Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Non-uniform work-groups at 3 ranks fails for SYCL (acc on DPC++) #7

Open
tom91136 opened this issue May 2, 2024 · 0 comments
Open
Assignees

Comments

@tom91136
Copy link
Member

tom91136 commented May 2, 2024

+ srun --cpu-bind=map_cpu:49,57,17,25,1,9,33,41 -n 3 --cpu-bind none --mem-bind none sh -c 'export ROCR_VISIBLE_DEVICES=$((SLURM_PROCID % N_GPUS)) && echo "# ROCR_VISIBLE_DEVICES=$ROCR_VISIBLE_DEVICES@$(hostname):$SLURM_PROCID" && /users/tomlin/performance-portability/benchmarking/2023/cloverleaf/results/cloverleaf-mi250x_oneapi-2024.1_sycl-acc/CloverLeaf/build/sycl-acc-cloverleaf --file /users/tomlin/performance-portability/benchmarking/2023/cloverleaf/results/cloverleaf-mi250x_oneapi-2024.1_sycl-acc/CloverLeaf/InputDecks/clover_bm256_300.in --out ../out/clover_np3_mi250x_oneapi-2024.1_sycl-acc_256_300_stage_false.out --staging-buffer false '
# ROCR_VISIBLE_DEVICES=0@nid006622:0
# ROCR_VISIBLE_DEVICES=1@nid006622:1
# ROCR_VISIBLE_DEVICES=2@nid006623:2
---
Devices:
 0: gfx90a:sramecc+:xnack-
CloverLeaf:
 - Ver.:     2.000
 - Deck:     /users/tomlin/performance-portability/benchmarking/2023/cloverleaf/results/cloverleaf-mi250x_oneapi-2024.1_sycl-acc/CloverLeaf/InputDecks/clover_bm256_300.in
 - Out:      ../out/clover_np3_mi250x_oneapi-2024.1_sycl-acc_256_300_stage_false.out
 - Profiler: deck-specified
MPI:
 - Enabled:     true
 - Total ranks: 3
 - Header device-awareness (CUDA-awareness):  unknown
 - Runtime device-awareness (CUDA-awareness): unknown
 - Host-Device halo exchange staging buffer:  false
Model:
 - Name:      SYCL (accessors)
 - Execution: Offload (device)
 - Indexing: RANGE2D_LINEAR
 - SYCL device: gfx90a:sramecc+:xnack-
   - Type    : gpu
   - Version : gfx90a:sramecc+:xnack-
   - Vendor  : AMD Corporation
   - Driver  : HIP 50221.15
# ---- 
Output: |+1
 Output file clover.out opened. All output will go there.
 Args: --file /users/tomlin/performance-portability/benchmarking/2023/cloverleaf/results/cloverleaf-mi250x_oneapi-2024.1_sycl-acc/CloverLeaf/InputDecks/clover_bm256_300.in --out ../out/clover_np3_mi250x_oneapi-2024.1_sycl-acc_256_300_stage_false.out --staging-buffer false
 Using input: `/users/tomlin/performance-portability/benchmarking/2023/cloverleaf/results/cloverleaf-mi250x_oneapi-2024.1_sycl-acc/CloverLeaf/InputDecks/clover_bm256_300.in`
 Problem initialised and generated
 Launching hydro
[SYCL] Exception : `Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)`
terminate called after throwing an instance of 'sycl::_V1::exception'
  what():  Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)
[SYCL] Exception : `Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)`
[SYCL] Exception : `Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)`
terminate called after throwing an instance of 'sycl::_V1::exception'
  what():  Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)
terminate called after throwing an instance of 'sycl::_V1::exception'
  what():  Non-uniform work-groups are not supported by the target device -54 (PI_ERROR_INVALID_WORK_GROUP_SIZE)
srun: error: nid006622: tasks 0-1: Aborted
srun: launch/slurm: _step_signal: Terminating StepId=6994308.5
srun: error: nid006623: task 2: Aborted (core dumped)
@tom91136 tom91136 self-assigned this May 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant