steve_v

Related to bug http://bugs.kerbalspaceprogram.com/issues/11382? Fullscreen in settings.cfg? Fix unity prefs file and set it read-only?

What kernel version are you running, and has it been upgraded to >=4.8 recently? As of 4.8 several distros have deprecated the vsyscall interface on amd64, you might try adding 'vsyscall=emulate' to your kernel boot parameters.

Sure you can:

In light of the recent console port debacle, I think this is a terrible idea. Performance on a high-end PC is... not fantastic. Performance on the switch would be awfull, and If history is any guide, the port would be a mess.

Rather than reducing gimbal range, what about just disabling yaw and roll on the vectors? You should have enough yaw/roll authority from your aero surfaces, at least for a while...

Er, that link be 404.

I get very good mileage out of RS Components online, cheap as and they do free (and often overnight) shipping where I live. Might be worth checking if this applies to you too.

Sure, for Windows. Which I don't really care much about. But so what? It's still slower under every other API (including Vulkan, which is the future) and any game you might want to play today. As well as that, it's less power efficient... to the extent of pulling more than it's stated specification under load and exceeding the PCIE spec for want of another aux power connector.

How hard did you have to look to find a benchmark that puts the 460 "neck and neck" with a 1050? I did a bunch of research before purchasing my 1070 and I couldn't find a single review that (averaging across benchmarks) has a 460 matching a 1050, or a 480 matching a 1060... And as far as I can tell, AMD has nothing at all that can touch the 1070 or 1080. ---- Ed. I can see why you only linked the image, not the full review... The 460 loses out in every other test. Look, I can cherry-pick a the same review too: ---- Speculative "driver updates" are speculative, and I'll believe the "eventually will outperform" bit when if it "eventuates". Why are you insisting that these cards have equivalent performance, when the rest of the internet says otherwise? They're certainly cheaper, but benchmarks say they're also slower... And not exactly better value.

Any chance of doing the same for M3X?

Land on the moon? Of course we did, that's where cheese comes from. Land in the moon... I think not. Landing in the moon would be most silly, not unlike this thread. Unless the moon is hollow, but we'd have to ask Wells... I vote to not vote, the question is far too silly.

If you get them from a proper electronics supplier, they will specify intensity in millicandela or lumens, if not a full data sheet. Your problem is that you are buying electronic components from a model shop.

-nolog

Less energy, yes. By explosive range I mean explosive at a wide range of concentrations - roughly 18-60% in air. Hydrogen also has a fairly low ignition energy, so a small spark is all that is required to ignite such a mixture. Don't. Make a little, make it go pop, move on. You won't generate enough to run a rocket, and I can't really think of any other use for it. A small quantity of hydrogen (and only hydrogen) might be stored in a plastic bag or balloon for a short time (it will leak through many plastics anyway), but storing larger quantities or for longer periods is highly dangerous and requires specialised equipment. Many. Pure oxygen will make things that normally won't burn a serious fire hazard, and things that normally burn explode. Again, fine in small amounts, but don't try to manufacture or store it in quantity at home. All in all, a small electrolysis experiment is fairly harmless, just keep volumes small and your work area well ventilated.

Dude, industrial rig running 6 thousand amperes. That failure mode is just not possible with a few 9v batteries. I would suggest keeping the bag small. I've not tried it with hydrogen, but I know someone who blew out all the windows in one side of a house with a garbage bag of acetylene/oxygen mix. Not smart. A bread bag of oxy/acetylene on the other hand... Hydrogen isn't as energy dense as most other fuel gasses, but it has a very broad explosive range, so be careful with it.

FWIW, the metadata on CKAN has M2X depending on "Interstellar Fuel Switch", rather than "Interstellar Fuel Switch Core" - so when installed via CKAN it will pull in the full version of IFS, complete with parts and "complementary" patches.

So long as you stay under about 60v you can't give yourself a shock, though you may get a tingle if your hands are wet. As for fire risk, I assuime you're talking about carbon-zinc or alkaline 9v batteries - in which case the internal resistance should limit current enough to prevent anything catastrophic (although you could probably still start a fire if you tried). So yeah, don't series any more than 6 or so, try not to short them, and you'll be fine.

Eh, it's not that big an outlay next to two current-gen cards. If you're going to SLI 10xx series cards, you're probably not going for cost as a primary concern. Benchmarks are showing a significant performance benefit to the added bandwidth of the HB bridge, and if there's that much data being transferred, it's a pretty good call not putting it on the PCIE bus IMO. If you're willing to forgo the extra bandwidth, the old flexible bridges still work too - and they're only ~10USD.

RAID0 stripes data across multiple devices, so reading or writing happens in parallel, with only part of the file read from/written to each drive. 2 drives = 1/2 the work for each device = 2x performance. It's simple physics, nothing theoretical about it. Why not? If the storage subsystem is faster, things waiting on storage get faster... whether that makes a difference to you depends on where the bottleneck is. What's wrong with bridges? If you have no dedicated inter-GPU channel then that data is going on the PCIE bus, and you're going to be using more PCIE bandwidth. I don't see how this could possibly be an advantage, except in cutting production cost. Then again, AMD does seem to like to abuse that PCIE slot... On the whole though, SLI/Crossfire is a bit of a mixed bag, sometimes it's great, sometimes it doesn't work properly (or at all), and sometimes it brings new and strange bugs. Heavily dependent on drivers and application support.

Assuming RAID0, you'll get ~2x the performance of a single SSD or two separate SSDs. As to whether you actually need a storage system that fast... in most cases you don't. Maybe if you do a lot of video editing.

It's not data-loss that's the issue, it's MTTR. With softraid or a redundant filesystem it's simply a case of plugging the disks into another machine, or replacing the controller with whatever comes to hand (or you can do full multipath failover if uptime is really important). With a sufficiently large array restoring a backup can be a lengthy process, particularly if it's write constrained parity-raid and doubly so if that backup is off-site or on tape.

This is most certainly true of spinning rust, though NCQ and plenty of RAM for disk cache can alleviate this problem somewhat. IME it's not really an issue on a decent SSD unless you are running properly I/O intensive applications. If I had two or more identical drives in my desktop, I'd almost certainly use them in RAID - RAID0 will give a much bigger performance boost than a separate OS disk (as will RAID1, for reads at least), and like any sane person, I have a robust (and well tested) backup solution. I'm not doing so at the moment because the SSDs I have are not the same size, and the second is a recent addition... also because I am too lazy to change it right now. I'm not sure I'd recommend RAID to Windows users though, my experiences with Windows soft-raid (at least in non-server editions) have been unpleasant to say the least. Most reasonably priced addon RAID controllers (or BIOS softRAID), aren't much better IME, and introduce dependency on a specific controller - if your card (or motherboard) dies, so does your array... permanently if you can't find a compatible replacement. On that note, I have been recommended DrivePool as an alternative to softraid on Windows (for redundancy, not performance), but I could never get it to meet all my requirements - primarily hotspares, automatic "just plug it in" disk replacement and graceful (read: fully automatic resync) detach / re-attach of redundant disks. Anyone used it seriously? I know how to do what I want with both MDRAID and ZFS, but I'm not seeing a solution for Windows that doesn't involve expensive controllers or expensive server licences.

It does, but only if you actually have multiple GPUs, which I don't. And yes, GPU dynamic clocking is working as expected.

There are many ways to circumvent corporate poxys, but where there is a filtering poxy, there's usually a policy that specifically forbids circumventing it. How alert the IT department and their monitoring software is for such things varies considerably. While it's possible to make it pretty much indistinguishable from legitimate traffic, that generally requires software on the client... And going to these lengths implies landing in rather hot water if caught. I feel your frustration, and I've often been in this situation myself. With a locked down machine I usually find the only real option (other than grovelling to the admin) is either a remote desktop to a machine I control (if allowed) or a separate data connection, i.e. tethering a phone, or bringing in a laptop with a cell modem.

AFAIK, already implemented: Automagic dewarp radio button in GUI.