First post, by jakethompson1
- Rank
- Oldbie
I recently posted about not being able to get cache read faster than 3-2-2-2 and write faster than 1 WS on my board. It's running a DX2-80 and has a UMC 491F chipset.
I just replaced all cache chips plus the tag ram with 12 ns, 256 kilobit (32Kx8) chips. They are from Digi-Key and therefore should be legitimate. No dice. I still can't get the timings any faster. I experimented with swapping out VLB VGA for an ISA one, and setting the VLB and CPUCLK jumpers for 40 MHz operation. Doesn't help. On 3-1-1-1 sometimes I can get it to boot partially into DOS and then have it complain about can't find command.com, etc., or suddenly AMISETUP claims the BIOS is password-protected. Very erratic operation. I can't get it to boot at all with the read timing on 2-1-1-1 or write timing on 0 WS. The DRAM is 60ns and is set (successfully) to 0 WS.
Doesn't seem like there is any point in having cache on this board if it can't go any faster than this.
With L2 disabled, CACHECHK reports 16 us/KB up to 8 KB (L1) cache, and 22 us/KB for all bigger sizes.
With L2 enabled, CACHECHK reports 16 us/KB up to 8KB, 20 us/KB from 16 to 256 KB, and 37 us/KB on anything bigger. So I'd be trading 10% faster for things in the cache for an added 68% miss penalty on anything not in the cache. Ouch!
My board is maxed out at 256KB cache. Any chance it would work at 2-1-1-1 and 0 WS with a smaller cache? Problem is, the smaller cache sizes call for 8Kbx8 or 16Kbx8 chips. Can I jumper it as if I've swapped the chips for smaller ones or will that not work? I'm inspired to try out smaller caches by some of the commentary in Feipoa's Biostar MB-8433UUD manual.
Presumably the chipset is just too slow? Perhaps it wasn't actually designed for 40 MHz operation? I haven't seen how much I can push it if I lower to 33 MHz but that might be an experiment for another day.
Was PC Chips onto something? haha
Thanks