abort: packed_vector.hpp:465: void dyn::packed_vector::remove(uint64_t): Assertion `(size_ / int_per_word_ == words.size() #24
Comments
|
First invocation: and in gdb: This is very nicely reproducible. |
|
I wonder if the recent merge of stuff from my repo caused a regression.
…On Sat, Jul 18, 2020 at 7:29 PM Steffen Möller ***@***.***> wrote:
First invocation:
$ ./obj-x86_64-linux-gnu/wm_string
SPEED
access:252ms
rank:331ms
wm_string: /home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp:200: ulint dyn::wm_string<dynamic_bitvector_t>::select(ulint, ulint) const [with dynamic_bitvector_t = dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256, 16> >; ulint = long unsigned int]: Assertion `rank > 0' failed.
Aborted (core dumped)
and in gdb:
Core was generated by `./obj-x86_64-linux-gnu/wm_string'.
Program terminated with signal SIGABRT, Aborted.
#0 __GI_raise ***@***.***=6) at ../sysdeps/unix/sysv/linux/raise.c:50
50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) bt
#0 __GI_raise ***@***.***=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#1 0x00007f08b36ef55b in __GI_abort () at abort.c:79
#2 0x00007f08b36ef42f in __assert_fail_base (fmt=0x7f08b3855b48 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n", assertion=0x56266529ce16 "rank > 0",
file=0x56266529bb40 "/home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp", line=200, function=<optimized out>) at assert.c:92
#3 0x00007f08b36fe092 in __GI___assert_fail (assertion=0x56266529ce16 "rank > 0", file=0x56266529bb40 "/home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp", line=200,
function=0x56266529c628 "ulint dyn::wm_string<dynamic_bitvector_t>::select(ulint, ulint) const [with dynamic_bitvector_t = dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256, 16> >; ulint = long unsigned int]")
at assert.c:101
#4 0x0000562665294d2f in dyn::wm_string<dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256u, 16u> > >::select (this=<optimized out>, rank=0, c=<optimized out>) at /usr/include/c++/9/bits/stl_vector.h:1058
#5 0x000056266528c474 in speed_select (num=100000, num_of_alphabet=<optimized out>) at ./wm_string.cpp:74
#6 0x0000562665290ada in speed_test (num=100000, num_of_alphabet=10000000) at /usr/include/c++/9/bits/char_traits.h:335
#7 0x000056266528aa2e in main () at ./wm_string.cpp:394
This is very nicely reproducible.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#24 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABDQEMRB4UO6XYOJ4AIWDDR4HLW5ANCNFSM4PAEUNXA>
.
|
|
Actually, it was from vgteam, correct?
On Sat, Jul 18, 2020 at 7:47 PM Erik Garrison <[email protected]>
wrote:
… I wonder if the recent merge of stuff from my repo caused a regression.
On Sat, Jul 18, 2020 at 7:29 PM Steffen Möller ***@***.***>
wrote:
> First invocation:
>
> $ ./obj-x86_64-linux-gnu/wm_string
> SPEED
> access:252ms
> rank:331ms
> wm_string: /home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp:200: ulint dyn::wm_string<dynamic_bitvector_t>::select(ulint, ulint) const [with dynamic_bitvector_t = dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256, 16> >; ulint = long unsigned int]: Assertion `rank > 0' failed.
> Aborted (core dumped)
>
> and in gdb:
>
> Core was generated by `./obj-x86_64-linux-gnu/wm_string'.
> Program terminated with signal SIGABRT, Aborted.
> #0 __GI_raise ***@***.***=6) at ../sysdeps/unix/sysv/linux/raise.c:50
> 50 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
> (gdb) bt
> #0 __GI_raise ***@***.***=6) at ../sysdeps/unix/sysv/linux/raise.c:50
> #1 0x00007f08b36ef55b in __GI_abort () at abort.c:79
> #2 0x00007f08b36ef42f in __assert_fail_base (fmt=0x7f08b3855b48 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n", assertion=0x56266529ce16 "rank > 0",
> file=0x56266529bb40 "/home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp", line=200, function=<optimized out>) at assert.c:92
> #3 0x00007f08b36fe092 in __GI___assert_fail (assertion=0x56266529ce16 "rank > 0", file=0x56266529bb40 "/home/moeller/git/med-team/dynamic/xxsds-dynamic-1.0~alpha.1/include/dynamic/internal/wm_string.hpp", line=200,
> function=0x56266529c628 "ulint dyn::wm_string<dynamic_bitvector_t>::select(ulint, ulint) const [with dynamic_bitvector_t = dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256, 16> >; ulint = long unsigned int]")
> at assert.c:101
> #4 0x0000562665294d2f in dyn::wm_string<dyn::succinct_bitvector<dyn::spsi<dyn::packed_bit_vector, 256u, 16u> > >::select (this=<optimized out>, rank=0, c=<optimized out>) at /usr/include/c++/9/bits/stl_vector.h:1058
> #5 0x000056266528c474 in speed_select (num=100000, num_of_alphabet=<optimized out>) at ./wm_string.cpp:74
> #6 0x0000562665290ada in speed_test (num=100000, num_of_alphabet=10000000) at /usr/include/c++/9/bits/char_traits.h:335
> #7 0x000056266528aa2e in main () at ./wm_string.cpp:394
>
> This is very nicely reproducible.
>
> —
> You are receiving this because you are subscribed to this thread.
> Reply to this email directly, view it on GitHub
> <#24 (comment)>, or
> unsubscribe
> <https://github.com/notifications/unsubscribe-auth/AABDQEMRB4UO6XYOJ4AIWDDR4HLW5ANCNFSM4PAEUNXA>
> .
>
|
|
That is what I presume since I requested a pull from vgteam/DYNAMIC. |
I thought I had solved this bug but apparently it's alive and well. I'll look into it next week. |
nicolaprezza
added a commit
that referenced
this issue
Jul 22, 2020
|
Hi, I solved the first part of the issue: Assertion `(size_ / int_per_word_ == words.size() || !(words[size_ / int_per_word_] >> ((size_ % int_per_word_) * width_))) && "uninitialized non-zero values in the end of the vector"' failed. Actually, the code was correct and only the assert had a bug : -) @ekg : the second issue happens in wm_string.cpp. The failed assert happens because the code calls select(rank, c) at wm_string.hpp:200 with rank == 0. Could it be simply a bug of the checker wm_string.cpp? select is not supposed to be called with rank == 0. |
|
That does look like a bug in the check in wm_string. I'll take a look.
…On Wed, Jul 22, 2020 at 12:44 PM Nicola Prezza ***@***.***> wrote:
Hi, I solved the first part of the issue:
Assertion `(size_ / int_per_word_ == words.size() || !(words[size_ /
int_per_word_] >> ((size_ % int_per_word_) * width_))) && "uninitialized
non-zero values in the end of the vector"' failed.
Actually, the code was correct and only the assert had a bug : -)
@ekg <https://github.com/ekg> : the second issue happens in
wm_string.cpp. The failed assert happens because the code calls
select(rank, c) at wm_string.hpp:200 with rank == 0. Could it be simply a
bug of the checker wm_string.cpp? select is not supposed to be called with
rank == 0.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#24 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABDQEJZJAEW4QBBYLQRSZLR427KNANCNFSM4PAEUNXA>
.
|
|
There these automated build tests spreading throughout github that are executed for every commit. Is this something that should be introduced to DYNAMIC, too? |
|
Yes, it should be done here too.
…On Wed, Jul 22, 2020, 13:09 Steffen Möller ***@***.***> wrote:
There these automated build tests spreading throughout github that are
executed for every commit. Is this something that should be introduced to
DYNAMIC, too?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#24 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABDQEOXOKEIIXHGAX7L23LR43CGNANCNFSM4PAEUNXA>
.
|
|
All dependencies of DYNAMIC are in Debian/Ubuntu already - I guess this would be rather straight-forward to implement. Ping me if you need help. |
|
@nicolaprezza I'm not able to reproduce the problem with wm_string. It's at least not happened in the 20-30 runs that I've done (running it in a loop and watching...) |
|
Latest checkout: Maybe because of other compiler settings? The benchmark seems stable now. Would you have other ideas what I should test over here?
I can get a VM online somewhere for a common playground to investigate the wm_string issue. |
|
I can reproduce it. Have you enabled asserts? with cmake, it's simply cmake .. -DCMAKE_BUILD_TYPE=Debug |
|
Thanks :) That's my mistake. I can reproduce it by enabling the assert! |
|
It does look like a possible problem with the input. There is a quirky PRNG being used. I might switch to modern c++ random distributions. |
|
Thanks @ekg! now it works. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello,
ran into
And peeking a boo in gdb:
but found it to reproduce only every 2nd to 5th invocation. Spotted with version 1.0-alpha.1.
Anything you want me to do about it?
Cheers,
Steffen
The text was updated successfully, but these errors were encountered: