intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-26 23:33:20 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add clEnqueueNDRangeKernelINTEL API

Browse Source 
Related-To: NEO-2712

Change-Id: If1d16d9d626871a9dc4b19282f9edc5786ffa398
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
This commit is contained in:
Filip Hazubski
2019-11-19 16:54:47 +01:00
committed by sys_ocldev
parent 7be937c226
commit 82bc594af0
26 changed files with 764 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
runtime/command_queue/enqueue_common.h
View File

@@ -6,6 +6,7 @@
*/
#pragma once
#include "core/program/sync_buffer_handler.h"
#include "core/utilities/range.h"
#include "runtime/built_ins/built_ins.h"
#include "runtime/built_ins/builtins_dispatch_builder.h"
@@ -642,6 +643,15 @@ CompletionStamp CommandQueueHw<GfxFamily>::enqueueNonBlocked(
blocking = true;
printfHandler->makeResident(getGpgpuCommandStreamReceiver());
}
if (multiDispatchInfo.peekMainKernel()->usesSyncBuffer()) {
auto &gws = multiDispatchInfo.begin()->getGWS();
auto &lws = multiDispatchInfo.begin()->getLocalWorkgroupSize();
size_t workGroupsCount = (gws.x * gws.y * gws.z) /
(lws.x * lws.y * lws.z);
device->syncBufferHandler->prepareForEnqueue(workGroupsCount, *multiDispatchInfo.peekMainKernel(), getGpgpuCommandStreamReceiver());
}
if (timestampPacketContainer) {
timestampPacketContainer->makeResident(getGpgpuCommandStreamReceiver());
timestampPacketDependencies.previousEnqueueNodes.makeResident(getGpgpuCommandStreamReceiver());