February 13, 20187 yr I am glad you are entertained Allen...just make me fart so badly Edited February 13, 20187 yr by H2o vs. Ln2
February 13, 20187 yr For me it's interesting to see that Vince's run is faster for all finished batches after the first one with an advantage of up to 28 ms after batch 9 finished. From "Batch 19 finished" to "PI value output" however his system needs 114 ms, while Slinky is noticably faster with 98 ms. At first you might think that the higher-clocked 7980X with faster RAM outperformed the lower clocked 7960X with slower RAM in this phase of the benchmark in some way, but watching at the statistics segment it seems that some small hickup with Titan V #2 lead to the loss of the WR. Slinky's values are way more even in comparison. Anyway, congratulations!
February 13, 20187 yr Yes it's just an efficiency thing. Nothing to see here folks, Vince rerun in 3,2,1...
February 13, 20187 yr ..., but watching at the statistics segment it seems that some small hickup with Titan V #2 lead to the loss of the WR. Slinky's values are way more even in comparison. Anyway, congratulations!Actually card #1 and #3 are the problem, card #2 was the most efficient. You have to look at the percentage of batches calculated, #2 did 32% of the whole calculation in about the same time as #1 did only 21%. On the software part, there was only a minor upgrade of the CUDA toolkit between 3.0 and 3.1, so there shouldn't be much difference. I think this is a good case for hardware efficiency through better stability.  Well done, H2o! Try GPUPI 3.2 as well, it should have a little less overhead, because it doesn't use a physical log file on a disk anymore. That could scrape off another ms maybe.
February 15, 20187 yr same old ek-r5e, r4e or r4be I can't confirm sorry, have to be mod as only one size will fit the screw
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.