Jump to content
HWBOT Community Forums

ADVenturePO

Members
 View Profile  See their activity

  • Posts

    191

  • Joined

  • Last visited

 Content Type 

Posts posted by ADVenturePO

    GPUPI 3.2 and 3.3.2 (64bit) not working with AMD Ryzen 3900x

    in Ryzen | Bristol Ridge AM4

    Posted

    @KaRta

    vcredist which one - those are Visual Studio - I have 2012 .2013. 2015 in x86 and x64.  :) 
    Please check which one do You have in the system for AMD Ryzen 3900x GPUPI 3.2 and 3.3.2. 
    Thank You. 

    @StingerYar - the card is NVidia GT 710, drivers are present and up to date :) 

    -----------
    I have tried with 2012 versions only, 2013 versions only and 2015 versions only. 
    2013 x86 is required to run Legacy versions. 
    None of the standard version works :/ Hmmm.

    GPUPI 3.2 and 3.3.2 (64bit) not working with AMD Ryzen 3900x

    in Ryzen | Bristol Ridge AM4

    Posted

    Dear Fellow OC'ers.
    @StingerYar @safedisk
    I find myself in a strange position with this CPU / system W10.

    Everything is up to date. But I can't run GPUPI 3.2 
    Despite having Visual Studio 2012 , 2013 & 2015 I'm still unable to run standard version.
    After Installing Visual Studio 2013 x64 I have run Legacy version.

    I have seen You both running those 64bit versions just fine. 
    Is there something else I should've installed?

    Cheers

     

    Peter
     

    GPUPI 3.2 fail.png

    ADVenturePO - 4x GeForce RTX 2080 Ti @ 2100/7600MHz - 0sec 978ms GPUPI - 1B

    in Result Discussions

    Posted

    Dear @avalanche 
    Please, get some knowledge... from Google maybe?
    https://hwbot.org/submission/3786224_h2o_vs_ln2_gpupi___1b_4x_titan_v_0sec_991ms/  - 2 times more pricey cards and chiller.
    https://hwbot.org/submission/3769407_kingpin_gpupi___1b_4x_titan_v_0sec_996ms/ - 2 times more pricey cards and Liquid Nitrogen

    Please compare specs. 
    I wrote you how many different timings and CPU speeds I have tested to get best workflow with those cards.
     

    Render Config

      RTX 2080 Ti TITAN V
    ROPs 88 96
    SM Count 68 80
    Shading Units 4352 5120
    TMUs 272 320
    Tensor Cores 544 640

    Theoretical Performance

      RTX 2080 Ti TITAN V
    FP16 (half) performance 26,895 GFLOPS (2:1) 29,798 GFLOPS (2:1)
    FP32 (float) performance 13,448 GFLOPS 14,899 GFLOPS
    FP64 (double) performance 420.2 GFLOPS (1:32) 7,450 GFLOPS (1:2)
    Pixel Rate 136.0 GPixel/s 139.7 GPixel/s
    Texture Rate 420.2 GTexel/s 465.6 GTexel/s

     

    Also: be aware, that in Octane, Vray and Redshift for a single card records are mine and done with Titan V.
    FE: Octane : 467 pts with Titan V vs 367 with RTX 2080 TI. 

    Judging by the price and specs Titan V is better. But I never had 4x Titan V with LC at once. 
     
    Cheers. 

    GPUPI Multi GPU records

    in Submission & member moderation

    Posted · Edited by Leeghoofd


    And even more @avalanche
    I was beating this GPUPI scores for 3 days.
    I have invested in CPU LC Loop over 1000$ more to beat them.
    I have tested those GPUs with CPU at multipliers 44, 45 , 46 and 47 with RAMs 3200 - 4000Mhz with Timings like
    14-14-14-28, 16-16-16-28, 16-16-16-32, 17-17-17-34

    And my score is taken by 10 and 16 cards. 
    No... there was absolutely no effort in defeating H2o vs LN2 and K|NGP|N.

     

    GPUPI Multi GPU records

    in Submission & member moderation

    Posted

    #avalanche, 
    Really?
    You're 100% not in that business. If you'd be, you'd know why this 10 card 2080 Ti server didn't make a score in 32GB. (overheating or too low power supply)
    You would also know how much time it take to even start a benchmark with 16 cards connected to single system. :)

    My business is my business, my passion is something else. 
    I'm not selling servers. ?

    If my product is weak than why the hell 10 cards didn't make it 2x faster? ?
    Why did I defeat 4 Quadro 6000, 8000 nad Tesla cards in all rendering engines?

    I repeat: it was shitty construction with no OC. No effort and only one benchmark. 

    Read this avalanche: I earned 1st places correctly. Why the hell am I not getting points due to the gap to wrongly made score?

    If my desire to be in the 1st place is wrong, why the hell are we even trying to beat those scores here?

    Funny that you're making that point here... well you are my direct competition with 1362 points in Enthusiast League... (1311 is my score)
    If I'll get those points for 1st places in those benchmarks I will be higher in points than you are.
    Something smells bad here avalanche.

    Aren't you playing your own game here?

     

    GPUPI Multi GPU records

    in Submission & member moderation

    Posted

    Hello OCers and moderators. @Leeghoofd @Massman @richba5tard

     

     @Christian Ney @TheMadDutchDude @Turrican

    @k|ngp|n  @H2o vs. Ln2

     

    I have a question : who is the programmer responsible for 3 moths of having 16 card  result sitting in 4 card competition?
    If someone submits 16 core CPU result into 4 core CPU category the sub is instantly blocked after reporting. 
    Here we have strange situation: 16- and 10-card machines are sitting in 4 cards World Record category, destroying other scores with gap...
    and no one is doing anything about it.
    Strange... isn't it. 

    How do you find that score OK?
    https://hwbot.org/submission/3991289_x_computers.eu_gpupi___1b_4_geforce_rtx_2080_ti_0sec_639ms
    There are categories of  5, 6 , 7 ,and 8 GPUs - you can find it here:
    https://hwbot.org/benchmark/gpupi_-_1b/

    Or this one: 
    https://hwbot.org/submission/4159675_macdaffy_gpupi___32b_4_geforce_gtx_1080_ti_52sec_442ms
    Again You have up to 8 cards category
    https://hwbot.org/benchmark/gpupi_-_32b/
    And MacDaffy sits in 4 cards category while using 16 cards!!! 

    What is happening? Move those results to max cards categories - easy.
    Why don't you do it?


    Cheers


    Peter

×
×
  • Create New...