Hi,

I'd run HOWMESSY to see how long the worst and average chains
are on the detail (and associated masters) in question.
I suspect a lot of time is spent chasing down the chains.

Most of the traces show that it's walking synonym chains in a
master during DBFIND.

>          export stub: 2b1.00d5c600 setupsmcb_312+$22c
> NM  9) SP=4184bf00 RP=2b1.00d5cc90 get'free'list'buffer_314+$1ec
> NM  a) SP=4184be98 RP=2b1.00d5d45c ti_getblock+$200
> NM  b) SP=4184be48 RP=2b1.00d89f04 get'block_314+$f4
> NM  c) SP=4184bd40 RP=2b1.00d8a57c get'record_315+$4f4
> NM  d) SP=4184bce0 RP=2b1.00d908c4 findchainhead_332+$294
> NM  e) SP=4184bc88 RP=2b1.00d91230 putdtlentry_333+$298

> NM  8) SP=41848998 RP=a.00585ed0 ?disc_sm_start_read+$8
>          export stub: 2b1.00d5c6b4 setupsmcb_312+$2e0
> NM  9) SP=418488d8 RP=2b1.00d5cc90 get'free'list'buffer_314+$1ec
> NM  a) SP=41848870 RP=2b1.00d5d45c ti_getblock+$200
> NM  b) SP=41848820 RP=2b1.00d640dc ti'bmfindrecord+$27c
> NM  c) SP=41848718 RP=2b1.00d7ac2c nmdbfind+$b1c


> NM  a) SP=41848870 RP=2b1.00d5d45c ti_getblock+$200
> NM  b) SP=41848820 RP=2b1.00d640dc ti'bmfindrecord+$27c
> NM  c) SP=41848718 RP=2b1.00d7ac2c nmdbfind+$b1c
> NM  d) SP=41848620 RP=2b1.00d31ca8 dbfind+$180
> NM  e) SP=41848478 RP=2b1.00d31b04 ?dbfind+$8
>          export stub: 2b1.001b6860 BB_DBFIND+$70

> NM  a) SP=41848870 RP=2b1.00d5d45c ti_getblock+$200
> NM  b) SP=41848820 RP=2b1.00d640dc ti'bmfindrecord+$27c
> NM  c) SP=41848718 RP=2b1.00d7ac2c nmdbfind+$b1c
> NM  d) SP=41848620 RP=2b1.00d31ca8 dbfind+$180
> NM  e) SP=41848478 RP=2b1.00d31b04 ?dbfind+$8

--
Stan Sieler                                          [log in to unmask]
                                     http://www.allegro.com/sieler.html