Yeah, that 'negative' silkscreen heatsink
is kinda cool...
Found a moment to read the ROM from my GC - attached. At first I thought it didn't work - all "00"s up to A000. I haven't fired my card up yet, and may not be able to for a while, so let me know if it works for you.
Thank you very much!
My card happily runs from this ROM
(the ROM I had, 341-0266, seems to occupy only the first half of the memory; it's attached below.)
341-0266 is listed as rev B in Apple's docs, but 341-0812 is not mentioned explicitly:
web.archive.org
The differences are not what I had guessed: TL;DR this is probably not the final Rev (C?) ROM, more likely the Rev A ROM. 16-bit color is missing, and everything else is essentially the same (also corroborated by the copyright dates on the stickers.)
Two points from the LEM page could be merged/amended based on these results:
- Rev. B adds support for 832 x 624 resolution and 16-bit color
- The Rev. B ROM (ROM part number 341-0266 or 341-0812) is required for Macintosh 16″ Color Display 832 x 624 support.
•16-bit color is available for 640x480 on 341-0266 but not on 341-0812.
•Using a DIP adapter to set 832x624 produces a working image up to 16-bit color on 341-0266 but only a powered-on black screen on 341-0812 (the system will still boot and allow a soft power-down using the keyboard.) Using a slave display confirms the master display is not recognized at all.
Things I hoped would be different/fixed but aren't:
•Netscape displays 'Back' in every button across the top bar and 'What's New?' in every bookmark on the bookmarks bar whenever acceleration is enabled.
•A trouble message prevents acceleration on slave displays whenever the GC doesn't also have a display plugged in. This can be remedied for a single start-up by shutting down, removing the GC card, reinserting it, and starting back up. Acceleration can then be used on the slave display without a display present on the GC until the following restart.
•With acceleration enabled in 7.5.3, background artifacts are frequently left when moving large windows or accessing the same menu more than once in quick succession. 7.6.1 seems more stable by contrast, though the same glitches still do occur.
Performance difference is within margin of error (SSW 7.5.3, IIfx) 40MHz / 50MHz:
•Speedometer 4.01 QuickDraw tests vs Q605: 341-0812 (341-0266)
Mono: 0.648 (0.645) / 0.732 (0.732)
Two bit 0.668 (0.661) / 0.752 (0.749)
Four Bit 0.711 (0.704) / 0.795 (0.791)
Eight Bit 0.781 (0.774) / 0.865 (0.862)
Sixteen Bit n/a (1.137) / n/a (1.203)
•Daystar PowerDemo tests: 341-0812 (341-0266)
DAYSTARS 8.92 sec (8.92 sec) / 7.40 sec (7.47 sec)
Sieve 44.8 sec ((45.0 sec) / 40.3 sec (40.5 sec)
Color Sort 10.17 sec (9.80 sec) / 8.78 sec (8.68 sec)
Disk I/O 2.8 sec (2.8 sec) / 2.4 sec (2.4 sec)
Fractal 510.5 sec (510.6 sec) / 419.5 sec (417.5 sec)
Lines 11.3 lines/sec (11.5 lines/sec) / 13.3/sec (13.5/sec)
•Personal Diagnostics vs IIci: 341-0812 (341-0266)
Black & White 4.97x (5.09x) / 5.60x (5.52x)
4-color 4.88x (4.88x) / 5.05x (5.05x)
16-color 4.21x (4.15x) / 4.28x (4.24x)
256-color 3.54x (3.59x) / 3.44x (3.55x)
Thousands n/a (2.07x) / n/a (2.08x)
Millions 3.66x (3.62x) / 3.68x (3.67x)