Geforce GTX 1050 for mining



  • Hi all,

    my pc is a low power pc with an ASRock J3455-ITX which consumes only about 8-10 watts. It has 8GB of RAM.
    I tried out different graphic cards in pc to mine with it. Currently I am mining with 50 TB and I got the following speed rates with different cards.

    Nvidia Geforce GT 730 = average read 250 MB/s.
    Nvidia Geforce GT 1030 = average read 300 MB/s
    Nvidia Geforce GTX 1050 = average read 310 MB/s. Time to read 50 TB is about 48 secs.

    Can someone tell me where the bottleneck is? Okay the GT 730 and GT1030 are not that good but I thought the GTX 1050 would be much faster.

    I also put all my harddisk with the same USB 3.0 Hubs to my mac book pro 2015 with a Intel Iris Pro and it runs like a shamp. With over 1000 MB/s and time read time is about 18 secs.

    Can someone give me an advice?

    Thanks a lot.



  • On my pc I am running Ubuntu Server 16.04 LTS with jminer 4.11



  • 48s for 50 TB on such a low end system is fairly good.

    The ASRock J3455-ITX has just 1 PCIe 2.0 x1 -> 500 MB/s maximum.
    I don't have the time to read up on that chipset, but I guess the USB3 controller is also on a single lane - so the overall throughput is reasonable.

    Connecting storage via USB (as opposed to PCIe->SATA) has a massive protocol overhead;

    -USB transmissions
    -USB-attached-SCSI protocol conversion
    -your target is "talking" SCSI, not AHCI, but still way slower than directly-connected-AHCI

    Looking at the system process table you will find a great amount of time spent in kernel/drivers/interrupts.

    @ciscler said in Geforce GTX 1050 for mining:

    I also put all my harddisk with the same USB 3.0 Hubs to my mac book pro 2015 with a Intel Iris Pro and it runs like a shamp. With over 1000 Mbit/s and time read time is about 18 secs.

    -?- So your Mac does what - "1000 Mbit/s" may translate to "125 Mbyte/s" -?-



  • sorry I mean the macbook is doing 700 - 1000 Mbytes/s. But I think you are right "PCIe 2.0 x1 -> 500 MB/s maximum" this is the problem.