Comments (11)
syamajala
commented on August 16, 2024
I guess there are other build errors after fixing that:
[ 7%] Building CXX object runtime/CMakeFiles/RealmRuntime.dir/realm/deppart/partitions.cc.o
In file included from /lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.cu:24:
In file included from /lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion.h:59:
In file included from /lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.h:1836:
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.inl:141:14: error: unknown type name '__forceinline__'
__device__ __forceinline__
^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.inl:148:14: error: unknown type name '__forceinline__'
__device__ __forceinline__
^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.inl:159:14: error: unknown type name '__forceinline__'
__device__ __forceinline__
^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/legion/legion_redop.inl:160:10: error: expected ';' after top level declarator
uint8_t __uint2ubyte(unsigned int value, unsigned offset)
^
;
[ 29%] Building CXX object runtime/CMakeFiles/RealmRuntime.dir/realm/hip/hip_access.cc.o
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/hip/hip_internal.cc:1135:47: error: no member named 'hip_fold_excl_fn' in 'Realm::ReductionOpUntyped'
(redop_info.is_exclusive ? redop->hip_fold_excl_fn : redop->hip_fold_nonexcl_fn) :
~~~~~ ^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/hip/hip_internal.cc:1135:73: error: no member named 'hip_fold_nonexcl_fn' in 'Realm::ReductionOpUntyped'
(redop_info.is_exclusive ? redop->hip_fold_excl_fn : redop->hip_fold_nonexcl_fn) :
~~~~~ ^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/hip/hip_internal.cc:1136:47: error: no member named 'hip_apply_excl_fn' in 'Realm::ReductionOpUntyped'
(redop_info.is_exclusive ? redop->hip_apply_excl_fn : redop->hip_apply_nonexcl_fn));
~~~~~ ^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/hip/hip_internal.cc:1136:74: error: no member named 'hip_apply_nonexcl_fn' in 'Realm::ReductionOpUntyped'
(redop_info.is_exclusive ? redop->hip_apply_excl_fn : redop->hip_apply_nonexcl_fn));
~~~~~ ^
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/hip/hip_internal.cc:1500:20: error: no member named 'hip_apply_excl_fn' in 'Realm::ReductionOpUntyped'; did you mean 'cpu_apply_excl_fn'?
if(!redop->hip_apply_excl_fn)
^~~~~~~~~~~~~~~~~
cpu_apply_excl_fn
/lustre/orion/cmb138/scratch/seshuy/legion_s3d_flow_control/legion/runtime/realm/redop.h:62:14: note: 'cpu_apply_excl_fn' declared here
void (*cpu_apply_excl_fn)(void *lhs_ptr, size_t lhs_stride,
from legion.
elliottslaughter
commented on August 16, 2024
Yes, CI is now failing on master commit 423ddfb: https://code.olcf.ornl.gov/ci/csc335/dev/legion/-/pipelines/9890
It previously worked on master commit 13d4101: https://code.olcf.ornl.gov/ci/csc335/dev/legion/-/pipelines/9862
from legion.
elliottslaughter
commented on August 16, 2024
I'm not immediately seeing how any of this is related, but here are the commits in this range:
$ git log --graph --decorate --oneline 13d4101cccc75ae67e880068617608e90fb8c3e4..423ddfb3943c6e86cd269b045b67ba99e3463502
* 423ddfb39 (HEAD -> master, origin/master, origin/HEAD) Merge branch 'rpath-fixes' into 'master'
|\
| * 63eeeda1b (origin/rpath-fixes) Don't use lib/ blindly, ask cmake what to use
| * 110a83c35 Add libpython.so directory to BUILD_RPATH for librealm.so
| * 065d1cc05 Set RPATH on legion_python to @ORIGIN/../lib
| * 22b554eef Make sure we append to all RPATHs, instead of setting them
* 4024b4723 Merge branch 'test_args' into 'master'
* bcfd3e031 realm: update the machine_config unit test to fully test command line parser and machine config APIfrom legion.
eddy16112
commented on August 16, 2024
I created a fix for the hip module https://gitlab.com/StanfordLegion/legion/-/merge_requests/908, but I can not reproduce the redop error with HIP_TARGET=CUDA. the function is there https://gitlab.com/StanfordLegion/legion/-/blob/master/runtime/realm/redop.h#L94
from legion.
elliottslaughter
commented on August 16, 2024
Other users are hitting this, so I went ahead and merged https://gitlab.com/StanfordLegion/legion/-/merge_requests/908. I'm not sure what to tell you about the latest NUMA failure, but hopefully this unblocks @syamajala and others.
from legion.
elliottslaughter
commented on August 16, 2024
New CI for master here: https://code.olcf.ornl.gov/ci/csc335/dev/legion/-/pipelines/10240
from legion.
elliottslaughter
commented on August 16, 2024
Fixed the initial issue. I'll open a new bug for the NUMA one.
from legion.
elliottslaughter
commented on August 16, 2024
Follow-on issue is #1547.
from legion.
syamajala
commented on August 16, 2024
I made a local branch that merges master into control_replication and I am still not able to build on frontier. It doesnt look like the CI actually passed and failed in the prep step?
The same errors are still there.
from legion.
elliottslaughter
commented on August 16, 2024
I think the CI build errors are in the filesystem? It seems to be intermittent, this one passed (the build at least):
https://code.olcf.ornl.gov/ci/csc335/dev/legion/-/pipelines/10250
from legion.
elliottslaughter
commented on August 16, 2024
FYI, CI is passing now. The only thing that fails is #1547.
https://code.olcf.ornl.gov/ci/csc335/dev/legion/-/pipelines/10316
I'm pretty sure this is I/O instability.
from legion.
Related Issues (20)
- Crash in checkpointed Circuit HOT 7
- Regent: Performance regression in Terra 1.2.0 / LLVM 18 on NVIDIA GPUs
- Legion: uninitialized data error HOT 1
- [BUG] Legion Multinode Crash UBSAN Error HOT 1
- Realm: Crash in checkpointed SNAP HOT 11
- Fuzzer: incorrect application of reductions HOT 4
- Legion: collective instance freeze on slingshot-11 HOT 7
- Remove CUDART hijack HOT 12
- Unify most memory kinds HOT 1
- Realm GPU Profiling Is Not Precise HOT 10
- Optimize gather copies in Moya/FleCSI HOT 25
- [BUG] `PhysicalManager::get_use_event()` race condition HOT 8
- Fuzzer: assertion in multi-node DCR HOT 2
- Legion: profile mpi handshake HOT 5
- [UBSAN] `VirtualChannel::package_message()`: store to misaligned address HOT 2
- Legion built without zlib generates uncompressed logs without warning
- Profiler: Show "truly-in-use" memory usage line
- Performance degradation when using compact instances HOT 4
- Add online hang detection in Realm HOT 2
- Barrier Profiling
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from legion.