diff options
author | Kevin Ryde <user42@zip.com.au> | 2000-08-01 02:30:33 +0200 |
---|---|---|
committer | Kevin Ryde <user42@zip.com.au> | 2000-08-01 02:30:33 +0200 |
commit | ba4afbe7927f4de0bf5c4898290055b22ae5f819 (patch) | |
tree | fe72cae7818be1aa855bfdb1d6cce1e2667baffd | |
parent | 719a2e97aac003a622c8c3f619bf1b24611f3b6b (diff) | |
download | gmp-ba4afbe7927f4de0bf5c4898290055b22ae5f819.tar.gz |
Under ultrasparc umul_ppmm, remove the unfounded assertion that
something better than 90 cycles should be possible.
-rw-r--r-- | doc/tasks.html | 5 |
1 files changed, 2 insertions, 3 deletions
diff --git a/doc/tasks.html b/doc/tasks.html index ebd7de5c4..c549bf176 100644 --- a/doc/tasks.html +++ b/doc/tasks.html @@ -153,9 +153,8 @@ longlong.h umul_ppmm is suspected to be causing sqr_basecase to be slower than mul_basecase. <li> UltraSPARC: Write <code>umul_ppmm</code>. Important in particular for - <code>mpn_sqr_basecase</code>. The generic C code turns into sensible - enough "<code>mulx</code>"s, but it's about 90 cycles and something better - should be possible, maybe with floating point. + <code>mpn_sqr_basecase</code>. Using four "<code>mulx</code>"s either + with an asm block or via the generic C code is about 90 cycles. <li> Implement <code>mpn_mul_basecase</code> and <code>mpn_sqr_basecase</code> for important machines. Helping the generic sqr_basecase.c with an <code>mpn_sqr_diagonal</code> might be enough for some of the RISCs. |