Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
20 check-ins using file examples/ac-mq-blinker.pal version 748eb7ccdc
|
2017-01-12
| ||
| 06:43 | Another URL fix in the new licensing material check-in: e0241770dd user: tangent tags: trunk | |
| 06:40 | Added copyright messages to the Markdown files that reference SIMH-LICENSE.md so that it is clear who is granting the license. check-in: 160483044a user: tangent tags: trunk | |
| 06:37 | Added license section to end of examples/README.md check-in: 7024cc51a2 user: tangent tags: trunk | |
| 06:35 | Clarifications to the DEC hobbyist license claim in COPYING.md. check-in: 08b984e585 user: tangent tags: trunk | |
| 06:25 | Small link fix for previous checkin check-in: f360190dd4 user: tangent tags: trunk | |
| 06:24 | Many more software licenses chased down and added to the appropriate files. Where direct editing was not possible, created license files nearby, and pointed to them from a new COPYING.md file. check-in: fc7e5f2a91 user: tangent tags: trunk | |
|
2017-01-09
| ||
| 22:27 | Added the new SIMH-style license grant to the header comment of the gpio*.c modules, and added Ian Schofield to the gpio.c module (only!) and to the pdp8_cpu.c.in module, per this private email reply to my licensing proposal: "This looks absolutely fine to me. Do add this license to the tarball if we all agree." check-in: 06829f12b1 user: tangent tags: trunk | |
| 22:09 | SIMH doesn't support THROTTLE values below 1k (or maybe 10k) so changed the --throttle=human and =trace configure script options to use a different syntax which allows specifying speed in terms of instructions/ms. check-in: 26a6b02367 user: tangent tags: trunk | |
| 04:03 | Lowercased the "SET [NO]THROTTLE" commands to match the style of the existing boot/*.script commands. SIMH doesn't care, but it's inconsistent. check-in: 0e15f11ea2 user: tangent tags: trunk | |
| 02:56 | Added the license notice to src/scanswitch.c as well check-in: ffa1390cb8 user: tangent tags: trunk | |
|
2017-01-08
| ||
| 15:53 | The *BSDs all understand sysctl hw.ncpu, so added tests for those to tools/corecount, so it'll return the correct value on them, too. check-in: 4be8a6510e user: tangent tags: trunk | |
| 15:36 | We weren't cleaning up after a few of the recently-added *.in files in "make clean". Fixed. check-in: 8cd4016117 user: tangent tags: trunk | |
| 15:03 | A prior checkin accidentally changed the way HLT instructions are handled. It was an attempt to fix bug [f961906a5c] which wasn't intended to be checked in; the fix doesn't work, so it's now back to the way the code worked yesterday, which is still broken. check-in: 6155a4b535 user: tangent tags: trunk | |
| 14:39 | Adjusted the single-core SIMH throttle value down from 3.3 MIPS to 1.6 MIPS. The prior value gave about 50% CPU usage for all threads in pidp8i-sim on a Pi 3, so cut it in half for a Pi 1. Also had to change the notation from "3.3M" to "1600k" because SIMH refuses to handle fractional values; LAME. check-in: 0770a4a2ec user: tangent tags: trunk | |
| 14:26 | All of the boot scripts except for those generated from assembly code now have a SIMH throttle value set based on either detected CPU cores at configuration time or the value of the new --throttle configure script value. Added README-throttle.md to describe the values allowed for this option, as they're pretty extensive. On a multi-core machine, the default is to have the prior behavior, which is no throttling: run flat-out, taking up a whole CPU core for pidp8i-sim's main thread, leaving the rest for the GPIO thread and whatever else the Pi does. On a single-core machine, the default is to set a value that should give something like 50% CPU usage to pidp8i-sim on the slowest Raspberry Pis we support, allowing enough CPU power left over for the incandescent lamp simulator thread. Consequently, removed README-single-core.md and the references to it. Closes [dfd864da90]. check-in: 48b84d4bd9 user: tangent tags: trunk | |
| 11:11 | The mkrel process now attempts to do all the work automatically, with no user input. The prior release showed that the gap between releases is long enough that I can't remember what to write as release comments, and because the comment editor takes over the screen, I can't figure it out from the context of previously-executed commands, since the editor hides that. check-in: 58e0f138a0 user: tangent tags: trunk | |
| 11:05 | Replaced the "make -j11" hard-coded values with a call to a new helper script, tools/mmake which calls another newly-added script (tools/corecount) to get the core count on this system and multiply it by 1.5 and pass it to make -j. It currently knows how to do this on Linux and Mac OS X. For all other systems, it returns 1, so it gives you make -j1, since the math truncates, rather than rounds. This is not just a nicety: "make -j11" can lock up a Pi in some cases. It appears to be worst when most of the files in the system have to be rebuilt. It's probably running the system out of RAM, and it isn't handling it gracefully. check-in: cffeb65eb8 user: tangent tags: trunk | |
| 10:53 | Style improvments to the coverage of ./configure script options in README.md. check-in: e30b30d2ef user: tangent tags: trunk | |
| 10:31 | Backed out part of [842fb3ef61] since now all of boot/*.script are generated in one way or another. Therefore, we no longer need to install @srcdir@/boot/*.script. (This doesn't revert the change, it effectively *inverts* the way this install step worked prior to that checkin: we used to install only from the srcdir, but now we install only from the builddir.) check-in: 71586b261d user: tangent tags: trunk | |
| 10:27 | Added tools/mkbootscript, a Perl script that parses palbart listing files generated during the normal build process and turns them into SIMH boot scripts. This in turn means we can remove boot/1.script and boot/5.script, since they can now be generated from corresponding programs in examples/*.pal. Closes [87cb9ca2f0] check-in: d77e169826 user: tangent tags: trunk | |