| --- Log opened Sat Oct 29 00:00:39 2016 |
| 01:42 | | catadroid [catalyst@Nightstar-0mshtr.dab.02.net] has joined #code |
| 01:42 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [Connection closed] |
| 02:50 | | macdjord|wurk is now known as macdjord |
| 02:59 | | Vash [Vash@Nightstar-uhn82m.ct.comcast.net] has quit [[NS] Quit: Quit] |
| 03:38 | | catadruid [catalyst@Nightstar-65o9q3.dab.02.net] has joined #code |
| 03:41 | | catadroid [catalyst@Nightstar-0mshtr.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 04:15 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 04:15 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 04:38 | | himi [sjjf@Nightstar-jcoeh9.staticip.rima-tde.net] has quit [Ping timeout: 121 seconds] |
| 04:46 | | himi [sjjf@Nightstar-jcoeh9.staticip.rima-tde.net] has joined #code |
| 04:46 | | mode/#code [+o himi] by ChanServ |
| 04:51 | | himi [sjjf@Nightstar-jcoeh9.staticip.rima-tde.net] has quit [Ping timeout: 121 seconds] |
| 05:03 | <&McMartin> | Hum |
| 05:03 | <&[R]> | Ho |
| 05:03 | | * McMartin tries to use mkrescue, but at the "editing partition table" step just gets "Unexpected error! Aborting!" |
| 05:06 | <&McMartin> | OK, a few extra tries and it fits together |
| 05:06 | <&McMartin> | So let's see if it works. |
| 05:09 | <&McMartin> | Nope. |
| 05:09 | <&McMartin> | "Failed to load COM32 file /boot/syslinux/menu.c32" |
| 05:09 | <&McMartin> | CDs are the next challenge, I guess |
| 05:09 | <&McMartin> | That should fail to run afoul of USB support vanishing partway through the boot process |
| 05:10 | <&[R]> | Can't build a FreeDOS kernel with USB support monolithically installeD? |
| 05:10 | | * [R] hasn't used FreeDOS in a while |
| 05:10 | <&McMartin> | That sounds like more work than making an iso |
| 05:10 | | Derakon is now known as Derakon[AFK] |
| 06:47 | <&McMartin> | ToxicFrog: Do your mkrescue sticks successfully boot on Thoth? Because it's my Thoth-class laptop that's having the issues |
| 06:53 | | gizmore [kvirc@Nightstar-g7eok3.dip0.t-ipconnect.de] has joined #code |
| 07:08 | | Kindamoody[zZz] is now known as Kindamoody |
| 08:02 | | Kindamoody is now known as Kindamoody|afk |
| 08:25 | | celticminstrel [celticminst@Nightstar-pvpf2s.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:55 | <&McMartin> | early experiments with bootable FreeDOS CDs are good, but I can't burn an image until I get my payload right |
| 08:56 | <&McMartin> | VBox is happy with the isos booting and running stuff though |
| 09:19 | <&McMartin> | "According to legend, the El Torito CD/DVD extension to ISO 9660 gained its name because its design originated in an El Torito restaurant in Irvine, California" |
| 09:23 | <&McMartin> | AS FORETOLD BY THE PROPHECY |
| 10:08 | | catadroid [catalyst@Nightstar-5a1nvt.dab.02.net] has joined #code |
| 10:11 | | catadruid [catalyst@Nightstar-65o9q3.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 10:19 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 11:27 | < ToxicFrog> | McMartin: yes, but I have to switch it from "Legacy+UEFI" to "Legacy" boot mode to do so |
| 11:27 | < ToxicFrog> | I've never gotten the former to actually boot any "legacy" boot media |
| 12:12 | | catadroid [catalyst@Nightstar-5a1nvt.dab.02.net] has quit [[NS] Quit: Bye] |
| 13:58 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has quit [[NS] Quit: Leaving] |
| 13:58 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has joined #code |
| 13:58 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 14:46 | | gizmore [kvirc@Nightstar-g7eok3.dip0.t-ipconnect.de] has quit [Ping timeout: 121 seconds] |
| 15:53 | | celticminstrel [celticminst@Nightstar-pvpf2s.dsl.bell.ca] has joined #code |
| 15:53 | | mode/#code [+o celticminstrel] by ChanServ |
| 16:54 | | Fault [rouk@Nightstar-deqg8j.fbx.proxad.net] has joined #code |
| 16:55 | < Fault> | Any risks on performing this on my Arch machine? |
| 16:55 | < Fault> | http://askubuntu.com/questions/16998/switch-between-nvidia-current-and-nouveau-w ithout-a-reboot |
| 16:56 | < Fault> | The script probably works, I just want to know if it could e damaging to the hardware. |
| 17:11 | < Fault> | Folks at #nouveau say GPU freeze. |
| 17:26 | | Fault [rouk@Nightstar-deqg8j.fbx.proxad.net] has quit [[NS] Quit: leaving] |
| 19:09 | | catadroid [catalyst@Nightstar-25qdni.dab.02.net] has joined #code |
| 19:34 | | Kindamoody|afk is now known as Kindamoody |
| 19:43 | | Derakon[AFK] is now known as Derakon |
| 20:08 | | gizmore [kvirc@Nightstar-66794m.dip0.t-ipconnect.de] has joined #code |
| 21:14 | <&McMartin> | Hmmm. I can't actually find that BIOS config option. |
| 21:14 | | * McMartin burns a CD, runs his experiments the way he should have in the first place. |
| 21:15 | <&McMartin> | The results suggest QEMU's overhead is about 10x |
| 21:45 | <&McMartin> | Also, it's pretty great to see an old program whose comment says "This function takes approximately 45 minutes to plot" |
| 21:46 | <&McMartin> | and then plotting it in 2.04 sec |
| 21:46 | <@Tamber> | Ah, the march of progress |
| 21:49 | <&McMartin> | Not just of the processors but also of the interpreter used! I was using a GW-BASIC terp there that was probably a good five to ten years more modern than the program itself |
| 21:49 | <&McMartin> | Which among other things would mean that it likely would deign to use hardware floating point |
| 21:50 | <&McMartin> | But what was actually really interesting was in VirtualBox, or on the metal itself |
| 21:51 | <&McMartin> | There wasn't a tremendous difference in performance between enabling and disabling hardware floating point, nor again in swapping between various kinds of software floating point |
| 21:51 | <&McMartin> | (This was done with a translation to a compiled language with more options, a task I seem to have undertaken back in high school) |
| 21:52 | <&McMartin> | But in DOSBox, shifting between kinds of software floating point produces a 10x swing in performance |
| 21:52 | <@Tamber> | yikes |
| 21:52 | <&Derakon> | Look, math is hard, guys. :( |
| 21:53 | <&McMartin> | Well, I have a pretty strong idea of why this is, but this is the most dramatic example of it I've ever seen |
| 21:53 | <&McMartin> | DOSBox emulates the x86, but it - openly - makes no effort whatsoever to actually match historical CPU speeds |
| 21:53 | <&McMartin> | I suspect based on the stuff I've read that DOSBox having each instruction take a constant amount of time. |
| 21:54 | <&McMartin> | And I bet the huge swing is because one of those software FP implementations uses a large number of very fast instructions.' |
| 21:54 | <&McMartin> | Very fast *if executed for real*, that is. |
| 21:55 | <&McMartin> | QEMU does a binary translation technique that would roughly preserve it, and of course the hardware is the hardware[[*] |
| 21:55 | <&McMartin> | [*] actually the hardware is emulating the instructions in microcode at this point and has been since like 1987 but that is neither here nor there |
| 21:57 | <&McMartin> | Also, and perhaps cynically of me, I would also note that one of the other things that produced sudden and extremely dramatic performance gains Only In Dosbox was to switch to a then-contemporary C compiler |
| 21:58 | <&McMartin> | Which makes me wonder if they optimized for the case of "hey, this game I like is running too slowly" >_> |
| 21:58 | <&McMartin> | (But this also intrinsically involved a switch from 48-bit floating point to 64-bit) |
| 22:18 | < ToxicFrog> | McMartin: what model is your thoth-equivalent? Mine doesn't even have a CD drive, |
| 22:18 | <&McMartin> | Oh huh |
| 22:18 | < ToxicFrog> | Oh wait |
| 22:18 | <&McMartin> | It's a gazp7 |
| 22:18 | < ToxicFrog> | is it still the-- right |
| 22:19 | < ToxicFrog> | I retired that thoth and thoth is now a T550 |
| 22:19 | < ToxicFrog> | The gazp7 is now ancilla |
| 22:19 | <&McMartin> | Whoops, sorry for the confusion |
| 22:19 | < ToxicFrog> | And yes, it doesn't have a UEFI/BIOS toggle, it doesn't really speak UEFI |
| 22:19 | < ToxicFrog> | That said, I have successfully booted that stick on the gazp7 as well |
| 22:20 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 22:20 | <&McMartin> | Hm. I wonder if there's something bizarre about the *stick* |
| 22:20 | <@Alek> | hrm. I'm wondering. how good would a pi be for a learn-to-install-and-admin-linux unit? |
| 22:20 | <&McMartin> | Depends on which bits you want to learn, I suspect. |
| 22:21 | <&McMartin> | It's an ARM system with very particular hardware, so if you want to learn how to deal with installing Linux in a heterogeneous mostly-x86 environment, probably not great |
| 22:21 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 22:21 | <&McMartin> | If we're talking "how much of /usr/bin do I need to know to do work", though... |
| 22:21 | <@Alek> | basically, I need a second box because my main one needs to be windows, but I don't have room for a full size box, nor money for one or a laptop. |
| 22:22 | <&McMartin> | Hm |
| 22:22 | <&McMartin> | Is VirtualBox also not an option? |
| 22:22 | <&McMartin> | ToxicFrog: Do you at least recognize the "could not load COM32 file menu.c32" error I was getting at the boot-select screen? |
| 22:23 | <@Alek> | I'm talking learning to: install and configure a kernel and gui, run linux-based servers, stuff like that. |
| 22:23 | <&McMartin> | Alek: If you've got a few tens of GB free on the Windows box, setting up an Ubuntu or Fedora VM inside VirtualBox is the absolute lowest-cost option. |
| 22:23 | < ToxicFrog> | Alek: virtualbox and dual booting are both options that don't require an additional machine; the former is better for immersion, the latter lets you run it in parallel with windows and ssh into it from windows and stuff. |
| 22:23 | <@Alek> | although the ARM-only stuff vs x86 might be a problem. |
| 22:23 | <&McMartin> | For server operations, I'm not sure how suitable the Pi would be either. |
| 22:23 | < ToxicFrog> | If it's specifically servery things you're interested in, virtualbox will be a better choice than dual boot, I suspect, because then you can just run it "in the background" as if it were a separate machine you ssh into. |
| 22:23 | <@Alek> | I have enough trouble with the main box that I don't want to trust it with a VM, and I need to keep it running all the time so no dualboot. otherwise I would. |
| 22:23 | <&McMartin> | TF: Surely you got your former/latter switched |
| 22:24 | < ToxicFrog> | McMartin: I did, thank you |
| 22:24 | < ToxicFrog> | As for that error, I understand it but I have not personally encountered it |
| 22:24 | < ToxicFrog> | That said, it's entirely possible that my script messed up somehow! did it in fact put menu.c32 in the right place? |
| 22:24 | < ToxicFrog> | (that place being /boot/syslinux, IIRC) |
| 22:25 | <&McMartin> | There is in fact a 56-KB file named menu.c32 in /boot/syslinux |
| 22:25 | <@Alek> | hrm. my desk has been vibrating periodically all day, badly enough that I suspected even more HDD trouble, but then I realized I was feeling it everywhere in the house. now I blame the rollers grading the road outside. o_o |
| 22:26 | < ToxicFrog> | Alek: "don't want to trust it with a VM" in what way? |
| 22:26 | < ToxicFrog> | Like, you're planning to actually serve stuff and you don't want your server to go down if the host flakes out? |
| 22:28 | <@Alek> | both of my HDDs are old and somewhat iffy. the system one goes up to 100% load quite frequently when it shouldn't be. the second one is now my storage unit, but I don't exactly trust it either. and my CPU and ram are old too, the CPU older by a bit, and slow. |
| 22:29 | <&McMartin> | ToxicFrog: A more careful examination of the drive suggests that the issue is that mkfs.vfat is too clever for its own good |
| 22:29 | <@Alek> | medium-low-grade GPUs as well. |
| 22:29 | <&McMartin> | So, next question: How big was the stick you used for this? |
| 22:29 | <@Alek> | but that's not as important. |
| 22:29 | <@Alek> | actually, my GPUs are the newest and most powerful thing in the box, and they're still crap. :P |
| 22:29 | < ToxicFrog> | McMartin: ummm. I initially did development on a 64MB stick, but lately it's all been stuff in the 4GB-64GB range. |
| 22:29 | < ToxicFrog> | Why? |
| 22:30 | <&McMartin> | Because I've already had attempts at this fail due to FAT16 vs FAT32 and checking post-attempted-reformat this 128MB stick appears to be FAT16 |
| 22:30 | <@Alek> | ... |
| 22:30 | <&McMartin> | Which means syslinux may be doing a reinterpret-cast on its directory entries in a way that does not give valid information. |
| 22:31 | < ToxicFrog> | Alek: so, this doesn't really answer my question, which I guess I could rephrase as: are you planning to use linux for something that you need to be reliable? Or are you concerned that the VM will put too much load on an already-overloaded system? Or do you just worry that performance will be so bad as to be unpleasant to use? |
| 22:31 | <@Alek> | well, I also had puppy linux on a CD, but that's only as an emergency OS when my HDD breaks down. which has happened twice already. |
| 22:31 | <@Alek> | TF: all of the above. but more B+C than A. |
| 22:32 | <&McMartin> | C turns out to not be an issue as long as your computer is less than, oh, ten years old or so |
| 22:32 | <&[R]> | VMs can be given a maximum amount of RAM |
| 22:32 | <@Alek> | um. I built it in '10 from a combination of super-cheap parts and leftovers from prior machines. |
| 22:33 | <&McMartin> | The Intel VTx extensions (AMD has an equivalent but I forget what they are called) mean 32- or 64-bit OSes run with zero overhead except for hardware access |
| 22:33 | < ToxicFrog> | AMD-V |
| 22:33 | <@Alek> | then replaced the GPU later. latest was in... 14? maybe. |
| 22:34 | < ToxicFrog> | Initially released in 06 with the Athlon 64 and Pentium 4 |
| 22:34 | < ToxicFrog> | So if you have something more recent than that you're good |
| 22:35 | <&[R]> | I'm pretty sure cat /proc/cpuinfo... oh wait |
| 22:35 | <&[R]> | :/ |
| 22:35 | <&McMartin> | You can do that from inside the VM~ |
| 22:35 | <&McMartin> | ... I think |
| 22:35 | <&McMartin> | Maybe not |
| 22:35 | <&McMartin> | Recursive virtualization came later, IIRC |
| 22:36 | < ToxicFrog> | You definitely can on modern processor/vbox combinations |
| 22:36 | <&McMartin> | Yeah |
| 22:36 | < ToxicFrog> | VB might lie about how many cores there are but it will tell you the CPU family and feature flags and stuff. |
| 22:36 | <&McMartin> | But I'm less confident saying that that goes all the way back to the P4 |
| 22:38 | <&McMartin> | Because I remember we had some issues with this even in the 2010 timeframe |
| 22:38 | <@Alek> | my ram is stable, I think, but slow. 8gb of 400MHz DDR2. old dual-core E5400, 2.7GHz. and somehow HDD is often the bottleneck. |
| 22:39 | <&McMartin> | But, I don't remember if that's because the extensions have to Get Better, the Hypervisor needs to be smart enough to support it, or if in that era machines routinely disabled VTx by default in their BIOSes |
| 22:39 | <&McMartin> | Alek: Well, I mean, the HDD *is* the highest-latency device on that list. |
| 22:39 | < ToxicFrog> | This will, I think, be fine. The VM really doesn't add much overhead. |
| 22:39 | <&[R]> | HDD is pretty much always the bottleneck on Windows. Their HDD access is hilariously slow. |
| 22:41 | <&[R]> | Especially notable when it starts swapping |
| 22:41 | < ToxicFrog> | To be fair, everything is hilariously slow once it starts swapping |
| 22:41 | < ToxicFrog> | Unless you're using something like ramzswap |
| 22:41 | <@Alek> | McM: yeah well, I often get 100% load but the task manager doesn't say why, all the listed tasks are way low on disk usage. process manager gives a bit more detail, SUPER high (multi-second) latency, and apparently there's SUPERfrequent antivirus checking making most of the load, when I'm only swapping windows in chrome or something. |
| 22:42 | <@Alek> | my RAM load is usually pretty stable at about 4 or 5 GB of 8. |
| 22:43 | < ToxicFrog> | Anyways. You should be fine. Give the VM, say, 1 core, 1-2GB of RAM and 16GB of hard drive space, do a text-only install and go wild. |
| 22:44 | < ToxicFrog> | (this is not to say that a Pi can't be fun to play with, but you can get away without it) |
| 22:44 | <&[R]> | Be careful if you chose Ubuntu to start with though, it has a default option to install all optional packages, which means you'll quickly end up installing all the GUI crap. |
| 22:44 | < ToxicFrog> | This reminds me, at some point I need to fire up a VM and experiment with Nix. |
| 22:44 | <&[R]> | Ubuntu servers were the bane of my existance when I was doing PCI/DSS compliance. |
| 22:45 | <&[R]> | NixOS? |
| 22:45 | < ToxicFrog> | That's why I recommend SUSE :P |
| 22:45 | <&[R]> | Or just *NIX? |
| 22:45 | < ToxicFrog> | Yeah. |
| 22:45 | < ToxicFrog> | NixOS. |
| 22:45 | < ToxicFrog> | I have ample experience with UNIX :P |
| 22:45 | <&[R]> | :p |
| 22:45 | <@Alek> | hmmmm. thanks, I just might. would have to be on my second drive though, because of the load. |
| 22:45 | < ToxicFrog> | (although it's pretty out of date these days -- SysV and SCO) |
| 22:46 | <&[R]> | Isn't SysV the last of the old initds? |
| 22:46 | < ToxicFrog> | [R]: "sysvinit" is so named because it's the init system used on SysV UNIX. |
| 22:46 | <&[R]> | I /think/ that's what CRUX/Slackware use. |
| 22:46 | < ToxicFrog> | https://en.wikipedia.org/wiki/UNIX_System_V |
| 22:46 | <&[R]> | Oh you mean actual UNIXes |
| 22:46 | < ToxicFrog> | Yes. |
| 22:47 | <&[R]> | I thought you mean software running on them. |
| 22:47 | <&[R]> | :p |
| 22:58 | < ToxicFrog> | Alek: anyways, my recommendation would be SUSE, but I'm biased towards that because it's what I run |
| 23:08 | | catadruid [catalyst@Nightstar-15jcf8.dab.02.net] has joined #code |
| 23:11 | | catadroid [catalyst@Nightstar-25qdni.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 23:12 | <&[R]> | OpenSUSE presumably? |
| 23:13 | <&[R]> | Systemd? Y/N. Minimal install possible? Y/N. Package manager is? |
| 23:45 | < ToxicFrog> | OpenSUSE, yes. Yes systemd, yes minimal install possible, package manager is zypper, package backend is rpm. |
| 23:46 | < ToxicFrog> | Rolling release is optional (SUSE Tumbleweed); user repos are available via OBS. |
| --- Log closed Sun Oct 30 00:00:54 2016 |