From: Pancho.Jones@proton.me   
      
   On 21/02/2024 20:51, druck wrote:   
   > On 21/02/2024 17:42, Pancho wrote:   
   >> I got rid of my desire for an nvme with my opi5. Due to being PCIe 1   
   >> lane, it wasn't that much faster,   
   >   
   > That surprises me, could anyone with an NVMe base run this script which   
   > measures the maximum sequential and random access read/write speeds.   
   > You'll need to install the fio package and be in a directory on the disc   
   > under test.   
   >   
      
   I posted some stats back in December ( hdparm -Ttv), on the opi5 the   
   nvme was only giving 186 MB/s seq read, not that much more than a USB3   
   SSD at 137 MB/s.   
      
   I think, like the oPi5, the rPi5 also has the handicap that it only   
   allocates PCIe 3.0 x 1 to nvme.   
      
   So yes, nvme is faster in benchmarks, but nowhere near the 3000MB/s you   
   might see on a PC.   
      
   Perhaps more pertinently, the user experience improvement using the Pi   
   as an HTPC was negligible. I guess the performance I care about wasn't   
   constrained by disk read speed. So now I just run a SD card for the OS,   
   and pretty much everything else is NAS anyway.   
      
   I'm actually really happy with the rPi5 as a HTPC. I'm sitting in front   
   of the TV typing this.   
      
   > #!/bin/bash   
   > TESTFILE=fio-tempfile.dat   
   > fio --name SeqRead --eta-newline=5s --filename=$TESTFILE --rw=read\   
   >   --size=500m --io_size=10g --blocksize=1024k --ioengine=libaio\   
   >   --fsync=10000 --iodepth=32 --direct=1 --numjobs=1 --runtime=30\   
   >   --group_reporting | grep bw=   
   > fio --name SeqWrite --eta-newline=5s --filename=$TESTFILE --rw=write\   
   >   --size=500m --io_size=10g --blocksize=1024k --ioengine=libaio\   
   >   --fsync=10000 --iodepth=32 --direct=1 --numjobs=1 --runtime=30\   
   >   --group_reporting | grep bw=   
   > fio --name RndRead --eta-newline=5s --filename=$TESTFILE --rw=randread\   
   >   --size=500m --io_size=10g --blocksize=4k --ioengine=libaio --fsync=1\   
   >   --iodepth=1 --direct=1 --numjobs=1 --runtime=30 --group_reporting\   
   >   | grep bw=   
   > fio --name RndWrite --eta-newline=5s --filename=$TESTFILE--rw=randwrite\   
   >   --size=500m --io_size=10g --blocksize=4k --ioengine=libaio --fsync=1\   
   >   --iodepth=1 --direct=1 --numjobs=1 --runtime=30 --group_reporting\   
   >   | grep bw=   
   > rm $TESTFILE   
   >   
   >> and there was significant additional effort installing the OS,   
   >> compared to a micro SD. Obviously, you may have a different use-case.   
   >   
   > More than dd'ing the SD card image on to the NVMe, using gparted to   
   > expand the linux partition, and then changing the boot device in   
   > /boot/cmdline.txt and /etc/fstab ?   
   >   
      
   On the oPi5 there was DD (which I hate), and editing files on the boot   
   partition, on the onboard persistent ram. Not difficult, but still   
   enough to make mistakes. Whereas the Raspberry Pi Imager makes writing a   
   SD card really easy.   
      
   I've got at least one spare nvme sitting around, I'm not going to put it   
   on the rPi5.   
      
   My oPi5 is also booting off micro SD even though it has an nvme drive.   
   If Armbian ever rollout a stable version for the oPi5 I will stick it on   
   the nvme again, but I did it so many times I got pissed off.   
      
   In stark contrast, Pi OS has been solid on the rPi5.   
      
      
      
   > ---druck   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)