Last fall, I tried to set up my J700 in a similar manner with a Sonnet 800Mhz
G4 card. I got as close as you did, too. OS X would sometimes run, sometimes
for about an hour, but eventually the system would completely lock up and
nothing short of powering the system down and back up would help. And since
the crash was unexpected, it seemed to leave some of the apps in a bad state,
often resulting in apps that would never start up again unless I reinstalled OS X.
After over a month of tweaking XPostFacto, swapping memory modules in and
out, I finally called Sonnet and asked the tech support person point-blank if
there was a known problem with their Crescendo 800 Mhz G4 card and the
J700.
The tech person told me that they were having lots of reported problems with this, and he said that one of the problems may be due to the fact that the Crescendo 800 Mhz card requires that memory NOT be interleaved.
I still find it odd that this accelerator card, which works fine under 9.1 would not
work under OS X. This is the first piece of hardware I've encountered which
became incompatable wtih installed memory due to an OS upgrade.
The reason you find it odd, is because you are a perceptive person and Sonnet tech support doesn't have a clue. It's exceedingly unlikely that interleaving is the problem and if it were, all of the machines for which the G4/700 & 800 were built would be having the same problem whenever memory was interleaved. Such a symptom would reduce the memory capacity of these machines by 25% and slow memory access by 10 - 15%.
However, it's a conveniently difficult to check this claim by Sonnet. Much more likely is that all the extra logic they loaded onto the card to handle the six slot machines (9500, 9600, not S900) and make the caches work properly are causing some low level problem with OSX's kernel and they probably didn't test with X during development.
Changing their "magic voodoo logic" now would mean changing hardware and that's expensive, so they'll just continue to sell the card the way it is, in it's broken state.
Your suggestion of disabling the S900's onboard cache is a good idea for a test. If the root of the problem is the logic that makes the 7455's cache levels work properly, then the onboard cache could be interfereing. However, if this were the problem, then the 9500 and original 9600 should have the same problem, because their caches are soldered down too and they don't have a convenient jumper for disabling the cache. You have to remove resistor R31 on those two machines.
Jeff Walther
-- SuperMacs is sponsored by <http://lowendmac.com/> and...
Small Dog Electronics http://www.smalldog.com | Refurbished Drives | Service & Replacement Parts [EMAIL PROTECTED] | & CDRWs on Sale! |
Support Low End Mac <http://lowendmac.com/lists/support.html>
SuperMacs list info: <http://lowendmac.com/supermacs/list.shtml> --> AOL users, remove "mailto:"; Send list messages to: <mailto:[EMAIL PROTECTED]> To unsubscribe, email: <mailto:[EMAIL PROTECTED]> For digest mode, email: <mailto:[EMAIL PROTECTED]> Subscription questions: <mailto:[EMAIL PROTECTED]> Archive: <http://www.mail-archive.com/supermacs%40mail.maclaunch.com/>
---------------------------------------------------------------
The Think Different Store
http://www.ThinkDifferentStore.com ---------------------------------------------------------------
