Probably the easiest way would be to set $/ = \16 so that you read the file in 16 bytes chunks.

Ie. $/ = \16; while( my $line = <INFILE> ) { willresult in $line containing 16 bytes each time, which will give you your 8 values per output line.

Simplistically, that make the program something like:

#! perl -lw
use strict;

## Note the -l above which makes print add newlines.

open IN, '<:raw:perlio', $ARGV[0] or die $!;
open OUT, '>', 'junk.out' or die $!;
$/ = \16; ## read 16 bytes at a time;

print join ',', map{ sprintf '0x%04x', $_ } unpack 'v*', $_ while <IN>
+;
close OUT;
close IN;
[download]

If there was (still) some way to binmode *ARGV, it could be reduced to a one-liner:

perl -nle"BEGIN{$/=\16}print join',',map{sprintf'0x%04x',$_}unpack'v*'
+,$_" binfile >outfile
[download]

but without binmode, that fails if the file contains a ^Z (control-Z; ascii 26) character.

Or you can go the other way and 'PBP-up' the above.

Well, the $ctr = 0 doesn't do anything useful, and the $ctr += 1 could be merged into the if conditional.

my @converted_list = map { sprintf('0x%04x', $_) } unpack('v*', $line)
+;
    
my $ctr;
foreach (@converted_list) {
   print OUTFILE "$_, ";
   print OUTFILE "\n" unless ++$ctr % 8;
}
[download]

You could merge of all that together if you so desired, but that's probably a bit odd.

my $ctr;
print map { ++$ctr % 8 ? "$_, " : "$_\n" }
      map { sprintf('0x%04x', $_) }
      unpack('v*', $line);
[download]

The simplest solution is probably to omit the counter altogether by only reading in 16 bytes at a time. As a bonus, it doesn't leave trailing comma when the file isn't an exact multiple of 16 bytes long.

# Each line has 8 16-bit words, so 16 bytes.
local $/ = \16;

while (<INFILE>) {
   print join ', ',
         map { sprintf('0x%04x', $_) }
         unpack('v*', $_);
   print("\n");
}
[download]