ATA guppi_59447_17004_6377981_frb180916_0001.0000.raw --> Segmentation fault

[texadactyl]

Tried to assign this to @RocketRoss but was not able to. guppi_59447_17004_6377981_frb180916_0001.0000.raw is a 4-bit file per @wfarah .

Stdout:

rawspec -f 41920 -t 2 -i 1.0 guppi_59447_17004_6377981_frb180916_0001
rawspec using librawspec 2.5.0+2@gb2026d5-dirty cuFFT 10.2.1.245
working stem: guppi_59447_17004_6377981_frb180916_0001
opening file: guppi_59447_17004_6377981_frb180916_0001.0000.raw
number of bits per sample must be 8 or 16 (not 4), using 8 bps
Using the single antenna-weight (1.000000) for all antennas in the incoherent-sum.
Segmentation fault

Note rawspec_gpu.cu, starting at line 462:

  if(ctx->Nbps == 0) {
    ctx->Nbps = 8;
  } else if(ctx->Nbps != 8 && ctx->Nbps != 16) {
    fprintf(stderr,
        "number of bits per sample must be 8 or 16 (not %d), using 8 bps\n",
        ctx->Nbps);
    fflush(stderr);
    NbpsIsExpanded = ctx->Nbps == 4;
    ctx->Nbps = 8;
  }

[radonnachie]

Ah right, I should blast that set of raw files from disk:

I believe that the issue there is that the rawfile has a PIPERBLK of 10480 which has prime factors of [2, 5, 131], the latter being exceptional and beyond scope (see conclusion to #20)

rawspec using librawspec 2.5.0+5@g7c0bf4f-dirty cuFFT 10.5.1.100
working stem: /mnt/datax-netStorage-40G/guppi_59447_17004_6377981_frb180916_0001
opening file: /mnt/datax-netStorage-40G/guppi_59447_17004_6377981_frb180916_0001.0000.raw
number of bits per sample must be 8 or 16 (not 4), using 8 bps
got error CUFFT_NOT_SUPPORTED at rawspec_gpu.cu:1089
rawspec initialization failed

I'll furnish you with 4bit data that has a power of 2 PIPERBLK.

[texadactyl]

Received and verified.