Wikipedia:Reference desk/Archives/Computing/2014 February 14

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

February 14

After a memory upgrade.

I have Dell R5400, it is a rack mount 64 bit workstation. For years I had 4 GB of RAM, then I began setting virtual machines and as a result came to the conclusion that this memory is not enough. So I ordered 8 GB of memory (two chips 4 GB each from Kingston via Amazon). Now I installed the chips however I am not sure I put the new chips into correct slots. The existing 2 GB chips were locked into two receptacles with white braces (staggered arrangement), I moved them both to the outer slots (one white, the other black brace). I put two new chips (4 GB each) together close to the center of the motherboard and they ended up in two slots, one with white, the other with black brace which lock them in place.

I checked the BIOS, everything seems to be OK however a momentary warning message first appeared and I lost it. Now I am looking at Control Panel==>System and Security==>System. This is what it says:

Installed memory (RAM): 12 GB (8 GB usable).

The question is why? Did I mess up the installation? I've been planning to purchase another 8 GB to make it 16 on brand new chips but now have second thoughts.

I am looking at the Dell Setup and Quick Reference Guide. It says

Memory capacities 256 MB, 512 MB, 1 GB, 2 GB, 4 GB, and 8 GB; Minimum memory 256 MB; Maximum memory 32 GB

Thus I am within legal limits. Why is "usable memory" not 12 GB?

Please help. Thanks --AboutFace 22 (talk) 01:28, 14 February 2014 (UTC)

The DIMM slots are in color coded pairs- a black pair and a white pair. You need to put the 2G DIMMs in one pair and the 8G DIMMs in the other pair. The DIMMs is each pair of slots must be identical. It is recommended that all four be the same for best performance. --  Gadget850 ^talk 01:43, 14 February 2014 (UTC)

Thank you much. You are very fast. --AboutFace 22 (talk) 02:04, 14 February 2014 (UTC)

What makes Linux the OS of choice for the TOP500?

Almost all of the Top500 computers use Linux. Not even Unix seems as a serious competitor, and Windows certainly not. What's under the hood, that makes Linux excel over all the others? OsmanRF34 (talk) 13:11, 14 February 2014 (UTC)

Did you follow the link to Supercomputer operating systems? It looks like a pretty good overview. In the end, the key is that Linux is free and open source, so it is easy to tweak to the exact needs of their systems. They don't usually run straight Linux, but a customized Linux-based system. Katie R (talk) 13:53, 14 February 2014 (UTC)

The question is not what caused it, but what is in Linux that runs differently. And as a second point, that still doesn't explain why is Unix so underrepresented, when it was, not long ago, the OS of choice in the field. Here we are talking about corporations, which have access to resources to tweak Unix as much as they want. OsmanRF34 (talk) 14:24, 14 February 2014 (UTC)

Yes, but it's still a lot more expensive to pay for Unix variants. Large corporations have the money to spend if they want, but will still be thrifty when there's no downside. Also, for any of the computers that are used in a research setting, many academic researchers put a value on using open source tools, and may be obligated to release code to the public when they publish results (many sources of public funding, such as NIH, NSF in the USA, come with obligation to release data and code). Distributing code that is modified from a Unix proprietary system would be problematic. SemanticMantis (talk) 15:10, 14 February 2014 (UTC)

It's not quite the same as supercomputer requirements, but the fully open source is why I am using Linux for a new project at work. I need a system based on off-the-shelf PC hardware that can communicate with a PCIe board we're developing, and in some situations the response time must be on the order of a few microseconds. The completely open kernel lets me account for every bit of delay along the execution path if I run into situations where things are running slow. With various UNIXes this is possible, but to get that level of source access or to modify sources would require expensive contracts. Integrating GPL code from somewhere else into a closed-source kernel would be impossible, so if I find something that does what I need in the open-source world I wouldn't be able to integrate it into our product, instead we would have to find a vendor that offers a non-GPL solution or develop it in-house. I would expect that the same kinds of issues would push supercomputer software engineers towards Linux as well, they just may be tweaking the kernel for different reasons than I do. Katie R (talk) 17:51, 14 February 2014 (UTC)

Because Microsoft keeps putting one foot in front of the other and unloading a full belt of machinegun ammo into it. http://www.theregister.co.uk/2013/12/12/feature_microsoft_caught_in_virtual_monkey_trap/ Hcobb (talk) 18:01, 14 February 2014 (UTC)

Spacing in MAC finder windows (Maverick)

Hello all! I recently switched from Mountain Lion to Maverick. The issue that I'm having (well one of the issues, but I don't want to state them all b/c forest for trees and all that) is that when I open up a finder window, the categories of data are hugely widely spaced. For example, I open up a window and see my folders, then, farther to the right in the finder window I have categories for "Date modified", "Size" and "Kind" and others but they are way to the right with huge whitespace in between each one, and I have to drag the window much wider and then manually drag the categories closer together so they fit in one view in a reasonably sized finder window. Is there any way to set how far over they are when a finder window opens so I don't have this issue? Thanks!--71.167.166.18 (talk) 15:07, 14 February 2014 (UTC)

I don't regularly use a Mac, but I searched for Mac Mavericks Finder list view columns to investigate a little bit for you. From what I'm reading, it looks like this is a bug. Finder is supposed to remember the column widths when you re-open the window but doesn't. Until Apple fixes the bug, here are some ideas I found from others having the same issue:

• In this thread (original seems to be down, so see Google cache), there's a suggestion that pressing Esc twice (to toggle to Full Screen view and back) will restore the Name column to a normal size, however, that person notes that only works for folders on internal drives not external drives.
• In this thread, a person seems to be describing adding many additional columns (right-click on a column header and check several items). It sounds like this will cause the overall space to be divided among more columns making each one smaller, but I'm confused by their description.

--Bavi H (talk) 04:01, 15 February 2014 (UTC)

Windows folder column widths

I've had the same problem with Windows windows. Ideally they would adjust their column widths to be no wider than the widest item in that column, when the window is first opened. Alternatively each window could remember the final column spacing the last time it was opened, and recall that. However, differing data widths each run might make that approach less useful. StuRat (talk) 17:38, 14 February 2014 (UTC)

I split the Windows comments into a subsection.

I mainly use Windows XP, so I'm not sure if my experience will be the same for newer version of Windows. I find that Windows does remember column widths and other view settings for each folder, but there's a limited number of folders it can remember. After I've opened many folders, the least recently used ones forget their settings.

Filenames can sometimes get really long. If opening a folder automatically resized the columns to show everything you might end up often resizing them smaller. Perhaps that's a reason why Microsoft hasn't implemented that.

Here are some tips for automatically resizing columns:

• You can double-click on the "resizing line" at the end of a column to automatically resize just that column.
• You can press Ctrl and the plus key on the number pad to automatically resize all columns.
• In Windows 7 and Windows 8, you can right click a column header and choose "Size Column to Fit" or "Size All Columns to Fit".

--Bavi H (talk) 04:01, 15 February 2014 (UTC)

I just remembered there's a setting in Windows XP that controls if Windows remembers each folder's settings. In Windows XP, open any folder, go to the Tools menu, choose Folder Options, go to the View tab, and look for "Remember each folder's view settings". StuRat, maybe this option is turned off on your computer. On my Windows 8 computer, this option isn't there, so I think Windows 8 will remember each folder's settings and won't let you disable it. --Bavi H (talk) 04:20, 15 February 2014 (UTC)

How about on Windows 7 ? StuRat (talk) 04:31, 15 February 2014 (UTC)

I guessed other Windows versions inbetween would either behave like Windows XP (have the setting) or Windows 8 (not have the setting but always remember). When I search for "Windows 7" "Remember each folder's view settings" I see message threads saying Windows 7 doesn't have the setting and doesn't remember folders' settings. How frustrating. :(

In this thread, I found a suggestion that if you add a folder to the Favorites in Explorer, then the folder's view settings are remembered. Otherwise, it sounds like Windows 7 was designed to only let you customize five folder "templates". --Bavi H (talk) 06:36, 15 February 2014 (UTC)

In the same thread, this post by NMGyrl sounds interesting:

Aside from the special case of Libraries, the folder view settings are associated with whatever path to the folder was in effect when you set them. Once you've set up the view you like for a given folder, use the same navigation path to get to it next time, and it will look the way you left it. Clearly using the Favorites shortcut will make this easier, but if you have a lot of folders where you want different views, it may be cleanest just to always use the "real" location path, i.e. starting with the drive letter, to get to everything.

But I'm unsure if this means navigating from My Computer to your desired folder every time, entering the folder's address in the Win+R Run box, and/or using a shortcut icon to the folder. --Bavi H (talk) 06:51, 15 February 2014 (UTC)

Yep, and if all that manual navigation is required, that would be more work than just readjusting the columns each time. StuRat (talk) 19:56, 15 February 2014 (UTC)

The easiest way to deploy a local DNS proxy/server?

I need recommendation on a simple solution for deploying a local DNS server or proxy. What I want to do is to allow local DNS names to be used, in addition to global ones. I can't spend too much time on setting it up, so a simple-to-deploy solution is better than one that is feature-rich but complicated. Thanks in advance. --108.36.90.200 (talk) 15:53, 14 February 2014 (UTC)

Create a hosts (file) on each machine. If you want to synchronize updates to this for n machines, set up a cron job or a scheduled task to copy the latest canonical hosts file from your server. Nimur (talk) 03:19, 15 February 2014 (UTC)

Using hosts files for name resolution is not an option for some of the devices on the network. That's one of the reasons I'm looking for a way to set up a local DNS server. --108.36.90.200 (talk) 07:32, 15 February 2014 (UTC)

You can run BIND on almost anything... on Linux, on OS X... but you might find it easiest to set up a NetBSD DNS server. Your DHCP server can point to that DNS server; alternately, you can manually configure each client. You might also read "Setting up TCP/IP in Practice from NetBSD's documentation; it goes over the DHCP set-up. Nimur (talk) 15:50, 15 February 2014 (UTC)