Recent content by bear

These analyses that armchair quarterbacks bandy back and forth over technical characteristics, while often interesting and sometimes correct, strike me as largely beside the point. When OS/2 1.0 was finally purchasable, after something like two years of anticipation and slipped schedules, was...

I assumed it was in a socket. I think gslick has it right. It could very well be relevant, but I would want to have some kind of evidence before assuming it was bad, just because it's "hot". I'd still suspect faulty DRAM first, unless you already know for sure all of it is good.

Oh, hm. I'm not sure what that is for. It may not be necessary at all; it's not populated on this board: https://cdn11.bigcommerce.com/s-a1x7hg2jgk/images/stencil/1280x1280/products/1137/229950/AST-AST-Advantage-ISA-Memory-Card-202051_7230__30031.1706979629.jpg?c=2 At least three of the other...

Finally found a data sheet for that, as a Harris part (Intersil has been sold several times). Yes, it's a 1K x 4 OTP PROM with tristate outputs. It absolutely does not belong in any of the DRAM sockets...

I can't think of any reason an AST Advantage! would have such a part anywhere on it. But, I am assuming it's mixed in with the rest of the DRAM, in which case somebody who had no idea what they were doing stuck it in there because it "looked cool" (or thereabouts). I give odds >99% it's not a...

I don't think so. I think that's an Intersil PROM. In which case I'm not surprised it failed testing.

In my case it was severe enough to notice that my new 33 MHz 386 was having a hard time keeping up with my old 10 MHz 286, in certain applications.

Beware that the SPARCstorage Array 100- and 200-series take SCSI disks and are ~250 mbit/s FCAL. The StorEdge A5000/A5200 take FCAL disks, are 1063 mbit/s, and are more likely what you want. Most of the Solaris bundled software (above the driver level) is licensed specially for the A5000 and...

I only ever noticed it being slow with EMS enabled. Once I realized I didn't need EMS for anything and turned that feature of EMM386 off, the slowness (that I had noticed) vanished. I didn't do any benchmarking to find out whether it was still relatively slow, only less perceptibly so.

it'll be on the compaq ecu disk(s), e.g. sp2397 or sp2489.

I am reasonably certain I recall the AT skips addresses between the top of base memory and 1024k while counting memory during the self test. Here are two possible correct configurations, depending on whether you want 512k or 640k base memory. In all cases SW1 1-5 will be OFF ON OFF OFF ON...

I suppose it depends on the error, but IME if the CFG file contains, say, a syntax error, the ECU will reject it. If your CFG is syntactically correct but writes the wrong configuration registers, then I guess your card will not be configured correctly. I don't know from experience, but I don't...

what I've found from my own silly benchmarking efforts is that just the compiler on its own can make more than an order of magnitude difference. http://typewritten.org/Articles/Benchmarks/primes.html?mx=evolution_iv/e&l=c&os=dos&null=true

The ECU itself is relatively universal. The .CFG files have been collected, but obscure ones are still troublesome to find. I couldn't say which category the one you'll need for your SBC will be in. Some system boards also want an .OVL to provide some extra functionality above and beyond the...

Uh. It's worth mentioning there's another possibility here. Since we don't know what host they were written on, we cannot assume the data format is standard QIC. The 3B2 (non-SCSI) QIC drive, for example, does not use a standard data format, and while I could theoretically read this format, I am...