VOGONS


First post, by LuigiThirty

User metadata
Rank Newbie
Rank
Newbie

I got a replacement motherboard for my 486 since the last one I had was flaky. It's another V4S471. I installed the hacked BIOS to replace the onboard WinBIOS. I've got an Am5x86 running at 4x40MHz, 128MB of RAM in the board and 256K of 15ns cache. (I bought a bunch of 611024 chips on eBay but they were dead, as it turns out.)

The motherboard detects the cache but it doesn't seem to be working properly. According to CACHECHK, only the 21st megabyte of RAM is being cached by the L2. Access speeds are the same for all RAM except inside the 21st megabyte - 7uS/KB up to 16KB, 33uS/KB up to 4096KB. The 21st megabyte reads like it's being L2 cached, with faster speeds up to 256KB access blocks before returning to 33uS/KB.

I checked the SiS471 configuration registers in the datasheet with DEBUG.COM and all of DRAM is set to be cacheable. I checked the SRAM ICs in my TL866 and they all read good. The jumpers are set properly for 32Kx8. L1 and L2 is set for write-back.

Reply 1 of 3, by mkarcher

User metadata
Rank l33t
Rank
l33t

The symptom sounds like the chipset is misconfigured into the "do not fill the cache on read misses, but keep serving read/write hits" mode. You should carefully check all the L2 cache enable bits in the chipset configuration.

Reply 2 of 3, by Anonymous Coward

User metadata
Rank l33t
Rank
l33t

Was it running properly with the WinBIOS?

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 3 of 3, by LuigiThirty

User metadata
Rank Newbie
Rank
Newbie

I put the WinBIOS back in and nope, same thing.

The BIOS detects L2 cache, HWINFO does not. Hm. I did figure out that the BIOS will report 256K cache even with one or more chips missing, so maybe I have a bad socket.