1 Copyright 1997, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
3 This file is part of the GNU MP Library.
5 The GNU MP Library is free software; you can redistribute it and/or modify
6 it under the terms of the GNU Lesser General Public License as published by
7 the Free Software Foundation; either version 3 of the License, or (at your
8 option) any later version.
10 The GNU MP Library is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
12 or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
13 License for more details.
15 You should have received a copy of the GNU Lesser General Public License
16 along with the GNU MP Library. If not, see http://www.gnu.org/licenses/.
22 This directory contains mpn functions for 64-bit V9 SPARC
24 RELEVANT OPTIMIZATION ISSUES
27 IANY = shift/add/sub/logical/sethi
28 IADDLOG = add/sub/logical/sethi
30 FA = fadd*/fsub*/f*to*/fmov*
33 UltraSPARC can issue four instructions per cycle, with these restrictions:
34 * Two IANY instructions, but only one of these may be a shift. If there is a
35 shift and an IANY instruction, the shift must precede the IANY instruction.
40 * IANY/IADDLOG/MEM must be insn 1, 2, or 3 in an issue bundle. Taken branches
41 should not be in slot 4, since that makes the delay insn come from separate
43 * If two IANY/IADDLOG instructions are to be executed in the same cycle and one
44 of these is setting the condition codes, that instruction must be the second
47 To summarize, ignoring branches, these are the bundles that can reach the peak
50 insn1 iany iany mem iany iany mem iany iany mem
51 insn2 iaddlog mem iany mem iaddlog iany mem iaddlog iany
52 insn3 mem iaddlog iaddlog fa fa fa fm fm fm
53 insn4 fa/fm fa/fm fa/fm fm fm fm fa fa fa
55 The 64-bit integer multiply instruction mulx takes from 5 cycles to 35 cycles,
56 depending on the position of the most significant bit of the first source
57 operand. When used for 32x32->64 multiplication, it needs 20 cycles.
58 Furthermore, it stalls the processor while executing. We stay away from that
59 instruction, and instead use floating-point operations.
61 Floating-point add and multiply units are fully pipelined. The latency for
62 UltraSPARC-1/2 is 3 cycles and for UltraSPARC-3 it is 4 cycles.
64 Integer conditional move instructions cannot dual-issue with other integer
65 instructions. No conditional move can issue 1-5 cycles after a load. (This
66 might have been fixed for UltraSPARC-3.)
68 The UltraSPARC-3 pipeline is very simular to he one of UltraSPARC-1/2 , but is
69 somewhat slower. Branches execute slower, and there may be other new stalls.
70 But integer multiply doesn't stall the entire CPU and also has a much lower
71 latency. But it's still not pipelined, and thus useless for our needs.
75 * mpn_lshift, mpn_rshift: The current code runs at 2.0 cycles/limb on
76 UltraSPARC-1/2 and 2.65 on UltraSPARC-3. For UltraSPARC-1/2, the IEU0
77 functional unit is saturated with shifts.
79 * mpn_add_n, mpn_sub_n: The current code runs at 4 cycles/limb on
80 UltraSPARC-1/2 and 4.5 cycles/limb on UltraSPARC-3. The 4 instruction
81 recurrency is the speed limiter.
83 * mpn_addmul_1: The current code runs at 14 cycles/limb asymptotically on
84 UltraSPARC-1/2 and 17.5 cycles/limb on UltraSPARC-3. On UltraSPARC-1/2, the
85 code sustains 4 instructions/cycle. It might be possible to invent a better
86 way of summing the intermediate 49-bit operands, but it is unlikely that it
87 will save enough instructions to save an entire cycle.
89 The load-use of the u operand is not enough scheduled for good L2 cache
90 performance. The UltraSPARC-1/2 L1 cache is direct mapped, and since we use
91 temporary stack slots that will conflict with the u and r operands, we miss
92 to L2 very often. The load-use of the std/ldx pairs via the stack are
93 perhaps over-scheduled.
95 It would be possible to save two instructions: (1) The mov could be avoided
96 if the std/ldx were less scheduled. (2) The ldx of the r operand could be
97 split into two ld instructions, saving the shifts/masks.
99 It should be possible to reach 14 cycles/limb for UltraSPARC-3 if the fp
100 operations where rescheduled for this processor's 4-cycle latency.
102 * mpn_mul_1: The current code is a straightforward edit of the mpn_addmul_1
103 code. It would be possible to shave one or two cycles from it, with some
106 * mpn_submul_1: Simpleminded code just calling mpn_mul_1 + mpn_sub_n. This
107 means that it runs at 18 cycles/limb on UltraSPARC-1/2 and 23 cycles/limb on
108 UltraSPARC-3. It would be possible to either match the mpn_addmul_1
109 performance, or in the worst case use one more instruction group.
111 * US1/US2 cache conflict resolving. The direct mapped L1 date cache of US1/US2
112 is a problem for mul_1, addmul_1 (and a prospective submul_1). We should
113 allocate a larger cache area, and put the stack temp area in a place that
114 doesn't cause cache conflicts.