Build failure on Linux ARM64

[martin-g]

Hello,

I've just tried to build h3-arm branch (against easel's develop branch) on openEuler OS 20.03 ARM64:

$ gcc --version
gcc (GCC) 7.3.0
...
$ autoconf
$ ./configure
...

HMMER configuration:
   compiler:             gcc -g -O2 -O3    -mlittle-endian -pthread 
   host:                 aarch64-unknown-linux-gnu
   linker:               
   libraries:              -lpthread
   DP implementation:    neon

$ make
...
$ make check
...
exercise    6 [      generic_viterbi] ...     FAILED [command failed]
...
1 of 307 exercises at level <= 2 FAILED.

System information:
ср апр 20 13:06:21 UTC 2022
Linux mgrigorov-arm64 4.19.90-2203.5.0.0143.oe1.aarch64 #1 SMP Wed Mar 30 17:26:54 UTC 2022 aarch64 aarch64 aarch64 GNU/Linux
make[1]: *** [Makefile:41: check] Error 1
make: *** [Makefile:118: check] Error 2

Please let me know if you need any specifics about my system and any test output files!

[npcarter]

Hello,

Could you run the command src/generic_viterbi_utest --vv in your hmmer directory and send us the results, please? I'm at a bit of a loss as to what could be going on here, as the generic_viterbi module doesn't contain any ARM-specific code, but hopefully running the utest with verbose turned on will tell us something useful.

-Nick

On Wed, Apr 20, 2022 at 9:16 AM Martin Grigorov @.***> wrote:

Hello,

I've just tried to build h3-arm branch (against easel's develop branch) on openEuler OS 20.03 ARM64:

...

exercise 6 [ generic_viterbi] ... FAILED [command failed]

...

1 of 307 exercises at level <= 2 FAILED.

System information:

ср апр 20 13:06:21 UTC 2022

Linux mgrigorov-arm64 4.19.90-2203.5.0.0143.oe1.aarch64 #1 SMP Wed Mar 30 17:26:54 UTC 2022 aarch64 aarch64 aarch64 GNU/Linux

make[1]: *** [Makefile:41: check] Error 1

make: *** [Makefile:118: check] Error 2

Please let me know if you need any specifics about my system and any test output files!

— Reply to this email directly, view it on GitHub https://github.com/EddyRivasLab/hmmer/issues/276, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABDJBZB7JXJGV5AVQWL23STVF77THANCNFSM5T4AMPPA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[martin-g]

The test does not run:

mgrigorov@mgrigorov-arm64 ~/g/h/h/hmmer (h3-arm)> ./src/generic_viterbi_utest --vv
trace score and Viterbi score don't agree.

Should I do something more before trying to execute it ?

[npcarter]

Just a progress update: tested this benchmark on our M1 Mac, and couldn't reproduce the problem. I'm looking for access to an ARM Unix machine to test on.

-Nick

On Wed, Apr 20, 2022 at 1:53 PM Martin Grigorov @.***> wrote:

@.*** ~/g/h/h/hmmer (h3-arm)> ./src/generic_viterbi_utest --vv trace score and Viterbi score don't agree.

— Reply to this email directly, view it on GitHub https://github.com/EddyRivasLab/hmmer/issues/276#issuecomment-1104245160, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABDJBZHWZDMDERLFN3TSTJLVGBABPANCNFSM5T4AMPPA . You are receiving this because you commented.Message ID: @.***>

[martin-g]

Hi Nick,

Thanks for looking into this issue!

I can provide you with a Linux ARM64 VM! Please contact me at martin.grigorov at Gmail com and send me your preferred spec (flavor, CPU, RAM, disk space,...)

On Mon, Apr 25, 2022, 22:03 Nick Carter @.***> wrote:

Just a progress update: tested this benchmark on our M1 Mac, and couldn't reproduce the problem. I'm looking for access to an ARM Unix machine to test on.

-Nick

On Wed, Apr 20, 2022 at 1:53 PM Martin Grigorov @.***> wrote:

@.*** ~/g/h/h/hmmer (h3-arm)> ./src/generic_viterbi_utest --vv trace score and Viterbi score don't agree.

— Reply to this email directly, view it on GitHub < https://github.com/EddyRivasLab/hmmer/issues/276#issuecomment-1104245160>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ABDJBZHWZDMDERLFN3TSTJLVGBABPANCNFSM5T4AMPPA

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/EddyRivasLab/hmmer/issues/276#issuecomment-1108931554, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABYUQSKK4HCFU3KNSFHX3LVG3UBTANCNFSM5T4AMPPA . You are receiving this because you authored the thread.Message ID: @.***>

[martin-g]

Hi,

I've started from scratch (clone, autoconf, ...) and today I am not able to reproduce the issue with the failing test! I executed make check many times and all tests passed successfully always.

Closing... I'll let you know if I face any issues!