I'm trying to benchmark nvBowtie but it is not working for me. It allocates memory in the GPU and then goes into a CPU loop with 0% GPU utilisation and no data are written.
Here is stdout:
info : nvBowtie... started
verbose : cuda devices : 1
verbose : device 0 has compute capability 5.2
verbose : SM count : 16
verbose : SM clock rate : 1177 Mhz
verbose : memory clock rate : 2.5 Ghz
verbose : chosen device 0
verbose : device name : GRID M60-8Q
verbose : compute capability : 5.2
visible : mapping reference index... started
info : file: "ref/hg38_no_alt.fa"
info : SequenceDataMMAP: error mapping file "/nvbio.ref/hg38_no_alt.fa.seq_info" (2)!
visible : mapping reference index... failed
visible : loading reference index... started
info : file: "ref/hg38_no_alt.fa"
visible : loading reference index... done
visible : FMIndexData: loading... started
visible : genome : ref/hg38_no_alt.fa
info : reading bwt... started
info : reading bwt... done
verbose : length: 3088286401
info : building occurrence table... started
info : building occurrence table... done
info : size: 386035808 words
info : reading rbwt... started
info : reading rbwt... done
verbose : length: 3088286401
info : building occurrence table... started
info : building occurrence table... done
visible : primary : 2655318798
visible : rprimary : 1993732142
info : reading SSA... started
info : reading SSA... done
info : reading SSA... started
info : reading SSA... done
visible : memory : 4417.8 MB
visible : FMIndexData: loading... done
visible : opening read file [1] ""
visible : opening read file [2] "/stornext/HPCScratch/PapenfussLab/mirror/ftp-trace.ncbi.nlm.nih.gov/giab/ftp/data/NA12878/Garvan_NA12878_HG001_HiSeq_Exome/NIST7035_TAAGGCGA_L001_R2_001_trimmed.fastq.gz"
verbose : mode = best
verbose : scoring = sw
verbose : score-min = log:0.00:10.00
verbose : alignment type = local
verbose : pe-policy = fr
verbose : seed length = 20
verbose : seed interval = (S, 1.000, 0.750)
verbose : seed rounds = 2
verbose : max hits = 100
verbose : max edit dist = 15
verbose : max effort = 15
verbose : substitutions = 0
verbose : mapq filter = 0
verbose : randomized = yes
visible : [0] nvBowtie cuda driver created on device 0
stats : [0] allocated device driver data (2.32 GB - 0.4s)
stats : [0] estimated allocation sizes: HOST 0 MB, DEVICE 3548 MB)
visible : [0] nvBowtie cuda driver... started
verbose : starting background paired-end input thread
stats : [0] device has 4584 of 8191 MB free
stats : [0] processing reads in batches of 1024K
stats : [0] allocating alignment buffers... started
stats : [0] estimated: HOST 0 MB, DEVICE 3548 MB)
verbose : [0] allocating 384 MB of string storage
verbose : [0] CIGARs : 128 MB
verbose : [0] MDs : 256 MB
verbose : [0] allocating 1182 MB of DP storage
stats : [0] allocating alignment buffers... done
stats : [0] allocated: HOST 0 MB, DEVICE 3548 MB)
stats : [0] ready to start processing: device has 1035 MB free
Here are the details of the card
Wed Jul 26 13:02:50 2017
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 367.43 Driver Version: 367.43 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GRID M60-8Q Off | 0000:02:01.0 Off | N/A |
| N/A N/A P8 N/A / N/A | 7155MiB / 8191MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 29140 C nvBowtie 6579MiB |
+-----------------------------------------------------------------------------+
I'm trying to benchmark nvBowtie but it is not working for me. It allocates memory in the GPU and then goes into a CPU loop with 0% GPU utilisation and no data are written.
Here is stdout:
Here are the details of the card