| 1 |
Copyright 1996 Free Software Foundation, Inc. |
| 2 |
|
| 3 |
This file is part of the GNU MP Library. |
| 4 |
|
| 5 |
The GNU MP Library is free software; you can redistribute it and/or modify |
| 6 |
it under the terms of either: |
| 7 |
|
| 8 |
* the GNU Lesser General Public License as published by the Free |
| 9 |
Software Foundation; either version 3 of the License, or (at your |
| 10 |
option) any later version. |
| 11 |
|
| 12 |
or |
| 13 |
|
| 14 |
* the GNU General Public License as published by the Free Software |
| 15 |
Foundation; either version 2 of the License, or (at your option) any |
| 16 |
later version. |
| 17 |
|
| 18 |
or both in parallel, as here. |
| 19 |
|
| 20 |
The GNU MP Library is distributed in the hope that it will be useful, but |
| 21 |
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY |
| 22 |
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
| 23 |
for more details. |
| 24 |
|
| 25 |
You should have received copies of the GNU General Public License and the |
| 26 |
GNU Lesser General Public License along with the GNU MP Library. If not, |
| 27 |
see https://www.gnu.org/licenses/. |
| 28 |
|
| 29 |
|
| 30 |
|
| 31 |
|
| 32 |
|
| 33 |
This directory contains mpn functions optimized for MIPS3. Example of |
| 34 |
processors that implement MIPS3 are R4000, R4400, R4600, R4700, and R8000. |
| 35 |
|
| 36 |
RELEVANT OPTIMIZATION ISSUES |
| 37 |
|
| 38 |
1. On the R4000 and R4400, branches, both the plain and the "likely" ones, |
| 39 |
take 3 cycles to execute. (The fastest possible loop will take 4 cycles, |
| 40 |
because of the delay insn.) |
| 41 |
|
| 42 |
On the R4600, branches takes a single cycle |
| 43 |
|
| 44 |
On the R8000, branches often take no noticeable cycles, as they are |
| 45 |
executed in a separate function unit.. |
| 46 |
|
| 47 |
2. The R4000 and R4400 have a load latency of 4 cycles. |
| 48 |
|
| 49 |
3. On the R4000 and R4400, multiplies take a data-dependent number of |
| 50 |
cycles, contrary to the SGI documentation. There seem to be 3 or 4 |
| 51 |
possible latencies. |
| 52 |
|
| 53 |
4. The R1x000 processors can issue one floating-point operation, two integer |
| 54 |
operations, and one memory operation per cycle. The FPU has very short |
| 55 |
latencies, while the integer multiply unit is non-pipelined. We should |
| 56 |
therefore write fp based mpn_Xmul_1. |
| 57 |
|
| 58 |
STATUS |
| 59 |
|
| 60 |
Good... |