Jump to content

KSP and hyperthreading


jkool702

Recommended Posts

In not sure if this sub-forum is the right place for this, but my question is:

Does hyperthreading (on intel CPU's) make a difference in performance in KSP?

Since KSP performance is more-or-less limited by single thread CPU performance it seems like splitting a single physical core into 2x logical cores might have detrimental effects on performance?

I looked into this issue for running matlab codes that were compute limited a while back and pretty much came to the answer that "System resource monitors will show 1/2 as much CPU utilization with hyperthreading enabled, but performance is essentially the same." Does anyone know how this effects KSP?

Link to comment
Share on other sites

Does hyperthreading (on intel CPU's) make a difference in performance in KSP?
Nope.
Since KSP performance is more-or-less limited by single thread CPU performance it seems like splitting a single physical core into 2x logical cores might have detrimental effects on performance?
It does but in my tests I didn't notice a difference.

Windows shifts a thread from one CPU core to the next, sometimes a hundred times per seconds. Depending on the program the performance can go down by a lot.

KSP has one thread per subsystem. A subsystem can be physics, audio, graphics, etc. The big ressource hog is the physics thread.

Unity 5 introduces multi-threaded physics. That means KSP 1.1 will probably run a bit faster that than pre-1.1.

Link to comment
Share on other sites

Thanks for the reply. I kind of expected that hyperthreading made little to no difference, but figured someone out there had actually tested it out.

And yes I'm very hopeful that 1.1 will bring nice performance gains. My i7 seems severely under-utilized right now :/

Link to comment
Share on other sites

Some Intel (and other) cores have extra hardware that doesn't (always) get used by a single thread. These cores can expose this extra hardware as a "virtual" core, but that core shares resources with the "real" core, meaning that depending on what is going on, the "virtual" core may or may not be able to be utilized fully. That's why your mathlab code didn't run any faster, because it either wasn't multithreaded, or if it was, the threads were competing for the shared resource (which is, I believe, generally an FPU, or Floating Point Unit, which does floating point math, and which KSP would use heavily...).

Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...