> -----Original Message----- > From: Peng Tao [mailto:bergwolf@xxxxxxxxx] > Sent: Friday, April 19, 2013 05:28 > To: Yan Burman > Cc: J. Bruce Fields; Tom Tucker; linux-rdma@xxxxxxxxxxxxxxx; linux- > nfs@xxxxxxxxxxxxxxx > Subject: Re: NFS over RDMA benchmark > > On Wed, Apr 17, 2013 at 10:36 PM, Yan Burman <yanb@xxxxxxxxxxxx> > wrote: > > Hi. > > > > I've been trying to do some benchmarks for NFS over RDMA and I seem to > only get about half of the bandwidth that the HW can give me. > > My setup consists of 2 servers each with 16 cores, 32Gb of memory, and > Mellanox ConnectX3 QDR card over PCI-e gen3. > > These servers are connected to a QDR IB switch. The backing storage on the > server is tmpfs mounted with noatime. > > I am running kernel 3.5.7. > > > > When running ib_send_bw, I get 4.3-4.5 GB/sec for block sizes 4-512K. > > When I run fio over rdma mounted nfs, I get 260-2200MB/sec for the same > block sizes (4-512K). running over IPoIB-CM, I get 200-980MB/sec. > > I got to these results after the following optimizations: > > 1. Setting IRQ affinity to the CPUs that are part of the NUMA node the > > card is on 2. Increasing > > /proc/sys/sunrpc/svc_rdma/max_outbound_read_requests and > > /proc/sys/sunrpc/svc_rdma/max_requests to 256 on server 3. Increasing > > RPCNFSDCOUNT to 32 on server > Did you try to affine nfsd to corresponding CPUs where your IB card locates? > Given that you see a bottleneck on CPU (as in your later email), it might be > worth trying. I tried to affine nfsd to CPUs on the NUMA node the IB card is on. I also set tmpfs memory policy to allocate from the same NUMA node. I did not see big difference. > > > 4. FIO arguments: --rw=randread --bs=4k --numjobs=2 --iodepth=128 > > --ioengine=libaio --size=100000k --prioclass=1 --prio=0 --cpumask=255 > > --loops=25 --direct=1 --invalidate=1 --fsync_on_close=1 --randrepeat=1 > > --norandommap --group_reporting --exitall --buffered=0 > > > On client side, it may be good to affine FIO processes and nfsiod to CPUs > where IB card locates as well, in case client is the bottleneck. > I am doing that - cpumask=255 affines it to the NUMA node my card is on. For some reason doing taskset on nfsiod fails. > -- > Thanks, > Tao ��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
