The Landmark AT (5170) test bios will test the RTC. It will also test the alarm feature and expect it to fire at the correct time.

Also CheckIT does it too.

Just managed to fix the CMOS date in the latest commit. Now both MS-DOS and CheckIt! Diagnostics detect it correctly.

The Date Comparison and Real-time clock alarm check out correctly(passed tests). The other two(time-based) tests fail(1.35 apart).

CheckIt! Diagnostics also displays the actual system time of Windows(which is taken at GMT). MS-DOS gives me a time a bit from the start of emulation, while the RTC gives the correct GMT time. Any idea what's going wrong? Why does MS-DOS count from midnight(when the emulator is started)?

Edit: The current source code: https://bitbucket.org/superfury/unipcemu/src/ … MOS.c?at=master

Edit: Btw, the CPU runs at about 75% speed on my AT emulation(using a Intel i7-4790K 4.0GHz CPU). Does this have anything to do with MS-DOS not taking the time set in the RTC(although the date is taken)?

The Landmark AT test BIOS says that the RTC Interrupt(Using the Periodic Interrupt) isn't working correctly ("1 FAILED" is reported). I see the timer (which runs at double the rate, at 64kHz to be able to create a square wave at 32kHz), which is timing out(generating interrupts) at increasing values from: 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x3FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0xFFF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x3FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x7FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x3FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x1FFF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x3FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x7FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF, 0x7F, 0x3FF, 0x7F, 0xFF, 0x7F, 0x1FF, 0x7F, 0xFF ... When testing the Real Time Clock Interrupt test of the BIOS.

Also a little question: what happens on an AT CMOS chip with 64 bytes of RAM installed, when bytes 64-127 are read? Does it simply wrap to 0-63? Or do they map to unexisting(always 0x00/0xFF)?