> The individual chips on the SIMMs should have id numbers ending
> in -70 or -7 for 70ns chips and -60 or -6 for 60 ns chips. I don't
> know how you check for EDO vs. regular SIMMS, but I would've
> thought your BIOS could detect that.
>
> I found that with my Cyrix 166+ on a SOYO board I had to 'de-tune'
> the BIOS setting to 70ns, despite have bought brand new 60ns EDO
> SIMMS. Try it and see if it helps.
Okay.
The id on the chips is 6, so i guess 60 ns. Your suggestion was really
fascinating. I have downgraded the memory timing from 60 to 70 ns and now
after 1:15 hour of stress-test with kernel compilation (make -j 10),
equivalent to 7 compilations, I have had just two (2):
gcc: Internal compiler error: program cc1 got fatal signal 7
Signal 7 ? What is this ?
Okay okay, no more SIGSEGV (signal 11) and the memory problem seems to be
gone, but this is SIGBUS. Why ? May be an out of memory condition ?
Just now another error showed up in the compilation:
audio.c: In function `audio_select':
audio.c:479: structure has no member named `flags'
audio.c:488: structure has no member named `flags'
but for sure the structure have a member named flags. Seems that the
compiler have chocked somewhere, may be something related with the
conditional compilation ?
Mah !
Ciao,
Riccardo.