectorised any of the 4 operands may be Scalar or -Vector, including `BC` meaning that multiple different dynamic -lookups may be performed with a single instruction. + Let j be the value of the concatenation of the + contents of bit i of CR Field BF with bit i of CR Field BFA. + + If bit i of msk is set to 1 then the value of bit j of + CR Field BFB is placed into bit i of CR Field BF. + + Otherwise, if bit i of msk is a zero then bit i of + CR Field BF is unchanged. *Programmer's note: just as with binlut and ternlogi, a pair of crbinlog instructions followed by a merging crternlogi may @@ -357,9 +379,12 @@ lookups may be performed with a single instruction. Appendix G Power ISA sorted by Compliancy Subset Appendix H Power ISA sorted by mnemonic -|Form| Book | Page | Version | mnemonic | Description | -|----|------|------|---------|----------|-------------| -|TLI | I | # | 3.2B | ternlogi | Ternary Logic Immediate | +|Form| Book | Page | Version | mnemonic | Description | +|----|------|------|---------|------------|-------------| +|TLI | I | # | 3.2B | ternlogi | GPR Ternary Logic Immediate | +|VA | I | # | 3.2B | binlog | GPR Binary Logic | +|CRB | I | # | 3.2B | crternlogi | CR Field Ternary Logic Immediate | +|CRB | I | # | 3.2B | crbinlog | CR Field Binary Logic | ----------------