Announcement

Collapse
No announcement yet.

Many Interesting Tests Are Coming For The Phoronix 10th Birthday

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Many Interesting Tests Are Coming For The Phoronix 10th Birthday

    Phoronix: Many Interesting Tests Are Coming For The Phoronix 10th Birthday

    To end out May on a high note and in getting June kick-started, when Phoronix turns ten years old on 5 June, there is a lot of interesting upcoming content and Linux/open-source performance benchmarks...

    http://www.phoronix.com/vr.php?view=MTcwMDk

  • #2
    CudaMiner and OpenCL miner benchmarks

    Comment


    • #3
      Originally posted by riklaunim View Post
      CudaMiner and OpenCL miner benchmarks
      Test profile additions welcome... As far as I know, those apps don't have an automated benchmark mode.
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #4
        Intel Bay Trail NUC vs Intel Haswell NUC vs AMD equivalent

        Comment


        • #5
          Originally posted by mendieta View Post
          Intel Bay Trail NUC vs Intel Haswell NUC vs AMD equivalent
          Bay Trail NUC is all I have with no Haswell NUC and no AMD equivalent, at least for that form factor.
          Michael Larabel
          http://www.michaellarabel.com/

          Comment


          • #6
            Originally posted by Michael View Post
            Test profile additions welcome... As far as I know, those apps don't have an automated benchmark mode.
            CudaMiner has --benchmark but I must check if it can be stopped after X calculations.

            Comment


            • #7
              Originally posted by riklaunim View Post
              CudaMiner has --benchmark but I must check if it can be stopped after X calculations.
              Well if you can check and verify that it will return gracefully after X calculations, that would be good, and could lead to it being added for testing.
              Michael Larabel
              http://www.michaellarabel.com/

              Comment


              • #8
                It's doable. It can be timeouted:

                Code:
                ./cudaminer  --benchmark --time-limit 20
                	   *** CudaMiner for nVidia GPUs by Christian Buchner ***
                	             This is version 2014-02-28 (beta)
                	based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
                	    Cuda additions Copyright 2013,2014 Christian Buchner
                	  LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm
                	  BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM
                	  YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4
                [2014-05-27 01:57:56] NVML GPU monitoring is not available.
                [2014-05-27 01:57:56] 1 miner threads started, using 'scrypt' algorithm.
                [2014-05-27 01:57:56] GPU #0: GeForce GT 740M with compute capability 3.5
                [2014-05-27 01:57:56] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
                [2014-05-27 01:57:56] GPU #0: 32 hashes / 4.0 MB per warp.
                [2014-05-27 01:57:56] GPU #0: Performing auto-tuning (Patience...)
                [2014-05-27 01:57:56] GPU #0: maximum total warps (BxW): 483
                [2014-05-27 01:58:11] GPU #0: 49466.21 hash/s with configuration T12x4
                [2014-05-27 01:58:11] GPU #0: using launch configuration T12x4
                [2014-05-27 01:58:17] GPU #0: GeForce GT 740M, 47.72 khash/s
                [2014-05-27 01:58:17] Total: 47.72 khash/s
                [2014-05-27 01:58:22] GPU #0: GeForce GT 740M, 47.42 khash/s
                [2014-05-27 01:58:22] Total: 47.42 khash/s
                [2014-05-27 01:58:27] GPU #0: GeForce GT 740M, 47.43 khash/s
                [2014-05-27 01:58:27] Total: 47.43 khash/s
                [2014-05-27 01:58:32] GPU #0: GeForce GT 740M, 47.44 khash/s
                [2014-05-27 01:58:32] Total: 47.44 khash/s
                [2014-05-27 01:58:37] GPU #0: GeForce GT 740M, 47.40 khash/s
                [2014-05-27 01:58:37] Total: 47.40 khash/s
                [2014-05-27 01:58:42] GPU #0: GeForce GT 740M, 47.42 khash/s
                [2014-05-27 01:58:42] Total: 47.42 khash/s
                [2014-05-27 01:58:43] workio thread dead, waiting for workers...
                [2014-05-27 01:58:43] worker threads all shut down, exiting.
                compute capability and the khash/s are the most important. Older Quadros on compute capability 1.1 got only ~27 khash/s on FX 3700 and around 7 on FX 580

                Going this rout there may be some CUDA kernels that could be adapter as a CUDA benchmarks. Someone familiar with them would have to help I guess.

                Comment


                • #9
                  More system benchmarks: real-world performance, a kernel 3.10 (at least) to 3.15 comparison, and what happened to the file system/block layer regressions found in past releases...

                  Comment


                  • #10
                    Originally posted by halo9en View Post
                    More system benchmarks: real-world performance, a kernel 3.10 (at least) to 3.15 comparison, and what happened to the file system/block layer regressions found in past releases...
                    Been thinking about benchmarking all Linux 3.x kernels...
                    Michael Larabel
                    http://www.michaellarabel.com/

                    Comment

                    Working...
                    X