intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-05 09:09:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

Refactor TimestampPacket class

Browse Source 
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
This commit is contained in:
Bartosz Dunajski
2021-03-22 15:14:49 +00:00
committed by Compute-Runtime-Automation
parent 689028992a
commit 3dc3ad36f8
11 changed files with 159 additions and 128 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
shared/source/command_stream/command_stream_receiver.h
View File

@@ -47,7 +47,7 @@ struct HwPerfCounter;
struct HwTimeStamps;
template <typename TSize>
struct TimestampPackets;
class TimestampPackets;
template <typename T1>
class TagAllocator;

8
shared/source/command_stream/command_stream_receiver_hw_base.inl
View File

@@ -1006,8 +1006,8 @@ uint32_t CommandStreamReceiverHw<GfxFamily>::blitBuffer(const BlitPropertiesCont
TimestampPacketHelper::programCsrDependencies<GfxFamily>(commandStream, blitProperties.csrDependencies, getOsContext().getNumSupportedDevices());
if (blitProperties.outputTimestampPacket && profilingEnabled) {
auto timestampContextStartGpuAddress = blitProperties.outputTimestampPacket->getGpuAddress() + offsetof(TimestampPacketStorage, packets[0].contextStart);
auto timestampGlobalStartAddress = blitProperties.outputTimestampPacket->getGpuAddress() + offsetof(TimestampPacketStorage, packets[0].globalStart);
auto timestampContextStartGpuAddress = TimestampPacketHelper::getContextStartGpuAddress(*blitProperties.outputTimestampPacket);
auto timestampGlobalStartAddress = TimestampPacketHelper::getGlobalStartGpuAddress(*blitProperties.outputTimestampPacket);
EncodeStoreMMIO<GfxFamily>::encode(commandStream, GP_THREAD_TIME_REG_ADDRESS_OFFSET_LOW, timestampContextStartGpuAddress);
EncodeStoreMMIO<GfxFamily>::encode(commandStream, REG_GLOBAL_TIMESTAMP_LDW, timestampGlobalStartAddress);
@@ -1017,8 +1017,8 @@ uint32_t CommandStreamReceiverHw<GfxFamily>::blitBuffer(const BlitPropertiesCont
if (blitProperties.outputTimestampPacket) {
if (profilingEnabled) {
auto timestampContextEndGpuAddress = blitProperties.outputTimestampPacket->getGpuAddress() + offsetof(TimestampPacketStorage, packets[0].contextEnd);
auto timestampGlobalEndAddress = blitProperties.outputTimestampPacket->getGpuAddress() + offsetof(TimestampPacketStorage, packets[0].globalEnd);
auto timestampContextEndGpuAddress = TimestampPacketHelper::getContextEndGpuAddress(*blitProperties.outputTimestampPacket);
auto timestampGlobalEndAddress = TimestampPacketHelper::getGlobalEndGpuAddress(*blitProperties.outputTimestampPacket);
EncodeMiFlushDW<GfxFamily>::programMiFlushDw(commandStream, 0llu, newTaskCount, false, false);

2
shared/source/helpers/blit_commands_helper.h
View File

@@ -29,7 +29,7 @@ template <typename TagType>
struct TagNode;
template <typename TSize>
struct TimestampPackets;
class TimestampPackets;
struct BlitProperties;
struct HardwareInfo;

42
shared/source/helpers/timestamp_packet.h
View File

@@ -1,5 +1,5 @@
/*
* Copyright (C) 2018-2020 Intel Corporation
* Copyright (C) 2018-2021 Intel Corporation
*
* SPDX-License-Identifier: MIT
*
@@ -12,6 +12,7 @@
#include "shared/source/helpers/aux_translation.h"
#include "shared/source/helpers/hw_helper.h"
#include "shared/source/helpers/non_copyable_or_moveable.h"
#include "shared/source/helpers/string.h"
#include "shared/source/utilities/tag_allocator.h"
#include "pipe_control_args.h"
@@ -30,7 +31,8 @@ constexpr uint32_t preferredPacketCount = 16u;
#pragma pack(1)
template <typename TSize>
struct TimestampPackets {
class TimestampPackets {
public:
struct Packet {
TSize contextStart = 1u;
TSize globalStart = 1u;
@@ -67,8 +69,27 @@ struct TimestampPackets {
implicitGpuDependenciesCount = 0;
}
void assignDataToAllTimestamps(uint32_t packetIndex, void *source) {
memcpy_s(&packets[packetIndex], sizeof(Packet), source, sizeof(Packet));
}
size_t getGlobalStartOffset() const { return ptrDiff(&packets[0].globalStart, this); }
size_t getContextStartOffset() const { return ptrDiff(&packets[0].contextStart, this); }
size_t getContextEndOffset() const { return ptrDiff(&packets[0].contextEnd, this); }
size_t getGlobalEndOffset() const { return ptrDiff(&packets[0].globalEnd, this); }
size_t getImplicitGpuDependenciesCountOffset() const { return ptrDiff(&implicitGpuDependenciesCount, this); }
uint64_t getContextStartValue(uint32_t packetIndex) const { return static_cast<uint64_t>(packets[packetIndex].contextStart); }
uint64_t getGlobalStartValue(uint32_t packetIndex) const { return static_cast<uint64_t>(packets[packetIndex].globalStart); }
uint64_t getContextEndValue(uint32_t packetIndex) const { return static_cast<uint64_t>(packets[packetIndex].contextEnd); }
uint64_t getGlobalEndValue(uint32_t packetIndex) const { return static_cast<uint64_t>(packets[packetIndex].globalEnd); }
void setPacketsUsed(uint32_t used) { packetsUsed = used; }
uint32_t getPacketsUsed() const { return packetsUsed; }
uint32_t getImplicitGpuDependenciesCount() const { return implicitGpuDependenciesCount; }
protected:
Packet packets[TimestampPacketSizeControl::preferredPacketCount];
uint32_t implicitGpuDependenciesCount = 0;
uint32_t packetsUsed = 1;
@@ -110,11 +131,20 @@ struct TimestampPacketDependencies : public NonCopyableClass {
struct TimestampPacketHelper {
static uint64_t getContextEndGpuAddress(const TagNode<TimestampPacketStorage> &timestampPacketNode) {
return timestampPacketNode.getGpuAddress() + offsetof(TimestampPacketStorage, packets[0].contextEnd);
return timestampPacketNode.getGpuAddress() + timestampPacketNode.tagForCpuAccess->getContextEndOffset();
}
static uint64_t getContextStartGpuAddress(const TagNode<TimestampPacketStorage> &timestampPacketNode) {
return timestampPacketNode.getGpuAddress() + timestampPacketNode.tagForCpuAccess->getContextStartOffset();
}
static uint64_t getGlobalEndGpuAddress(const TagNode<TimestampPacketStorage> &timestampPacketNode) {
return timestampPacketNode.getGpuAddress() + timestampPacketNode.tagForCpuAccess->getGlobalEndOffset();
}
static uint64_t getGlobalStartGpuAddress(const TagNode<TimestampPacketStorage> &timestampPacketNode) {
return timestampPacketNode.getGpuAddress() + timestampPacketNode.tagForCpuAccess->getGlobalStartOffset();
}
static uint64_t getGpuDependenciesCountGpuAddress(const TagNode<TimestampPacketStorage> &timestampPacketNode) {
return timestampPacketNode.getGpuAddress() + offsetof(TimestampPacketStorage, implicitGpuDependenciesCount);
return timestampPacketNode.getGpuAddress() + timestampPacketNode.tagForCpuAccess->getImplicitGpuDependenciesCountOffset();
}
static void overrideSupportedDevicesCount(uint32_t &numSupportedDevices);
@@ -128,7 +158,7 @@ struct TimestampPacketHelper {
auto compareAddress = getContextEndGpuAddress(timestampPacketNode);
auto dependenciesCountAddress = getGpuDependenciesCountGpuAddress(timestampPacketNode);
for (uint32_t packetId = 0; packetId < timestampPacketNode.tagForCpuAccess->packetsUsed; packetId++) {
for (uint32_t packetId = 0; packetId < timestampPacketNode.tagForCpuAccess->getPacketsUsed(); packetId++) {
uint64_t compareOffset = packetId * sizeof(TimestampPacketStorage::Packet);
EncodeSempahore<GfxFamily>::addMiSemaphoreWaitCommand(cmdStream, compareAddress + compareOffset, 1, COMPARE_OPERATION::COMPARE_OPERATION_SAD_NOT_EQUAL_SDD);
}
@@ -202,7 +232,7 @@ struct TimestampPacketHelper {
template <typename GfxFamily>
static size_t getRequiredCmdStreamSizeForNodeDependency(TagNode<TimestampPacketStorage> &timestampPacketNode) {
size_t totalMiSemaphoreWaitSize = timestampPacketNode.tagForCpuAccess->packetsUsed * sizeof(typename GfxFamily::MI_SEMAPHORE_WAIT);
size_t totalMiSemaphoreWaitSize = timestampPacketNode.tagForCpuAccess->getPacketsUsed() * sizeof(typename GfxFamily::MI_SEMAPHORE_WAIT);
return totalMiSemaphoreWaitSize + sizeof(typename GfxFamily::MI_ATOMIC);
}