#dynamic loading

[jbk405]

I know it's a heretical opinion in the Mass Effect fandom, but I actually really like the elevator mechanic in the first game. I wish they'd kept it in the sequels.

It really helped the verisimilitude of being there in the game. Some areas on the Citadel had multiple elevator connections to different levels, and it felt like you were really traversing a city since you could find shortcuts and different stuff along the way. Plus, I adore the squadmate conversations that would occur. They gave great character insights, and were often entertaining all on their own. The news reports were less engrossing, but still a nice touch.

I agree they could have structured them better so they weren't so noticeably long, but I think it was a great way to fill the loading time. And Rapid Transit was available for people who didn't want to run through multiple elevators.

[basicittopic]

Logical- Versus Physical-Address Space

Logical- Versus Physical-Address Space

Logical versus physical Address, The logical address is a virtual address. It can be viewed by the user. The user can’t see the physical address directly. The logical address is used as a reference, to access the physical address.

An address generated by the CPU is commonly referred to as a logical address or a virtual address whereas an address seen by the…

View On WordPress

[engineeringtldr]

We started looking at fluctuating loads last time - that is, loads that feature some combination of non-zero mean and alternating stresses - and how to account for them using a Goodman diagram. Let’s re-examine the bracket design problem we did earlier. This time, instead of a fully-reversed load, we’ll assume a fluctuating load with a mean force of 200 lbs, a minimum force of 50 lbs, and a maximum force of 350 lbs. We’ll say the dimensions of the bracket are those we calculated earlier that could handle the fully reversed load. (Problem adapted from Machine Design: An Integrated Approach, 4th Ed., by Robert L. Norton.)

Most of the calculations we did earlier will still hold. We won’t need to recalculate the endurance limit or stress concentration factors. The only new things we need to do are calculate the mean and alternating stresses and the new safety factors.

First step is to calculate the mean and alternating force.

From here, we get the mean and alternating moment.

We’re dealing with a situation of simple bending, so we can calculate mean and alternating stress using the basic bending stress equation.

The geometry of the part hasn’t changed, so we’ll apply the same stress concentration factors that we used before.

Great. We’ve got our new stresses. Now we need to figure out safety factors. As we mentioned earlier, this is now a slightly more complicated proposition. Which safety factor is appropriate will depend on how the alternating and mean stress behave in relation to each other. The possible failure states are shown as points A, B, C, and D on the Goodman diagram for this situation.

We’ll step through all the possible situations one by one using the new stresses we calculated and the endurance limit we got earlier.

Case 1: Constant alternating stress, variable mean stress.

Case 2: Variable alternating stress, constant mean stress.

Case 3: Alternating and mean stress are proportional to each other.

Case 4: Alternating and mean stress vary independently.

We take the worse case, with the failure state F being as close as possible to the current stress situation.

Our design will survive all four cases. Note that Case 4 is always the most conservative case - if you don’t know what your stresses are going to do, this is the one to go with.

[engineeringtldr]

We need to design a bracket to take some kind of repeated bending load. One end is clamped stationary between a couple of plates, and a uniaxial, fully-reversed load with a maximum of 300 lbf rides on the end at a distance of a away from where it’s clamped. There’s a fillet where it clamps into the wall to try and keep it from fretting on a sharp corner. We’ll say it’s machined steel with an ultimate tensile strength of 82000 psi and an elastic modulus of 28 x 10^6 psi. We need it to survive basically forever. We aren’t given any dimensions, so we need to figure out exactly what this part has to look like. (Problem adapted from Machine Design: An Integrated Approach, 4th Ed., by Robert L. Norton.)

This is a pretty complicated problem - although we’re given a basic shape, we need to design this piece more or less from scratch. Here’s the basic process we’ll follow for this and other fatigue design problems:

Initial estimated design

Calculate nominal stresses

Apply fatigue stress concentration factor

Calculate principal stresses

Calculate corrected endurance limit or fatigue strength

Calculate safety factor

Modify design

This gets long and ugly, but none of the individual pieces are too bad.

1. Initial Estimate

We’re trying to design this piece from the ground up - we know what it’s going to look like, but we have no dimensions. We need some numbers to start with, so we’ll just pick some dimensions that sound like they might work to start with. Here’s what we’ll go with for an initial guess:

2. Calculate nominal stresses

The only stress we have going on here is a bending stress. We’ll calculate the moment using the maximum force applied (the alternating stress, in this case) and apply the basic bending stress formula, using the moment of inertia for a rectangular cross-section.

3. Apply fatigue stress concentration factor

We’ll get a static stress concentration factor from the dimensions we’ve selected and pick it off a chart like this one:

We’ll also pick notch sensitivity, q, from a chart in this case.

Knowing these, we can calculate the fatigue stress concentration factor and apply it to the nominal bending stress we found earlier.

4. Calculate principal stresses

In this case, we don’t have to calculate principal stresses - we have a situation of uniaxial stress so our maximum principal stress will be equal to the nominal alternating bending stress we found. If we had a situation of combined loading (say, if we also had a torsional component to this load), we’d need to do this step.

5. Calculate corrected endurance limit

We’ll do this exactly the same way we did before. I won’t run through the details now, but you can see them here.

Note that this is only about a quarter of our ultimate tensile strength.

6. Calculate safety factor

It’s obvious that the first parameters we guessed won’t do the job. If we compare the maximum stress our part will experience to its fatigue strength, we get a safety factor of about 0.4:

7. Modify design

It’s clear that we need to make some changes. We’ve got a number of options. We could change the material if our dimensions were constrained for some reason. We could increase the radius of the fillet to try and take some of the edge off that stress concentration factor. We could increase the thickness and width of the beam to increase the moment of inertia, and consequently reduce the bending stress. In fact, we definitely want to start with that last one, since our nominal stress is greater than the material’s endurance limit, even before we apply a stress concentration factor. Suppose we go with these dimensions instead:

We’ll need to recalculate the nominal bending stress. This looks much better.

We also need to calculate a new stress concentration factor, since our dimensions have changed. Here’s what we get:

Nothing about the endurance limit has changed, so we can go ahead and calculate a new maximum stress taking into account the stress concentration factor and compare it to the endurance limit.

This try gets us a safety factor greater than one, so it looks like this version is a viable design.

[engineeringtldr]

Okay. Let’s try putting some of this fatigue stuff to use. We’re given a rod with a circular cross-section of 1.00″ diameter, subject to a cyclical, fully-reversed bending load in an environment that will see a maximum sustained temperature of 300 degrees Fahrenheit. The material is a steel with a 180 kpsi ultimate tensile strength and a machined surface finish. We need this part to have 99% reliability and to last forever. Given those conditions, what’s the maximum stress we could subject it to?

Since we need infinite life, we know we’re talking about a material with an endurance limit. That is, a material with an S-N diagram that will look something like this:

We need to find S_e, the stress it can survive for an infinitely long period of time. To do that, we’ll need to account for environmental factors and properties of the part that impact its ability to stand up to fatigue loading. We’ve got this general equation to start with:

We can get S_e’, the uncorrected tensile strength, pretty easily. For a material with a tensile strength of less than 200 kpsi, we can estimate it as half the ultimate tensile strength.

We’re left with a bunch of correction factors to figure out. We’ll tackle them one-by-one.

We’ve said that we’re dealing with a bending load, so C_load will be 1. We have a circular cross-section of 1.00″ diameter, so we can calculate C_size with the formula we had earlier.

We’re told that we have a machined surface and an ultimate tensile strength of 180 kpsi. Using those, we can look up C_surface using a chart. It’s about 0.66 here.

(Image from Machine Design: An Integrated Approach, 4th Ed., by Robert L. Norton.)

The temperature is under 840 degrees Fahrenheit, so C_temperature is just 1.

We’re told we need 99% reliability. We can get C_reliability from a chart. For us, it’ll be 0.814.

(Image from Machine Design: An Integrated Approach, 4th Ed., by Robert L. Norton.)

We’ve got all the correction factors we need, so now we just need to plug them in to get our endurance limit.

It’s worth noting just how small that 42 kpsi is compared to the ultimate tensile strength of the material. If we want this thing to survive forever, we can’t load it beyond about 23% of what it would survive in straight static tension. Suppose we decide that, well, okay, maybe we don’t really need infinite life. Maybe we can get away with just 10^4 cycles. How much would that increase the load we could put on it?

Let’s take another look at that S-N diagram.

10^4 cycles occurs in that sloped region between S_e and S_m. To figure out what kind of strength we can expect, we’ll need to figure out that slope.

We can estimate S_m based off the ultimate tensile strength. For a bending load, it’s about 90% of S_ut.

We know that the slope has this form:

We have formulas for a and b, so we can just plug in and get a general equation for that slope.

I rounded here, so this isn’t quite exact, but it’ll get us pretty close. If we plug in 10^4 cycles, here’s what we get:

[engineeringtldr]

More Than You Ever Wanted to Know About Mechanical Engineering, Part 29: S-N Diagrams

We’ve been talking about fatigue strength and endurance limits and the factors that affect them. However, we know that fatigue strength isn’t just a single number - it depends on the number of cycles the part is subjected to. A cyclical load will have to be higher to break a part at low cycles (higher fatigue strength) than at high cycles (lower fatigue strength). So we need to figure out some way to determine what the fatigue strength or endurance limit should be, taking into account the particular conditions affecting our part.

One of the ways to do this is with an S-N diagram. It’ll look like this:

Note that this is a chart where both vertical and horizontal axes are on a logarithmic scale.

Drawing these is pretty simple. You just need to know two points: Sm and Se or Sf. Sm is an estimated strength at 10^3 cycles. You can generalize for this. It’s usually safe to assume the following values:

Se and Sf are the corrected endurance limit or fatigue strength we figured out earlier. If it’s an endurance limit, you plot it at N = 10^6 cycles, since that’s usually where an endurance limit kicks in for materials that have them. If it’s a fatigue strength, you plot it whatever number of cycles the fatigue strength was given for.  You plot Sm at 10^3 cycles, and Se or Sf at the number of cycles appropriate to them and draw a straight line connecting them. If it’s Se, you draw a horizontal line after, and if it’s Sf, you just continue the original line you drew. Now you’ve got a chart that will let you pick out a particular stress and see how many cycles it corresponds to.

If you don’t want to draw the whole diagram out, you can also get an equation for the line between Sm and Se or Sf and figure out cycles for a particular strength algebraically. Since this is a log-log chart, the equation for the line is going to have this general form:

Or:

We just need the constants a and b to make this into something usable. Here are the formulas formulas for these.

Knowing these, you can just plug into that general equation and solve for a particular number of cycles.

[moccosk]

New Post has been published on Mocco

New Post has been published on http://mocco.sk/using-ngx_pagespeed-with-nginx-on-debian-wheezy/

Using ngx_pagespeed With nginx On Debian Wheezy

Version 1.0 Author: Falko Timme <ft [at] falkotimme [dot] com> Follow me on Twitter Last edited 04/29/2013

Google has just released PageSpeed Beta for nginx, the nginx equivalent of mod_pagespeed for Apache. This module applies web performance best practices to pages, and associated assets (CSS, JavaScript, images) and therefore speeds up your web site and reduces load times. This tutorial explains how to use PageSpeed with nginx on Debian Wheezy.

I do not issue any guarantee that this will work for you!

1 Building nginx With PageSpeed

nginx does not support dynamic loading of modules, therefore we must rebuild it with PageSpeed support. We will build nginx as a .deb package so that we can replace our existing nginx installation with it (or easily install it on other systems).

Create the build directory first:

cd /usr/src mkdir pagespeed && cd pagespeed

Make sure you have deb and deb-src lines for Wheezy in /etc/apt/sources.list:

vi /etc/apt/sources.list

[...] deb http://ftp.de.debian.org/debian/ wheezy main deb-src http://ftp.de.debian.org/debian/ wheezy main [...]

Update your packages list and install some prerequisites:

apt-get update

apt-get install dpkg-dev build-essential zlib1g-dev libpcre3 libpcre3-dev

Download the nginx sources and the build dependencies:

apt-get source nginx

apt-get build-dep nginx

Let’s check our nginx version:

ls -l

root@server1:/usr/src/pagespeed# ls -l total 2044 drwxr-xr-x 10 root root    4096 Apr 29 13:03 nginx-1.2.1 -rw-r–r–  1 root root 1360507 Aug  5  2012 nginx_1.2.1-2.2.debian.tar.gz -rw-r–r–  1 root root    2740 Aug  5  2012 nginx_1.2.1-2.2.dsc -rw-r–r–  1 root root  718161 Jun 27  2012 nginx_1.2.1.orig.tar.gz root@server1:/usr/src/pagespeed#

Our nginx version is 1.2.1. Let’s go to the debian/modules directory of the downloaded nginx sources…

cd /usr/src/pagespeed/nginx-1.2.1/debian/modules

… and download the PageSpeed sources:

git clone https://github.com/pagespeed/ngx_pagespeed.git

Let’s edit the debian/rules file:

vi /usr/src/pagespeed/nginx-1.2.1/debian/rules

In that file you will find four config.status sections (config.status.full, config.status.light, config.status.extras, config.status.naxsi). Add the line –add-module=$(MODULESDIR)/ngx_pagespeed \ to each of them, right before the $(CONFIGURE_OPTS) >$@ line, e.g. as follows:

[...] config.status.full: config.env.full config.sub config.guess cd $(BUILDDIR_full) && CFLAGS="$(CFLAGS)" CORE_LINK="$(LDFLAGS)" ./configure \ --prefix=/etc/nginx \ --conf-path=/etc/nginx/nginx.conf \ --error-log-path=/var/log/nginx/error.log \ --http-client-body-temp-path=/var/lib/nginx/body \ --http-fastcgi-temp-path=/var/lib/nginx/fastcgi \ --http-log-path=/var/log/nginx/access.log \ --http-proxy-temp-path=/var/lib/nginx/proxy \ --http-scgi-temp-path=/var/lib/nginx/scgi \ --http-uwsgi-temp-path=/var/lib/nginx/uwsgi \ --lock-path=/var/lock/nginx.lock \ --pid-path=/var/run/nginx.pid \ --with-pcre-jit \ --with-debug \ --with-http_addition_module \ --with-http_dav_module \ --with-http_geoip_module \ --with-http_gzip_static_module \ --with-http_image_filter_module \ --with-http_realip_module \ --with-http_stub_status_module \ --with-http_ssl_module \ --with-http_sub_module \ --with-http_xslt_module \ --with-ipv6 \ --with-sha1=/usr/include/openssl \ --with-md5=/usr/include/openssl \ --with-mail \ --with-mail_ssl_module \ --add-module=$(MODULESDIR)/nginx-auth-pam \ --add-module=$(MODULESDIR)/nginx-echo \ --add-module=$(MODULESDIR)/nginx-upstream-fair \ --add-module=$(MODULESDIR)/nginx-dav-ext-module \ --add-module=$(MODULESDIR)/ngx_pagespeed \ $(CONFIGURE_OPTS) >$@ touch $@ [...]

Now let’s build our new nginx packages:

cd /usr/src/pagespeed/nginx-1.2.1/ && dpkg-buildpackage -b

Afterwards, let’s take a look at the generated packages:

cd /usr/src/pagespeed ls -l

root@server1:/usr/src/pagespeed# ls -l total 35464 drwxr-xr-x 10 root root    4096 Apr 29 13:12 nginx-1.2.1 -rw-r–r–  1 root root   61006 Apr 29 13:12 nginx_1.2.1-2.2_all.deb -rw-r–r–  1 root root    5143 Apr 29 13:12 nginx_1.2.1-2.2_amd64.changes -rw-r–r–  1 root root 1360507 Aug  5  2012 nginx_1.2.1-2.2.debian.tar.gz -rw-r–r–  1 root root    2740 Aug  5  2012 nginx_1.2.1-2.2.dsc -rw-r–r–  1 root root  718161 Jun 27  2012 nginx_1.2.1.orig.tar.gz -rw-r–r–  1 root root   72632 Apr 29 13:12 nginx-common_1.2.1-2.2_all.deb -rw-r–r–  1 root root   74042 Apr 29 13:12 nginx-doc_1.2.1-2.2_all.deb -rw-r–r–  1 root root 4140470 Apr 29 13:12 nginx-extras_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 6075706 Apr 29 13:12 nginx-extras-dbg_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 3931344 Apr 29 13:12 nginx-full_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 4609688 Apr 29 13:12 nginx-full-dbg_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 3815222 Apr 29 13:12 nginx-light_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 3546806 Apr 29 13:12 nginx-light-dbg_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 3853762 Apr 29 13:12 nginx-naxsi_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root 3673646 Apr 29 13:12 nginx-naxsi-dbg_1.2.1-2.2_amd64.deb -rw-r–r–  1 root root  342008 Apr 29 13:12 nginx-naxsi-ui_1.2.1-2.2_all.deb root@server1:/usr/src/pagespeed#

We can now install nginx with PageSpeed support as follows:

dpkg –install nginx-common_1.2.1-2.2_all.deb nginx_1.2.1-2.2_all.deb nginx-full_1.2.1-2.2_amd64.deb

Let’s check if the PageSpeed module was built successfully:

nginx -V

You should see the ngx_pagespeed module in the output:

root@server1:/usr/src/pagespeed# nginx -V nginx version: nginx/1.2.1 TLS SNI support enabled configure arguments: –prefix=/etc/nginx –conf-path=/etc/nginx/nginx.conf –error-log-path=/var/log/nginx/error.log –http-client-body-temp-path=/var/lib/nginx/body –http-fastcgi-temp-path=/var/lib/nginx/fastcgi –http-log-path=/var/log/nginx/access.log –http-proxy-temp-path=/var/lib/nginx/proxy –http-scgi-temp-path=/var/lib/nginx/scgi –http-uwsgi-temp-path=/var/lib/nginx/uwsgi –lock-path=/var/lock/nginx.lock –pid-path=/var/run/nginx.pid –with-pcre-jit –with-debug –with-http_addition_module –with-http_dav_module –with-http_geoip_module –with-http_gzip_static_module –with-http_image_filter_module –with-http_realip_module –with-http_stub_status_module –with-http_ssl_module –with-http_sub_module –with-http_xslt_module –with-ipv6 –with-sha1=/usr/include/openssl –with-md5=/usr/include/openssl –with-mail –with-mail_ssl_module –add-module=/usr/src/pagespeed/nginx-1.2.1/debian/modules/nginx-auth-pam –add-module=/usr/src/pagespeed/nginx-1.2.1/debian/modules/nginx-echo –add-module=/usr/src/pagespeed/nginx-1.2.1/debian/modules/nginx-upstream-fair –add-module=/usr/src/pagespeed/nginx-1.2.1/debian/modules/nginx-dav-ext-module –add-module=/usr/src/pagespeed/nginx-1.2.1/debian/modules/ngx_pagespeed root@server1:/usr/src/pagespeed#

2 Configuring PageSpeed

To enable PageSpeed, open /etc/nginx/nginx.conf…

vi /etc/nginx/nginx.conf

… and add the lines pagespeed on; and pagespeed FileCachePath /var/ngx_pagespeed_cache; before any vhosts:

[...] pagespeed on; pagespeed FileCachePath /var/ngx_pagespeed_cache; ## # Virtual Host Configs ## include /etc/nginx/conf.d/*.conf; include /etc/nginx/sites-enabled/*; [...]

Create the cache directory and restart nginx (please note that whenever you modify the PageSpeed configuration, no matter if in the main nginx configuration or in a vhost, you must restart nginx – a simple reload does not work):

mkdir /var/ngx_pagespeed_cache chown -R www-data:www-data /var/ngx_pagespeed_cache

/etc/init.d/nginx restart

root@server1:~# /etc/init.d/nginx restart Restarting nginx: Setting option from (“on”) Setting option from (“FileCachePath”, “/var/ngx_pagespeed_cache”) nginx. root@server1:~#

Let’s load a page and check if PageSpeed is mentioned in the output:

curl -I -p http://localhost|grep X-Page-Speed

root@server1:~# curl -I -p http://localhost|grep X-Page-Speed   % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current                                  Dload  Upload   Total   Spent    Left  Speed   0     0    0     0    0     0      0      0 –:–:– –:–:– –:–:–     0 X-Page-Speed: 1.5.27.1-2845 root@server1:~#

Now we can configure PageSpeed individually or each vhost, e.g. like this:

vi /etc/nginx/sites-available/example.com.vhost

server { [...] # let's speed up PageSpeed by storing it in the super duper fast memcached pagespeed MemcachedThreads 1; pagespeed MemcachedServers "localhost:11211"; # Filter settings pagespeed RewriteLevel CoreFilters; pagespeed EnableFilters collapse_whitespace,remove_comments; # Ensure requests for pagespeed optimized resources go to the pagespeed # handler and no extraneous headers get set. location ~ "\.pagespeed\.([a-z]\.)?[a-z]{2}\.[^.]{10}\.[^.]+" { add_header "" ""; } location ~ "^/ngx_pagespeed_static/" { } location ~ "^/ngx_pagespeed_beacon$" { } location /ngx_pagespeed_statistics { allow 127.0.0.1; deny all; } location /ngx_pagespeed_message { allow 127.0.0.1; deny all; } [...] }

The important line is the pagespeed EnableFilters line which tells PageSpeed which filters it should apply. You can find a list of all filters here: http://ngxpagespeed.com/ngx_pagespeed_example/

Don’t forget to restart nginx afterwards:

/etc/init.d/nginx restart

When you open a page now and take a look at the headers (e.g. with the Live HTTP headers add-on for FireFox), you should see an x-page-speed line:

(JavaScript must be enabled in your browser to view the large image as an image overlay.)

You can check the HTML sources of your pages to see if the PageSpeed filters work as expected.

3 Links

About The Author

Falko Timme is the owner of Timme Hosting (ultra-fast nginx web hosting). He is the lead maintainer of HowtoForge (since 2005) and one of the core developers of ISPConfig (since 2000). He has also contributed to the O’Reilly book “Linux System Administration”.

Check out the original source here.