kmc_tools always in in1: 0%

[zhangrengang]

Hi, I have two nodes in a cluster. In one node, running kmc_tools will be completed in a few minutes. But in the other node, running kmc_tools will be always in in1: 0% status and never be completed. The %CPU is 100%. I have tested versions 3.1.1 and the lastest 3.2.1. Any way to fix it? My command is kmc_tools transform k_31 sort k_31.sorted

[zhangrengang]

Similarly, kmc_tools transform k_31 dump -s k_31.dump does not work but kmc_tools transform k_31 dump k_31.dump works.

[marekkokot]

This is not good :( Could you please provide your input data and how you counted k-mers (i.e. exact kmc command line). It might also be that something went wrong during k-mer counting (I hope not), so could you also provide your computed kmc database?. If the sizes are large, maybe you would be able to find a smaller set causing this behavior. I would really want to fix this issue, but it may be hard without reproducing. Also, how many cores does your node have?

[zhangrengang]

My command is kmc -m58 -t60 -k31 -ci4 -cs1000000 -fq @test.list test_31 test_31 > test_31.stats && kmc_tools transform test_31 sort test_31.sorted.

The files are too large for GitHub. Could you provide me a email address to send the files?

When I make a subset of the input data, I find that: 1) when dataset is small (head 100000-2000000 lines), it works; 2) when head 4000000 or 4500000, the status is stoped at in1: 1%; 3) when head >= 5000000 lines, the status is stoped at in1: 0%.

The database generated in the normal node also does not work in the abnormal node.

The CPU information of the abnormal node:

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              192
On-line CPU(s) list: 0-191
Thread(s) per core:  2
Core(s) per socket:  48
Socket(s):           2
NUMA node(s):        2
Vendor ID:           AuthenticAMD
CPU family:          23
Model:               49
Model name:          AMD EPYC 7642 48-Core Processor
Stepping:            0
CPU MHz:             2392.417
CPU max MHz:         2300.0000
CPU min MHz:         1500.0000
BogoMIPS:            4591.02
Virtualization:      AMD-V
L1d cache:           32K
L1i cache:           32K
L2 cache:            512K
L3 cache:            16384K
NUMA node0 CPU(s):   0-47,96-143
NUMA node1 CPU(s):   48-95,144-191
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate sme ssbd mba sev ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr wbnoinvd arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif umip rdpid overflow_recov succor smca

[marekkokot]

Hi,

I configured my e-mail to be displayed on github, so that you can send me your files. Could you also try to run kmc_tools with a small number of threads? Maybe the issue is related to the many threads used (of course it needs to be fixed anyway, but if it works for a smaller number of threads I will have some ideas about the causes).

[zhangrengang]

Hi, I have sent the files to your email. I have test kmc with -t1 and -t4, but it does not work.

[zhangrengang]

I find the size of dataset also has an effect by changing the option -ci: -ci5: in1: 1% -ci6: in1: 3% -ci7: in1: 4% -ci8: in1: 6% -ci9: in1: 100% (works)

[marekkokot]

Hi. Could you also test kmc_tools with lower number of threads? kmc_tools -t16 transform test_31 sort test_31.sorted I am able to reproduce this bug when I set -t190 (like the default on your node I guess). I will try to fix this as soon as possible, but I am currently quite busy - nevertheless, I will try to prioritize this.

[zhangrengang]

Yes, it works! Thank you. It is resolved for me.

[Shiye2023]

Is this problem solved?