Page 2 of 12 FirstFirst 123456 ... LastLast
Results 11 to 20 of 114
Like Tree45Likes

Thread: Revit 2016 .ini file

  1. #11
    Forum Co-Founder Twiceroadsfool's Avatar
    Join Date
    December 7, 2010
    Location
    Dallas, TX
    Posts
    9,673
    Current Local Time
    12:29 PM
    It occurs to me that this means we should ramp up the Benchmark. I could include a ridiculously complex Beck Model for benchmarking, and you could viewcube around a bunch of views. And get some decent numbers with HWA on and off.

  2. #12
    Administrator Gordon Price's Avatar
    Join Date
    December 7, 2010
    Location
    Rotterdam, South Holland
    Posts
    3,034
    Current Local Time
    07:29 PM
    Well, I have an idea for a "heavy" test. A template that contains 8 links of the final model in a grid around a central space where the new model gets built. Now if I can also make sure animated transitions are forced on, we may be getting close. That's my "travel game" on the way to RTCAUS.

    EDIT: The main concern I have is keeping the overall benchmark download reasonably small. Thus the idea of a single file with the final model, then linked in to the 3x3 grid. Maybe a 5x5 grid if needed to make things heavy enough. What do you think? Is a link a meaningful extra load, or does it have to be local to the test file to really push things? For that matter, maybe some "Benchmarking the Benchmark" is in order. Try it with a 5x5 grid of links and one with in-situ model, and see if the times are affected.

    Gordon

  3. #13
    Forum Co-Founder Twiceroadsfool's Avatar
    Join Date
    December 7, 2010
    Location
    Dallas, TX
    Posts
    9,673
    Current Local Time
    12:29 PM
    My main thought is that file doesn't have nearly enough stuff in it to be a convincing tests even if it's linked in 8 times. But I guess you're right about the download...

  4. #14
    Forum Addict elton williams's Avatar
    Join Date
    December 7, 2010
    Location
    Gold Coast
    Posts
    2,451
    Current Local Time
    04:29 AM
    Quote Originally Posted by Gordon Price View Post
    Custom colors also show up in the color picker, so you can seed those 16 squares with office vetted colors for lots of uses. I like providing 8 pastel/earthy colors and a matching 8 more vibrant variants, ready for use is all sorts of cases, especially custom room and area stuff. Users can quickly choose a color, and the color is part of a consistent graphic language.
    Brilliant, I never knew this, thanks!

  5. #15
    Forum Co-Founder iru69's Avatar
    Join Date
    December 7, 2010
    Location
    United States
    Posts
    1,615
    Current Local Time
    10:29 AM
    Regarding 2016 benchmarking (and maybe Gordon would like to split this off topic into another thread), I have a few thoughts...

    1. What makes the benchmark interesting to me is how a CPU (and GPU if that were possible) performs with certain Revit tasks (and in a few cases how less common CPUs performs vs. common CPUs, e.g. a Haswell-EP Xeon E5-series vs. a Haswell i7-series with similar clock speeds).

    2. Therefor, Id like to see more emphasis on the modeling portion of the benchmark currently, 90% of whats interesting about the benchmark (i.e. primarily the modeling portion) only makes up about a third of the total time of the benchmark run. On top of that, many of the sub-sections of the modeling test only last a handful of seconds... there's too much margin of error in such a small amount of time. And the two subsections of exporting DWGs and PNGs aren’t super critical relevant modeling tasks, yet take up a hugely disproportionate amount of what little modeling benchmark run there is.

    3. The “render” portion can actually be a little shorter if we needed to keep the total benchmark run down to a reasonable time. It’s a generic test that doesn’t indicate anything more than any multi-threaded CPU benchmark test does (e.g. Geekbench).

    4. OTTH, the experimental GPU "benchmark" portion is confusing in what it indicates, but often takes more than a third of the benchmark runtime. I would can it altogether for now and spend the benchmark time on expanding the modeling portion... or it could be reduced to one or two views and included as part of the modelling benchmark, since it's mostly testing CPU as far as I can tell.

    5. Since there's been a very large margin of error at times, it would be great if the benchmark would automatically run 2-3 times consecutively and only keep the lowest result. It doesn't need to do any fancy results averaging, etc. - we're only interested in the fastest data set, i.e. how fast a system is able to perform under optimal conditions. For a relatively new computer, the whole multi-run benchmark should be able to complete in about 15 minutes. It's of course understandable if that amount of "programming" is beyond the time Gordon wants to put into it.

    6. Posting in the results thread: I know some think embedding the results as a "code quote" looks cool or more professional or something, but as someone who actually examines the results on a regular basis, it makes it a major PITA to try to quickly find and compare results in different posts. Could we please discourage this practice in the 2016 results thread?

    Gordon, let me know if I can help.
    Last edited by iru69; April 26th, 2015 at 01:26 AM.

  6. #16
    Member
    Join Date
    March 31, 2011
    Posts
    62
    Current Local Time
    02:29 PM
    Please explain what animated transitions are. Thanks

  7. #17
    Member BD Mackey's Avatar
    Join Date
    November 1, 2013
    Posts
    466
    Current Local Time
    11:29 AM
    Quote Originally Posted by rboaz View Post
    Please explain what animated transitions are. Thanks
    When using the view cube, by clicking on a corner face etc. it will animate the rotation. With this off it will just change to the new location, slower machines can be slowed down by the animation.

  8. #18
    Moderator cellophane's Avatar
    Join Date
    August 9, 2011
    Location
    Louisville, KY
    Posts
    6,000
    Current Local Time
    01:29 PM
    Quote Originally Posted by Twiceroadsfool View Post

    [Directories]
    ProjectPath=C:\Temp
    ImportLineweightsNameDWG=\\dalfs\Revit\2016\Beck Project Template\ImportLayers-NCS40.txt

    [DirectoriesENU]
    FamilyTemplatePath=\\dalfs\Revit\2016\Beck Family Templates
    DataLibraryLocations=Families=\\dalfs\Revit\2016\B eck Family Content
    DefaultTemplate=Beck Project=\\dalfs\Revit\2016\Beck Project Template\Beck-Template-R2016.rte
    What's the difference between Directories and DirectoriesENU?

  9. #19
    Administrator Gordon Price's Avatar
    Join Date
    December 7, 2010
    Location
    Rotterdam, South Holland
    Posts
    3,034
    Current Local Time
    07:29 PM
    Quote Originally Posted by cellophane View Post
    What's the difference between Directories and DirectoriesENU?
    Directories is for common (non language specific) resources. DirectoriesENU is for content that is English specific.

  10. #20
    Moderator cellophane's Avatar
    Join Date
    August 9, 2011
    Location
    Louisville, KY
    Posts
    6,000
    Current Local Time
    01:29 PM
    Quote Originally Posted by Gordon Price View Post
    Directories is for common (non language specific) resources. DirectoriesENU is for content that is English specific.
    I knew that...

Page 2 of 12 FirstFirst 123456 ... LastLast

Similar Threads

  1. Revit 2016
    By Dsmelcer in forum Out There
    Replies: 78
    Last Post: May 27th, 2016, 01:55 AM
  2. Revit MEP 2016 Fabrication Parts
    By Chris.kershaw in forum MEP - General
    Replies: 10
    Last Post: May 4th, 2015, 06:22 PM
  3. Revit 2016 and .rcp link problem
    By mihan in forum Architecture and General Revit Questions
    Replies: 1
    Last Post: April 23rd, 2015, 02:45 PM
  4. Replies: 0
    Last Post: March 29th, 2015, 10:45 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •