The follow me tool uses only 1 profile edge at a time. You can make a zero thickness rect with the radius edges and use the follow me and then force the bottom face to form or if you do it with thickness then select only one edge to follow.
There is a plugin you could use but you should probably learn how to do this manually.
I did not check but you have many segs in the corner arcs and I don't know if that maybe contributing to your problem. I did re-draw your posted example.
You also need to check your models. Turn on hidden lines and you will note you have some that should not be there. Hint: in area where side is pulled up but not the only spot
If you clean your model things work fine. Learn how to get the tangent inference with the ARC tool. If you go to styles, edges and select color by axis it will help you find the areas where there are model problems.

FYI Su8 M2 release changes the may that works now. Look at release notes and search for that topic.
Soory I don't have the URL handy

There are a couple of tutorials in this forum aslo
http://forums.sketchucation.com/viewtopic.php?f=18&t=723
http://forums.sketchucation.com/viewtopic.php?f=18&t=752&p=79190&hilit=+How+a+pro+builds+a+house+in+Su+pt+2+#p79190

Some miss information floating around:

1. Windows memory manager controls the access to physical memory via working sets, most programs cannot touch physical memory, and in fact virtual memory can be larger that 4GB;
2. By setting the 3GB switch in the boot ini file for certain configurations of your system then SU can have 3GB of memory( Less what some device drivers use.) vs the normal two while the kernel gets reduced to 1 GB. I was able to open an ~350MB SU file ( Window Xp sp3, 4GB Ram) using this approach. Does not mean speed was great.
3. For the 64 bit systems SU is still limited to the 4GB Address space because it runs as a 32 bit app. However the operating system can remap memory to use RAM greater than 4GB so the app. can have the total 4GB. Your BIOS needs to be set correctly to max use of your 64 system. Make sure to read the manual.
   For those interested read these articles http://blogs.technet.com/b/markrussinovich/archive/2008/11/17/3155406.aspx and http://blogs.technet.com/b/markrussinovich/archive/2008/07/21/3092070.aspx

Here is the link. http://www.google.com/support/forum/p/sketchup/thread?tid=2979d2e0f28d0eed&hl=en

window, preferences, short cuts=> export. Wo3dan posted a ruby consul command that will give more but I need to search for his post.

It can be confusing so read the user manual very carefully , but it sounds to me they are doing what I would expect. There is a slight difference how they behave if you insert directly from the component browser vs to the dwg and then insert. One issue I seem to face all the time is when I set the axis how does it really get inserted. SU shows the greyed rectangle and the three axis with the cross but it shows the same if you set with a right vs left hand coordinates. The usual convention is either with a circle and center dot or the cross. The former indicates you are looking back toward the origin ( arrow head of the axis) or the other way around. I finally made a simple model and went thru a number of test cases. If not done already you may want to give that a try. You should also look at the capability of selecting inserted components , going to the component browser, selecting a component and then select "replace selected".

john2
Can you post your model.
When one is trying to clean a model there usually are tolerances associated with that operation which you can change and the results can be a function of the sequence of steps you follow. For example if I merge "close" vertices there is a tolerance with what close is and I can actually change my geo. with that operation. There use to be associated with the suddivide and smooth a video that shows that possibel artifact along model boundary. Like Thomthom notes under stand what changes are being made.

Whyile you are looking check to see if use memory remapping also there.

Bummer, like dealing with the government since it is almost assured they have all kinds of performance metrics etc. and they could probably easily set one specific to for you. For your info: From here to the site 400 sec test shows 30(16) hops and no drops at this time.

@thomthom said:

Well, as it is with shared hosting, there could be any number of sources for any issue. I've been monitoring the site and trying to put down some notes. Will try to make a test script that times the various execution time for different functions. Need some kind of pointer to give the support monkeys.

If you have not done so already I would strongly recommend you talk to the tech support. They may already have an idea because most will get some performance metrics or the can suggest TS steps??

Yes I understand that ,but at the same time others are OK. How does your postulated failure fit their results? Don't get stove piped and fool yourself. I would guess you host should have some test they can run or comment on TS approach and maybe correlate with any recent changes?? They can even run SNR test and make sure levels are ok. If you have ever had the chance to scope a high data rate line it is not what you think of as nice square waves. It is more sinusoidal with jitter at the zero crossing etc, so the sample clock must be such that you are not trying to work in the linear region. ON top of that teflon under goes a state change at room temp and can cause timing errors unless they are properly match for phase delay and lot control. The point is a lots of things can go on to affect a high rate system. Have you made any recent changes?

Don't think this is the problem but a lots of ACK-Nacks can cause that can it not? If the folks who downloaded and found no problem go thru the same sequence as you what does that tell you? Either they are lucky or there is some difference. For a while I was getting SQL errors on this site just because of two many users.
Good Luck

@thomthom said:

When I looked at the control panel and tried to access database related tools they seemed to be slow as well. So did downloading the databases - which I do regularly as backups. It took many seconds before I got any response.

Ping test of your site showed an average of ~53ms. For example this site shows 145 ms but probably can just be indicating the number of hops. Suprised either of you allow that to happen
You can also go to the control panel, admin tools, performance and you can check the TCP object and plot various items. I would look at retransmits and that would indicate if something in your loop is casuing data errors.

For the BIOS, run msconjfig, select the boot tab and advanced, You may have a selection box that is use max memory or such. Unselect that other wise it is telling the kernel to use max memory vs what it needs, That may free some more memory for SU to use. Is that a soild state drive you are showing

Have you tried running any ping test to get sates? http://www.livinginternet.com/i/ia_tools_ping.htm There are also sites that will do free test. You may have to run at different times etc. to get a good base line.

Couple of thoughts;

1. You state you have 6 GB of of RAM So I assune that means you have the 64 bit version of Windows 7 I hope. Other wise you can only address 4GB of that;
2. SU runs as a 32 bit app and will only use a single core. W7 64 bit and as ,with W 32 bits, device driver memory uses some of the RAM. With W7 64 SU can use 4 GB of Ram unlike only 2GB for w 32 with 4 Gb of RAM. I have seen posting here where 2 Monitors with 1 GB each is used . You can check out what is being used by using the device manager, selecting resource by connection and then memory. That will show what is using what. The down side is the report is in hex so use the windows calculator and you can convert the numbers to decimal.
   You are on the correct path=> rendering is the main cause of slowness because the GP must re-calcualte for every change in view. So use :monochrome as much as possible; layers turned off: keep poloygon count down; you can use scenes to help with the orbit to spots in your model. See this http://sites.google.com/site/sketchupsage/faster.These do not help with the startup because the program needs to load from disk which is relative slow.
   There are some settings in the BIOs that may also help but will have to run that down for you. ALso check task manager and see what RAM it is reporting?

@irwanwr said:

that is what those links Dave provided are all about, isn't it?

Yes but I thought it would be helpful to flag the use of the component browser capability also.

@jgb said:

Mac1; You are very close to the actual panel, except for a curve that you could not know of.

Please note; I am not looking for a solution. I have it. The original post was for Fredo to investigate why I get a black face but an empty group.

I've attached copies of the actual problem panel and the adjacent panel with notes. A 2nd copy has the problem face deleted and a C-Loft empty group attached.

FYI, these panels are carbon fiber composite honeycomb sandwich, 1/2 in thick. At this point in the design development, the stiffeners and structural attachments are not yet incorporated. To all you aero engineers, these panels are NOT primary structure, and are subject only to aerodynamic loads. Even so, they are stronger than they have to be, considering the planned stiffeners/internal structure that attaches them to the primary a/c structure. As such, they are considered fairings.

Note jpg of strut deleted by poster.12/28-2011 2240 MST
FYI some GY70 near zero CTE tapered and curved strut. In my experience based aeroshell and fairing two differnt designs. Spent many hours on antenna range with aeroshell to reduce radar cross section so that at post planetary entry and jettison landing radar did not think it was a big target.
Glad you have solution I still get same results face forms ok.
Gday

Is this some what the goal?
As I noted the 1/2" face is not flat so you cannot push pull a soften or curved face. TIGs extrude tools maybe better for this problem but need to see the total issue.
This was done loft by spline but just one side of the surface. To get the thickness it will require two steps and then reconnect the ends.

@jgb said:

@mac1 said:

I get the two lines also and is the reason I used the skinning option and get the same result as Fredo. BTW I measure 0.5 in apart and looks like by design ???

The 2 lines you refer to are the inner and outer edges of a very large 1/2 inch thick panel which is part of a fairing made of several adjoining panels. The 1st master panel itself (outer skin) was originally formed by C-Loft, then thickened with JPP to 1/2 inch. It formed all 4 edges. All the panels are rectangular (4 distinct corners) with that similar pronounced curve.

Then I selected 1 edge to become the curve control for this adjacent panel, which had a different overall shape at the top, but shared an edge. I formed that panel the same way, but had to modify it for the design. In doing so, I lost the 2 side edge faces and 1 top edge face (expected). I then closed the top 2 corners. One edge immediately skinned, but this edge and the top did not. I expected the top not to skin as it had a slightly irregular profile, but this edge should have auto skinned because it was skinned prior to the mod, and the top end was redrawn. I used C-Loft to re-skin the top edge (worked 1st time) but this edge persistently refused. I have since manually skinned it in the model.

If you want I can post the complete panel here, in its present state, but I doubt it will contribute anything to this discussion.

Please do.
I think we are working at cross purpose withoout seeing the whole picture