Results 1 to 15 of 15

Thread: Advice on render speed cpu change for Lightwave 3d

  1. #1
    Registered User
    Join Date
    Aug 2006
    Location
    UK
    Posts
    251

    Advice on render speed cpu change for Lightwave 3d

    Just wanted some advice on this. I currently use a i7 5820k cpu 6 core 12 threads. If I upgraded to say an e5 10 or 12 core would I see a vast improvement in render speed in Lightwave with more cores, or would it only be marginal? I know the E5's don't run as fast and you can't over clock them so wondered what the rest of you guys thought.

  2. #2
    Newbie Member
    Join Date
    Feb 2009
    Location
    Germany
    Posts
    51
    I checked your values on a cinebench results list: https://www.cpu-monkey.com/en/cpu_be...5_multi_core-8

    Intel Core i7-5820K
    6x 3.30 GHz (3.60 GHz) HT
    1,085

    VS

    Intel Xeon E5-2687W v4
    12x 3.00 GHz (3.50 GHz) HT
    2,043

    compared with
    AMD Ryzen Threadripper 3990X
    64x 2.90 GHz (4.30 GHz) HT
    10,542

    This means a larger result is better and you need to check which type of cpu within your budget gets most points.

    I would check several benchmarks if you run different applications on your machine. Cinebench is cpu based.
    And if it is ment to be an upgrade, please make sure the type of cpu fits into your machine!

    And you could check the LW 2019 benchmark post in this forum and compare the results. But you would need to check all posts there, because there is no compiled result list.
    https://forums.newtek.com/showthread...ight=benchmark
    Last edited by slacer; 05-22-2020 at 01:36 AM.

  3. #3
    Registered User
    Join Date
    Aug 2016
    Location
    a place
    Posts
    2,354
    running a dual e5 2670 48 threads. better for render speed but slower at everything else, including modeller due to single core speed. difference is noticeable compared to an i7 with 4ghz single core speed.

    id go for either :

    Intel Core i9-9900K/KF
    Intel Core i9-10980XE
    Intel Core i9-10900K

    AMD Ryzen 9 3950X
    AMD Threadripper 3970X
    AMD Threadripper 3990X

    this vray table should give some ideas about speed:
    https://www.cgdirector.com/vray-benchmark/

    this one for photoshop/after effects:
    https://www.cgdirector.com/best-comp...after-effects/

    gpu there too:
    https://www.cgdirector.com/octaneben...hmark-results/

    you will see how the cheaper but fast single core cpus do.

    Another one with a quick pic

    https://www.pugetsystems.com/recomme...ecommendations



    Click image for larger version. 

Name:	benchmark.PNG 
Views:	319 
Size:	173.8 KB 
ID:	147928
    Last edited by gar26lw; 05-22-2020 at 02:27 AM.

  4. #4
    3900x is probably your most cost-effective option as of today. I recently upgraded from an i7-5820k and it's literally 3x faster, plus single core performance is slightly higher as well (though not really noticeable in practice). Just swap out the CPU+Mobo and will be about $480 + the time it takes to replace those components.

    the price will also depend slightly on whether or not you want multi-GPU support, overclocking, and other misc features of fancier mobos.
    Professional-level 3d training: Ryan's Lightwave Learning
    Plugin Developer: RR Tools for Lightwave

  5. #5
    I'm running the 2990WX but LW 2015 seems to be very unstable when running all 32 cores so to use 2015 I have to reboot with 16 cores active and it runs fine. LW 2018, 2019 & 2020 all seem to run great and take full advantage of all 32 cores.
    Threadripper 2990WX, X399 MSI MEG Creation, 64GB 2400Mhz RAM, GTX 1070 Ti 8GB

    https://www.dynamicrenderings.com/

  6. #6
    Founding member raymondtrace's Avatar
    Join Date
    May 2003
    Location
    Ohio
    Posts
    1,155
    Quote Originally Posted by Nicolas Jordan View Post
    ...to use 2015 I have to reboot with 16 cores active...
    This can save you from needing to reboot...

    https://www.youtube.com/watch?v=ht4bV8IuWcM
    LW4, 7.5D, 2015, 2018, 2019, 2020 running portably on a USB drive on an Amiga 2500 running Wine.

  7. #7
    Registered User
    Join Date
    May 2003
    Location
    Europe
    Posts
    278
    New cheaper AMD mainboards coming soon, the B550:
    https://wccftech.com/amd-b550-mother...zen-4000-cpus/

    The AMD 12 cored 3900x is dropping in price now, to counter Intels "new" 10 core 5 years old 10900k:

    https://wccftech.com/amd-slashes-ryz...ration-launch/

    But the 16core 3950x is a good deal considering that it runs on cheap mainboards.

  8. #8
    Registered User
    Join Date
    Aug 2006
    Location
    UK
    Posts
    251
    Thanks for all the info guys, my head is buzzing lol. I think for now I'm just looking to see if any of those e5 2650 v3 that will fit my X99 board are worth it. I notice there are some 12 core ones going for just over £100 online. I was wondering if I would see any render speed benefit from swapping out my 5820k for something like this? Or whether it's not worth the effort?

  9. #9
    Registered User
    Join Date
    May 2003
    Location
    Europe
    Posts
    278
    Sell your rig to get cash for a cheap AMD rig, the 12 core 3900X is getting a faster friend, the 3900XT, so the 3900X is getting cheaper:

    https://wccftech.com/amd-ryzen-9-390...pus-confirmed/

    More than 2x faster than old xeon...

  10. #10
    Registered User
    Join Date
    May 2003
    Location
    Europe
    Posts
    278
    Have you overclocked your 5820k? They are supposed to reach 4.4Ghz easily, giving you 4,4x6=26,4GHz of rendering power. The Xeons usually stick to the base frequency at heavy loading, so I guess that gives you 10x2,3=23GHz of rendering power. Roughly the same and a lot slower single core performance for all other apps and operations. Overclock to the max or buy a new AMD CPU/mainboard.

  11. #11
    Registered User
    Join Date
    Aug 2006
    Location
    UK
    Posts
    251
    Quote Originally Posted by Free4Ever View Post
    Have you overclocked your 5820k? They are supposed to reach 4.4Ghz easily, giving you 4,4x6=26,4GHz of rendering power. The Xeons usually stick to the base frequency at heavy loading, so I guess that gives you 10x2,3=23GHz of rendering power. Roughly the same and a lot slower single core performance for all other apps and operations. Overclock to the max or buy a new AMD CPU/mainboard.
    So at a slower speed even though the xeon has 12 cores= 24 threads it will still be slower rendering than my 5820k 6 core 12 threads? I have in the past overclocked my 5820k to 4.1Ghz which it was stable at. It all depends on the chip and the cooling to get 4.4Ghz but I will try at 4.00Ghz again. The only reason I dropped back to stock is my Gigabyte X99 is a bit of a fickle board. It will happily run for months with no problem and then. Play dead once in a blue moon. The remedy has always been to reset the bios by removing the battery., but that's another subject. I also have 16gb of ram but I understand that more ram doesn't speed up renders.

  12. #12
    Registered User
    Join Date
    May 2003
    Location
    Europe
    Posts
    278
    Yes, 6 cores at 4,6GHz renders just as fast as 12 cores at half the speed of 2,3GHz. Google and YouTube how to overclock X99 and 5820k and save your money for now. The Gigabyte should overclock easy to 4.4Ghz below 1.30V for the CPU for both stability and decent temps. You just need to learn how to overclock and do stability testing.

  13. #13
    Registered User
    Join Date
    Aug 2006
    Location
    UK
    Posts
    251
    Thanks for that. Will have a search around for how to do this.

  14. #14
    Registered User
    Join Date
    Aug 2006
    Location
    UK
    Posts
    251
    I tried overclocking to 4.4Ghz but it only gave me a 2 second difference in rendering the same scene.
    I was surprised at how little difference it made in render speed. So I might as well stick at a safe stock speed.
    Last edited by genesis1; 05-24-2020 at 11:17 AM.

  15. #15
    Electron wrangler jwiede's Avatar
    Join Date
    Aug 2007
    Location
    San Jose, CA
    Posts
    6,970
    Quote Originally Posted by genesis1 View Post
    I tried overclocking to 4.4Ghz but it only gave me a 2 second difference in rendering the same scene.
    I was surprised at how little difference it made in render speed. So I might as well stick at a safe stock speed.
    It isn't just about frequency, it's also about amount of cache, cache architecture, memory interface channels & frequency, having modern instruction sets, and having other modern features (larger predictor cache, data streamers, write-combining ranges, etc.). You can't just base it on frequency comparisons, those other factors can have huge, significant impact on performance comparisons.
    John W.
    LW2015.3UB/2019.1.5 on MacPro(12C/24T/10.13.6),64GB RAM, NV 980ti

Bookmarks

Posting Permissions

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