Issue 395 -- Filenames are too long #424

akhilpotla · 2019-05-15T02:54:22Z

Realized there are some errors in this, I have final exams this week. I will have this corrected over the weekend. Apologize for the premature PR.

On Ubuntu 16.04 I get a failure when cloning this repository since some the file names are too long.
The problematic files are:

KokkosSparse_gauss_seidel_apply_eti_spec_inst_Kokkos_complex_double__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_complex_double__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_complex_double__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_complex_float__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_complex_float__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_complex_double__int_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_complex_double__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_complex_double__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_complex_float__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_complex_float__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp

The longest of these files has a name of 147 characters. Linux typically restricts filename length to 255 characters, but if your hard drive is encrypted (which is a lot of Linux users) this goes down to 143.

I abbreviated a few of the words in the filenames:
double --> dbl
float --> f
complex --> cmplx

After my changes the new names are:

KokkosSparse_gauss_seidel_apply_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_cmplx_f__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_numeric_eti_spec_inst_Kokkos_cmplx_f__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_dbl__int_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
kkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_dbl__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_f__size_t_int64_t_LayoutLeft_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp
KokkosSparse_gauss_seidel_symbolic_eti_spec_inst_Kokkos_cmplx_f__size_t_int64_t_LayoutRight_Cuda_CudaHostPinnedSpace_CudaHostPinnedSpace.cpp

akhilpotla · 2019-05-15T17:41:03Z

@srajama1
I was way more excited about this than I was for my final exam, so I ended up working on this instead.

I made the changes that I needed too.

I don't know how to run the unit tests, the README indicates the don't seem to be working, so do I need to run them locally?

srajama1

There are some odd changes here. Please see below.

Thanks for contributing but finals are important :)

scripts/generate_specialization_function.bash

scripts/generate_specialization.bash

...ons_cpp/abs/KokkosBlas1_abs_eti_spec_inst_Kokkos_cmplx_dbl__LayoutLeft_Cuda_CudaUVMSpace.cpp

crtrott · 2019-05-15T20:27:52Z

How about abreviating the Space names instead?
Like CudaSpace -> CudaS
CudaHostPinnedSpace -> CudaHPS
CudaUVMSpace -> CudaUS
in the file names.

srajama1 · 2019-05-15T20:31:18Z

Yeah, when we are doing this we could also abbreviate the spaces.

akhilpotla · 2019-05-15T20:51:40Z

Some of the comments you have made are outdated. I think I might have pushed my change when you were in middle of your review. I seem to have resolved most of them.

I will work on abbreviating the spaces.
Should I return the scalars to their original names? dbl --> double?

I will update my work this weekend, after my finals 😄.

srajama1 · 2019-05-15T21:55:24Z

I am ok with having dbl and flt, in addition to the space names @crtrott suggested. Yes, I found some of the changes unexplainable. You might have have cleaned it up already.

akhilpotla · 2019-05-22T05:23:21Z

I have made an update, I think this is ready for another review. I have abbreviated the space names in the filenames as @crtrott suggested, and I have changed f -> flt as you said @srajama1.

srajama1

Thanks for fixing these, please see below for two main questions

scripts/generate_specialization.bash

...cpp/abs/KokkosBlas1_abs_eti_spec_inst_Kokkos_complex_double__LayoutLeft_OpenMP_HostSpace.cpp

...cpp/abs/KokkosBlas1_abs_eti_spec_inst_Kokkos_complex_double__LayoutLeft_Serial_HostSpace.cpp

...ric_eti_spec_inst_Kokkos_complex_float__size_t_int_LayoutRight_Threads_HBWSpace_HBWSpace.cpp

akhilpotla · 2019-06-01T23:54:18Z

@srajama1

I fixed the issues in the review that you pointed out. The issue was that I forgot to add a conditional to see whether the MemSpace needed to be abbreviated or not; therefore, if it didn't need to be abbreviated the MemSpace would not be printed in the filename. I added a conditional to allow the unabbreviated MemSpace names to be printed, so git now knows those files were only renamed.

In addition, I had to delete the original files that were too long on Github's website because they couldn't be cloned on my computer, so each one of those deletions forced me to make a commit. I tried rebasing and squashing locally, but that wasn't possible because the names of the files were too long causing my computer to error out.

Also I would like to thank you for your quick responses and also your patience. This is my first open source contribution, so I really appreciate your guidance. Also if there are any issues you think I could tackle please assign them to me. I would be happy to dedicate part of my time to working on those!

akhilpotla · 2019-06-02T21:06:37Z

Also I am curious how have y'all not experience this issue? I am assumed you guys use linux machines. Do you use a custom OS that allow longer filename lengths for encrypted hard drives?

akhilpotla · 2019-08-09T16:44:01Z

@srajama1

Could you please take a look?

srajama1 · 2019-08-17T03:12:12Z

@akhilpotla : Our spotchecking is broken because of an inadvertent change. We will get this in as soon as we fix the spotchecks and test this one.

ndellingwood · 2019-09-04T16:45:45Z

I'll start a spot-check
@akhilpotla - thanks for working on this, can you rebase to squash the commits down to a single commit for this PR (will require a force push to update the PR)? I can provide some instructions if that would help, or I can try and submit a PR to your branch. This touches many files, if for some reason we have to back out the PR it would be much easier off a single commit.

ndellingwood · 2019-09-04T16:47:00Z

I changed the base of the PR to submit to develop instead of master.

srajama1 · 2019-09-11T15:14:00Z

@ndellingwood : Spot check works ? Also, can you or @akhilpotla squash it to one commit ?

ndellingwood · 2019-09-17T16:29:41Z

@srajama1 here's spot-check results, sorry for delay in posting

bash-4.2$ ../../scripts/test_all_sandia --spot-check
Running on machine: white
Going to test compilers:  gcc/6.4.0 gcc/7.2.0 ibm/16.1.0 cuda/9.2.88
Testing compiler gcc/6.4.0
Testing compiler gcc/7.2.0
  Starting job gcc-7.2.0-OpenMP-release
  Starting job gcc-6.4.0-OpenMP_Serial-release
  PASSED gcc-7.2.0-OpenMP-release
  Starting job gcc-7.2.0-Serial-release
  PASSED gcc-6.4.0-OpenMP_Serial-release
  PASSED gcc-7.2.0-Serial-release
Testing compiler ibm/16.1.0
Testing compiler cuda/9.2.88
  Starting job gcc-7.2.0-OpenMP_Serial-release
  Starting job ibm-16.1.0-Serial-release
  PASSED gcc-7.2.0-OpenMP_Serial-release
  PASSED ibm-16.1.0-Serial-release
  Starting job cuda-9.2.88-Cuda_OpenMP-release
  PASSED cuda-9.2.88-Cuda_OpenMP-release
#######################################################
PASSED TESTS
#######################################################
cuda-9.2.88-Cuda_OpenMP-release build_time=884 run_time=631
gcc-6.4.0-OpenMP_Serial-release build_time=426 run_time=964
gcc-7.2.0-OpenMP-release build_time=288 run_time=273
gcc-7.2.0-OpenMP_Serial-release build_time=515 run_time=943
gcc-7.2.0-Serial-release build_time=206 run_time=622
ibm-16.1.0-Serial-release build_time=978 run_time=849
#######################################################
FAILED TESTS
#######################################################

srajama1 · 2019-09-17T17:12:36Z

@ndellingwood : can you squash the commit too ? I am assuming kokk-dev and bowman are coming soon. I will commit as soon as you give green light.

akhilpotla · 2019-10-03T03:57:25Z

@srajama1, sorry for the slow response. I was moving.

I squashed the commits. Now there is one commit titled "Abbreviated spacenames and types in filenames".

I see there are some failing tests. Were those tests expected to break?

srajama1 · 2019-10-23T15:26:15Z

@akhilpotla I don't think any tests are failing. @ndellingwood : Is this ready to be pushed ?

ndellingwood · 2019-10-24T20:17:56Z

@srajama1 Yes, this is ready. We shouldn't mark the issue as InDevelop since newer capabilities are coming into develop for 3.0 that will need the same file renaming treatment (sptrsv, spiluk)

Follow-on to PR #424 for issue #395

ndellingwood · 2019-11-18T22:11:04Z

Ready to get this merged in. The conflicts with the CMake system updates are listed in PR #491 and I'll clean those up when updating with the develop branch. sptrsv and spiluk were added after this PR, so I'll put in a PR immediately after with updates to rename those files. I tested this branch with a merge of develop and the sptrsv + spiluk file renaming and all looks good.

Running on machine: kokkos-dev-2
Going to test compilers:  gcc/7.3.0 gcc/9.1 intel/18.0.5 clang/8.0 cuda/10.1
Testing compiler gcc/7.3.0
Testing compiler gcc/9.1
  Starting job gcc-7.3.0-OpenMP-release
  Starting job gcc-7.3.0-Pthread-release
  PASSED gcc-7.3.0-OpenMP-release
  Starting job gcc-9.1-OpenMP-release
  PASSED gcc-7.3.0-Pthread-release
Testing compiler intel/18.0.5
  Starting job gcc-9.1-Serial-release
  PASSED gcc-9.1-OpenMP-release
Testing compiler clang/8.0
  Starting job intel-18.0.5-OpenMP-release
  PASSED gcc-9.1-Serial-release
  Starting job clang-8.0-Cuda_OpenMP-release
  PASSED intel-18.0.5-OpenMP-release
Testing compiler cuda/10.1
  Starting job clang-8.0-Pthread_Serial-release
  PASSED clang-8.0-Cuda_OpenMP-release
  PASSED clang-8.0-Pthread_Serial-release
  Starting job cuda-10.1-Cuda_OpenMP-release
  PASSED cuda-10.1-Cuda_OpenMP-release
#######################################################
PASSED TESTS
#######################################################
clang-8.0-Cuda_OpenMP-release build_time=524 run_time=209
clang-8.0-Pthread_Serial-release build_time=273 run_time=296
cuda-10.1-Cuda_OpenMP-release build_time=585 run_time=183
gcc-7.3.0-OpenMP-release build_time=204 run_time=94
gcc-7.3.0-Pthread-release build_time=189 run_time=139
gcc-9.1-OpenMP-release build_time=224 run_time=85
gcc-9.1-Serial-release build_time=200 run_time=182
intel-18.0.5-OpenMP-release build_time=329 run_time=108

ndellingwood · 2019-11-18T22:12:41Z

As an additional follow-on: the associative arrays used in the generate_* bash files are only available for bash >= 4, however it seems by defaults macs have bash 3.2, so this should be updated.

ndellingwood · 2019-11-18T22:16:27Z

Thanks @akhilpotla for the PR!

akhilpotla changed the title ~~Issue 395 -- Filenames are too long~~ [WIP] -- Issue 395 -- Filenames are too long May 15, 2019

akhilpotla changed the title ~~[WIP] -- Issue 395 -- Filenames are too long~~ Issue 395 -- Filenames are too long May 15, 2019

srajama1 reviewed May 15, 2019

View reviewed changes

scripts/generate_specialization_function.bash Outdated Show resolved Hide resolved

scripts/generate_specialization.bash Outdated Show resolved Hide resolved

...ons_cpp/abs/KokkosBlas1_abs_eti_spec_inst_Kokkos_cmplx_dbl__LayoutLeft_Cuda_CudaUVMSpace.cpp Outdated Show resolved Hide resolved

srajama1 reviewed May 23, 2019

View reviewed changes

ndellingwood changed the base branch from master to develop September 4, 2019 16:46

Abbreviated spacenames and types in filenames

319839a

akhilpotla force-pushed the generated_specializations_too_long branch from d1fa75f to 319839a Compare October 3, 2019 03:54

akhilpotla requested a review from srajama1 October 25, 2019 16:59

ndellingwood mentioned this pull request Nov 18, 2019

Overhaul of CMake build system to provide both TriBITS and standalone systems #491

Merged

ndellingwood added a commit that referenced this pull request Nov 18, 2019

Rename sptrsv and spiluk

0082094

Follow-on to PR #424 for issue #395

ndellingwood self-requested a review November 18, 2019 22:12

ndellingwood approved these changes Nov 18, 2019

View reviewed changes

ndellingwood merged commit dd48175 into kokkos:develop Nov 18, 2019

ndellingwood mentioned this pull request Nov 18, 2019

Pr 424 update #497

Merged

bartlettroscoe mentioned this pull request Nov 19, 2019

Generated kokkos-kernels file names are too long and are crashing cloning Trilinos on Windows #395

Closed

ndellingwood mentioned this pull request Mar 24, 2020

Build error on MacOS ( ld: symbol(s) not found for architecture x86_64) trilinos/Trilinos#7035

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Issue 395 -- Filenames are too long #424

Issue 395 -- Filenames are too long #424

akhilpotla commented May 15, 2019 •

edited

Loading

akhilpotla commented May 15, 2019 •

edited

Loading

srajama1 left a comment

crtrott commented May 15, 2019

srajama1 commented May 15, 2019

akhilpotla commented May 15, 2019 •

edited

Loading

srajama1 commented May 15, 2019

akhilpotla commented May 22, 2019

srajama1 left a comment

akhilpotla commented Jun 1, 2019 •

edited

Loading

akhilpotla commented Jun 2, 2019

akhilpotla commented Aug 9, 2019

srajama1 commented Aug 17, 2019

ndellingwood commented Sep 4, 2019

ndellingwood commented Sep 4, 2019

srajama1 commented Sep 11, 2019

ndellingwood commented Sep 17, 2019

srajama1 commented Sep 17, 2019 •

edited

Loading

akhilpotla commented Oct 3, 2019 •

edited

Loading

srajama1 commented Oct 23, 2019

ndellingwood commented Oct 24, 2019

ndellingwood commented Nov 18, 2019

ndellingwood commented Nov 18, 2019

ndellingwood commented Nov 18, 2019

Issue 395 -- Filenames are too long #424

Issue 395 -- Filenames are too long #424

Conversation

akhilpotla commented May 15, 2019 • edited Loading

akhilpotla commented May 15, 2019 • edited Loading

srajama1 left a comment

Choose a reason for hiding this comment

crtrott commented May 15, 2019

srajama1 commented May 15, 2019

akhilpotla commented May 15, 2019 • edited Loading

srajama1 commented May 15, 2019

akhilpotla commented May 22, 2019

srajama1 left a comment

Choose a reason for hiding this comment

akhilpotla commented Jun 1, 2019 • edited Loading

akhilpotla commented Jun 2, 2019

akhilpotla commented Aug 9, 2019

srajama1 commented Aug 17, 2019

ndellingwood commented Sep 4, 2019

ndellingwood commented Sep 4, 2019

srajama1 commented Sep 11, 2019

ndellingwood commented Sep 17, 2019

srajama1 commented Sep 17, 2019 • edited Loading

akhilpotla commented Oct 3, 2019 • edited Loading

srajama1 commented Oct 23, 2019

ndellingwood commented Oct 24, 2019

ndellingwood commented Nov 18, 2019

ndellingwood commented Nov 18, 2019

ndellingwood commented Nov 18, 2019

akhilpotla commented May 15, 2019 •

edited

Loading

akhilpotla commented May 15, 2019 •

edited

Loading

akhilpotla commented May 15, 2019 •

edited

Loading

akhilpotla commented Jun 1, 2019 •

edited

Loading

srajama1 commented Sep 17, 2019 •

edited

Loading

akhilpotla commented Oct 3, 2019 •

edited

Loading