1 # RFC ls013 Min/Max GPR/FPR
5 * <https://libre-soc.org/openpower/sv/rfc/ls013/>
6 * <https://git.openpower.foundation/isa/PowerISA/issues/TODO>
7 * <https://bugs.libre-soc.org/show_bug.cgi?id=1057>
19 **Books and Section affected**:
22 Book I Fixed-Point and Floating-Point Instructions
23 Appendix E Power ISA sorted by opcode
24 Appendix F Power ISA sorted by version
25 Appendix G Power ISA sorted by Compliancy Subset
26 Appendix H Power ISA sorted by mnemonic
35 **Submitter**: Luke Leighton (Libre-SOC)
37 **Requester**: Libre-SOC
39 **Impact on processor**:
42 Addition of new GPR-based and FPR-based instructions
45 **Impact on software**:
48 Requires support for new instructions in assembler, debuggers,
55 GPR, FPR, min, max, fmin, fmax
60 Minimum/Maximum are common operations that can take an astounding number of
61 operations to implement in software. Additionally, Vector Reduce-Min/Max are
62 common vector operations, and SVP64 Parallel Reduction needs a single Scalar
63 instruction in order to effectively implement Reduce-Min/Max.
65 **Notes and Observations**:
67 1. SVP64 REMAP Parallel Reduction needs a single Scalar instruction to
68 work with, for best effectiveness. With no SFFS minimum/maximum instructions
69 Simple-V min/max Parallel Reduction is severely compromised.
70 2. Once one FP min/max mode is implemented the rest are not much more
72 3. There exists similar instructions in VSX (not IEEE754-2019 though).
73 This is frequently used to justify not
74 adding them. However SVP64/VSX may have different meaning from SVP64/SFFS,
75 so it is *really* crucial to have SFFS ops even if "equivalent" to VSX
76 in order for SVP64 to not be compromised (non-orthogonal).
77 4. FP min/max are rather complex to implement in software, the most commonly
78 used FP max function `fmax` from glibc compiled for SFFS is an
79 astounding 32 instructions.
83 Add the following entries to:
85 * the Appendices of Book I
86 * Book I 3.3.9 Fixed-Point Arithmetic Instructions
87 * Book I 4.6.6.1 Floating-Point Elementary Arithmetic Instructions
88 * Book I 1.6.1 and 1.6.2
94 # Floating-Point Instructions
96 This group is to provide Floating-Point min/max however with IEEE754 having advanced
97 to 2019 there are now subtle differences. These are selectable with a Mode Field, `FMM`.
99 ## `FMM` -- Floating Min/Max Mode
101 <a id="fmm-floating-min-max-mode"></a>
103 | `FMM` | Assembly Alias | Origin | Semantics |
104 |-------|-------------------------------|--------------------------------|-------------------------------------------------|
105 | 0000 | fminnum08[s] FRT, FRA, FRB | IEEE 754-2008 | FRT = minNum(FRA, FRB) (1) |
106 | 0001 | fmin19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minimum(FRA, FRB) |
107 | 0010 | fminnum19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minimumNumber(FRA, FRB) |
108 | 0011 | fminc[s] FRT, FRA, FRB | x86 minss or Win32's min macro | FRT = FRA \< FRB ? FRA : FRB |
109 | 0100 | fminmagnum08[s] FRT, FRA, FRB | IEEE 754-2008 (TODO: (3)) | FRT = minmaxmag(FRA, FRB, False, fminnum08) (2) |
110 | 0101 | fminmag19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minmaxmag(FRA, FRB, False, fmin19) (2) |
111 | 0110 | fminmagnum19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minmaxmag(FRA, FRB, False, fminnum19) (2) |
112 | 0111 | fminmagc[s] FRT, FRA, FRB | - | FRT = minmaxmag(FRA, FRB, False, fminc) (2) |
113 | 1000 | fmaxnum08[s] FRT, FRA, FRB | IEEE 754-2008 | FRT = maxNum(FRA, FRB) (1) |
114 | 1001 | fmax19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = maximum(FRA, FRB) |
115 | 1010 | fmaxnum19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = maximumNumber(FRA, FRB) |
116 | 1011 | fmaxc[s] FRT, FRA, FRB | x86 maxss or Win32's max macro | FRT = FRA > FRB ? FRA : FRB |
117 | 1100 | fmaxmagnum08[s] FRT, FRA, FRB | IEEE 754-2008 (TODO: (3)) | FRT = minmaxmag(FRA, FRB, True, fmaxnum08) (2) |
118 | 1101 | fmaxmag19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minmaxmag(FRA, FRB, True, fmax19) (2) |
119 | 1110 | fmaxmagnum19[s] FRT, FRA, FRB | IEEE 754-2019 | FRT = minmaxmag(FRA, FRB, True, fmaxnum19) (2) |
120 | 1111 | fmaxmagc[s] FRT, FRA, FRB | - | FRT = minmaxmag(FRA, FRB, True, fmaxc) (2) |
122 Note (1): for the purposes of minNum/maxNum, -0.0 is defined to be less than
123 +0.0. This is left unspecified in IEEE 754-2008.
125 Note (2): minmaxmag(x, y, cmp, fallback) is defined as:
128 def minmaxmag(x, y, is_max, fallback):
137 # equal magnitudes, or NaN input(s)
138 return fallback(x, y)
141 Note (3): TODO: icr if IEEE 754-2008 has min/maxMagNum like IEEE 754-2019's
142 minimum/maximumMagnitudeNumber
148 ## Floating Minimum/Maximum
153 * fminmax FRT, FRA, FRB, FMM
154 * fminmax. FRT, FRA, FRB, FMM
157 |0 |6 |11 |16 |21 |26 |31 |
158 | PO | FRT | FRA | FRB | FMM[0:3] / | XO | Rc |
161 Special Registers altered:
167 Compute the minimum/maximum of FRA and FRB, according to FMM, and store the
170 Assembly Aliases: see
171 [`FMM` -- Floating Min/Max Mode](#fmm-floating-min-max-mode)
175 ## Floating Minimum/Maximum Single
179 * fminmaxs FRT, FRA, FRB, FMM
180 * fminmaxs. FRT, FRA, FRB, FMM
183 |0 |6 |11 |16 |21 |26 |31 |
184 | PO | FRT | FRA | FRB | FMM[0:3] / | XO | Rc |
187 Special Registers altered:
195 Compute the minimum/maximum of FRA and FRB, according to FMM, and store the
198 Assembly Aliases: see
199 [`FMM` -- Floating Min/Max Mode](#fmm-floating-min-max-mode)
205 # Fixed-Point Instructions
207 These are signed and unsigned, min or max. SVP64 Prefixing defines Saturation
208 semantics therefore Saturated variants of these instructions need not be proposed.
210 ## Integer Min/Max Mode
212 * bit 0: set if word variant else dword
213 * bit 1: set if signed else unsigned
214 * bit 2: set if max else min
216 | `IMM` | Assembly Alias | Semantics |
217 |-------|------------------|----------------------------------------------|
218 | 000 | `minu RT,RA,RB` | `RT = (uint64_t)RA < (uint64_t)RB ? RA : RB` |
219 | 001 | `maxu RT,RA,RB` | `RT = (uint64_t)RA > (uint64_t)RB ? RA : RB` |
220 | 010 | `mins RT,RA,RB` | `RT = (int64_t)RA < (int64_t)RB ? RA : RB` |
221 | 011 | `maxs RT,RA,RB` | `RT = (int64_t)RA > (int64_t)RB ? RA : RB` |
222 | 100 | `minuw RT,RA,RB` | `RT = (uint32_t)RA < (uint32_t)RB ? RA : RB` |
223 | 101 | `maxuw RT,RA,RB` | `RT = (uint32_t)RA > (uint32_t)RB ? RA : RB` |
224 | 110 | `minsw RT,RA,RB` | `RT = (int32_t)RA < (int32_t)RB ? RA : RB` |
225 | 111 | `maxsw RT,RA,RB` | `RT = (int32_t)RA > (int32_t)RB ? RA : RB` |
227 ## Integer Min/Max MM-Form
229 * minmax RT, RA, RB, MMM
230 * minmax. RT, RA, RB, MMM
233 |0 |6 |11 |16 |21 |24 |25 |31 |
234 | PO | RT | RA | RB | MMM | / | XO | Rc |
240 if MMM[0] then # word mode
241 # shift left by XLEN/2 to make the dword comparison
242 # do word comparison of the original inputs
243 a <- a[XLEN/2:XLEN-1] || [0] * XLEN/2
244 b <- b[XLEN/2:XLEN-1] || [0] * XLEN/2
245 if MMM[1] then # signed mode
246 # invert sign bits to make the unsigned comparison
247 # do signed comparison of the original inputs
248 a[0] <- !a[0] # convert
250 if MMM[2] then # max mode
251 # swap a and b to make the less than comparison do
252 # greater than comparison of the original inputs
256 # store the entire selected source (even in word mode)
257 if a <u b then RT <- (RA)
261 Compute the integer minimum/maximum according to `MMM` of `RA` and `RB` and
262 store the result in `RT`.
264 Special Registers altered:
274 # Instruction Formats
276 Add the following entries to Book I 1.6.1.15 X-FORM:
279 |0 |6 |11 |16 |21 |26 |31 |
280 | PO | FRT | FRA | FRB | FMM[0:3] / | XO | Rc |
283 Add a new field to Book I 1.6.2 Word Instruction Fields:
287 Field used to specify minimum/maximum mode for fminmax[s].
298 Appendix E Power ISA sorted by opcode
299 Appendix F Power ISA sorted by version
300 Appendix G Power ISA sorted by Compliancy Subset
301 Appendix H Power ISA sorted by mnemonic
303 | Form | Book | Page | Version | mnemonic | Description |
304 |------|------|------|---------|----------|-------------|
305 | A | I | # | 3.2B | fminmax | Floating Minimum/Maximum |
306 | A | I | # | 3.2B | fminmaxs | Floating Minimum/Maximum Single |
307 | ??? | I | # | 3.2B | minmax | Minimum/max Signed/Unsigned |
309 ## fmax instruction count
311 32 instructions are required in SFFS to emulate fmax.
312 <https://gcc.godbolt.org/z/6xba61To6>
316 fmax(double, double):
354 .byte 0,9,0,0,0,0,0,0