Skip to content
This repository was archived by the owner on Mar 28, 2023. It is now read-only.

[SYCL] [L0] [PI] Ensure desired batching behavior independent of defaults. #943

Merged
merged 1 commit into from
Mar 23, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion SYCL/Plugin/interop-level-zero.cpp
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
// REQUIRES: level_zero, level_zero_dev_kit
// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %level_zero_options %s -o %t.out
// RUN: env SYCL_BE=PI_LEVEL_ZERO %GPU_RUN_PLACEHOLDER %t.out
// RUN: env SYCL_BE=PI_LEVEL_ZERO SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 %GPU_RUN_PLACEHOLDER %t.out
// UNSUPPORTED: ze_debug-1,ze_debug4

// Test for Level Zero interop API
Expand Down
2 changes: 1 addition & 1 deletion SYCL/Plugin/level_zero_batch_barrier.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
// REQUIRES: gpu, level_zero, TEMPORARILY_DISABLED

// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %level_zero_options %s -o %t.out
// RUN: env SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 %GPU_CHECK_PLACEHOLDER
// RUN: env SYCL_PI_TRACE=2 ZE_DEBUG=1 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 %GPU_RUN_PLACEHOLDER %t.out 2>&1 %GPU_CHECK_PLACEHOLDER

// Test that the wait with a barrier is fully batched, i.e. it doesn't cause
// extra submissions.
Expand Down
2 changes: 1 addition & 1 deletion SYCL/Plugin/level_zero_batch_event_status.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@
// RUN: %clangxx -fsycl -fsycl-unnamed-lambda -fsycl-targets=%sycl_triple %s -o %t.out

// Set batching to 4 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck %s

// level_zero_batch_test.cpp
//
Expand Down
14 changes: 7 additions & 7 deletions SYCL/Plugin/level_zero_batch_test.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -4,25 +4,25 @@
// UNSUPPORTED: ze_debug-1,ze_debug4

// Set batching to 4 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB4 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB4 %s

// Set batching to 1 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=1 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB1 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=1 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB1 %s

// Set batching to 3 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=3 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB3 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=3 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB3 %s

// Set batching to 5 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=5 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB5 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=5 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB5 %s

// Set batching to 7 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=7 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB7 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=7 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB7 %s

// Set batching to 8 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=8 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB8 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=8 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB8 %s

// Set batching to 9 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=9 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB9 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=9 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB9 %s

// level_zero_batch_test.cpp
//
Expand Down
14 changes: 7 additions & 7 deletions SYCL/Plugin/level_zero_batch_test_copy_with_compute.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -3,25 +3,25 @@
// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out

// Set batching to 4 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB4 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=4 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB4 %s

// Set batching to 1 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=1 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB1 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=1 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB1 %s

// Set batching to 3 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=3 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB3 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=3 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB3 %s

// Set batching to 5 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=5 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB5 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=5 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB5 %s

// Set batching to 7 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=7 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB7 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=7 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB7 %s

// Set batching to 8 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=8 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB8 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=8 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB8 %s

// Set batching to 9 explicitly
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=9 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB9 %s
// RUN: env SYCL_PI_LEVEL_ZERO_BATCH_SIZE=9 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKB9 %s

// level_zero_batch_test_copy_with_compute.cpp
//
Expand Down
2 changes: 1 addition & 1 deletion SYCL/Plugin/level_zero_dynamic_batch_test.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out

// Check that dynamic batching raises/lowers batch size
// RUN: env SYCL_PI_TRACE=2 ZE_DEBUG=1 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKDYN %s
// RUN: env SYCL_PI_TRACE=2 ZE_DEBUG=1 SYCL_PI_LEVEL_ZERO_DEVICE_SCOPE_EVENTS=2 SYCL_PI_LEVEL_ZERO_USE_IMMEDIATE_COMMANDLISTS=0 %GPU_RUN_PLACEHOLDER %t.out 2>&1 | FileCheck --check-prefixes=CKALL,CKDYN %s

// level_zero_dynamic_batch_test.cpp
//
Expand Down