Add `gdb` pretty-printers for rmm types #1088

upsj · 2022-08-17T16:18:48Z

Description

This PR adds a pretty-printer for device_uvector and pulls in Thrust pretty-printers that were added in NVIDIA/thrust#1631. CMake provides a convenience script to load all of the pretty-printers, to resolve the duplication concerns raised in rapidsai/cudf#11499.

Example output:

Details

$ gdb -q gtests/DEVICE_UVECTOR_TEST
Reading symbols from gtests/DEVICE_UVECTOR_TEST...
(gdb) b cudaMalloc
Function "cudaMalloc" not defined.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 1 (cudaMalloc) pending.
(gdb) run
Starting program: /home/nfs/tribizel/rapids/rmm/build/cuda-11.5.0/feature__pretty-printers/debug/gtests/DEVICE_UVECTOR_TEST 
warning: Error disabling address space randomization: Operation not permitted
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/usr/lib/x86_64-linux-gnu/libthread_db.so.1".
Running main() from gmock_main.cc
[==========] Running 95 tests from 5 test suites.
[----------] Global test environment set-up.
[----------] 19 tests from TypedUVectorTest/0, where TypeParam = signed char
[ RUN      ] TypedUVectorTest/0.MemoryResource
[New Thread 0x7f741efc1000 (LWP 86147)]

Thread 1 "DEVICE_UVECTOR_" hit Breakpoint 1, 0x00007f7426dddc80 in cudaMalloc () from /home/nfs/tribizel/rapids/compose/etc/conda/cuda_11.5/envs/rapids/lib/libcudart.so.11.0
(gdb) c
Continuing.
[New Thread 0x7f741e7c0000 (LWP 86148)]
[       OK ] TypedUVectorTest/0.MemoryResource (999 ms)
[ RUN      ] TypedUVectorTest/0.ZeroSizeConstructor
[       OK ] TypedUVectorTest/0.ZeroSizeConstructor (0 ms)
[ RUN      ] TypedUVectorTest/0.NonZeroSizeConstructor

Thread 1 "DEVICE_UVECTOR_" hit Breakpoint 1, 0x00007f7426dddc80 in cudaMalloc () from /home/nfs/tribizel/rapids/compose/etc/conda/cuda_11.5/envs/rapids/lib/libcudart.so.11.0
(gdb) finish
Run till exit from #0  0x00007f7426dddc80 in cudaMalloc () from /home/nfs/tribizel/rapids/compose/etc/conda/cuda_11.5/envs/rapids/lib/libcudart.so.11.0
rmm::mr::cuda_memory_resource::do_allocate (bytes=<optimized out>, this=<optimized out>) at /home/nfs/tribizel/rapids/rmm/include/rmm/mr/device/cuda_memory_resource.hpp:70
70          RMM_CUDA_TRY_ALLOC(cudaMalloc(&ptr, bytes));
(gdb) finish
Run till exit from #0  rmm::mr::cuda_memory_resource::do_allocate (bytes=<optimized out>, this=<optimized out>) at /home/nfs/tribizel/rapids/rmm/include/rmm/mr/device/cuda_memory_resource.hpp:70
0x00005587874feef0 in rmm::device_buffer::allocate_async (bytes=12345, this=0x7ffefebb46b0) at /home/nfs/tribizel/rapids/rmm/include/rmm/device_buffer.hpp:418
418         _data     = (bytes > 0) ? memory_resource()->allocate(bytes, stream()) : nullptr;
Value returned is $1 = (void *) 0x7f73ff000000
(gdb) finish
Run till exit from #0  0x00005587874feef0 in rmm::device_buffer::allocate_async (bytes=12345, this=0x7ffefebb46b0) at /home/nfs/tribizel/rapids/rmm/include/rmm/device_buffer.hpp:418
TypedUVectorTest_NonZeroSizeConstructor_Test<signed char>::TestBody (this=<optimized out>) at /home/nfs/tribizel/rapids/rmm/tests/device_uvector_tests.cpp:55
55        EXPECT_EQ(vec.size(), size);
(gdb) print vec
$2 = {_storage = {_data = 0x7f73ff000000, _size = 12345, _capacity = 12345, _stream = {stream_ = 0x0}, _mr = 0x558787561008 <rmm::mr::detail::initial_resource()::mr>}}
(gdb) source load-pretty-printers 
(gdb) print vec
$3 = rmm::device_uvector<signed char> of length 12345, capacity 12345 = {0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 
  0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000', 0 '\000'...}
(gdb)

Checklist

I am familiar with the Contributing Guidelines.
New or existing tests cover these changes.
The documentation is up to date with these changes.

bdice

Some thoughts on Python clarity -- I haven't used GDB pretty printers before so forgive me if I'm not aware of some limitations. Does GDB run the Python interpreter from the user environment, or a recent version like 3.8+?

scripts/gdb-pretty-printers.py

bdice · 2022-08-17T16:40:46Z

scripts/gdb-pretty-printers.py

+            if char == '>':
+                depth -= 1
+            if depth == 0:
+                return False


Trying to understand the logic here -- if you only care if the < brackets are balanced, you can just count those. What's the "not alias" part mean?

I guess an example is easiest: std::vector<std::optional<std::pair<int, std::optional<double>>>> should pass, but std::vector<std::optional<std::pair<int, std::optional<double>>>>::some_templated_type_alias<int> should fail. gdb sometimes doesn't resolve type aliases, so the pretty-printer might erroneously be called for such a some_templated_type_alias<int>, and would fail at accessing the expected members.

upsj · 2022-08-17T17:09:06Z

gdb runs in the user environment, so I wouldn't bet on recent Python being available unfortunately

harrism · 2022-08-17T22:34:16Z

scripts/load-pretty-printers.in

@@ -0,0 +1,2 @@
+source @Thrust_SOURCE_DIR@/scripts/gdb-pretty-printers.py


What is a .in file?

Not sure if this is a general convention, but in the projects I worked in, we use it for template files that will be expanded into an actual file by CMake - config headers, pkg-config files etc.

The CMake convention is <file>.suffix.in for the input to configure_file. ( .cpp.in => .cpp )

bdice

Needs a handful of Python improvements. The general approach is much clearer now, with load-pretty-printers.in and the like.

scripts/gdb-pretty-printers.py

bdice

Thanks @upsj!

bdice · 2022-08-26T22:49:12Z

@gpucibot merge

This applies rmm's configured pre-commit hooks to format the gdb pretty printer script that was recently added (#1088). I would like to move rmm to use pre-commit for all style checks (as was recently done for cudf) and this is a prerequisite step. Authors: - Bradley Dice (https://github.com/bdice) Approvers: - GALI PREM SAGAR (https://github.com/galipremsagar) - Mark Harris (https://github.com/harrism) URL: #1127

upsj added 3 commits August 16, 2022 14:23

add pretty-printers

46ef3dc

wip

15bedd1

pull in thrust pretty-printer sources

533774e

upsj requested a review from a team as a code owner August 17, 2022 16:18

github-actions bot added the CMake label Aug 17, 2022

bdice reviewed Aug 17, 2022

View reviewed changes

upsj added 3 commits August 17, 2022 18:20

provide script to load pretty-printers

69eb5d2

review updates

7c839a3

modernize python code

4171b2f

upsj requested a review from bdice August 17, 2022 18:41

consistent quotes

4e73504

harrism added feature request New feature or request non-breaking Non-breaking change labels Aug 17, 2022

harrism reviewed Aug 17, 2022

View reviewed changes

harrism approved these changes Aug 22, 2022

View reviewed changes

bdice reviewed Aug 23, 2022

View reviewed changes

upsj added 2 commits August 23, 2022 10:46

review updates

ccfb112

remove unnecessary error message

433fda0

bdice approved these changes Aug 23, 2022

View reviewed changes

upsj mentioned this pull request Aug 26, 2022

Add gdb pretty-printers for simple types rapidsai/cudf#11499

Merged

3 tasks

rapids-bot bot merged commit ed43650 into rapidsai:branch-22.10 Aug 26, 2022

bdice mentioned this pull request Oct 7, 2022

Format gdb script. #1127

Merged

3 tasks

		@@ -0,0 +1,2 @@
		source @Thrust_SOURCE_DIR@/scripts/gdb-pretty-printers.py

Add gdb pretty-printers for rmm types #1088

Add gdb pretty-printers for rmm types #1088

Uh oh!

Conversation

upsj commented Aug 17, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Checklist

Uh oh!

bdice left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bdice Aug 17, 2022

Choose a reason for hiding this comment

Uh oh!

upsj Aug 17, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

upsj commented Aug 17, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

harrism Aug 17, 2022

Choose a reason for hiding this comment

Uh oh!

upsj Aug 18, 2022

Choose a reason for hiding this comment

Uh oh!

robertmaynard Aug 19, 2022

Choose a reason for hiding this comment

Uh oh!

bdice left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

bdice left a comment

Choose a reason for hiding this comment

Uh oh!

bdice commented Aug 26, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Add `gdb` pretty-printers for rmm types #1088

Add `gdb` pretty-printers for rmm types #1088

upsj commented Aug 17, 2022 •

edited

Loading

upsj Aug 17, 2022 •

edited

Loading

upsj commented Aug 17, 2022 •

edited

Loading