AI/rocm-systems
2
0
Förgrening 0
Kod Ärenden Pull-förfrågningar Actions Packages Projekt Släpp Wiki Aktiviteter

P4 to Git Change 1505728 by wchau@wchau_OCL_boltzmann on 2018/01/19 13:13:53

Bläddra i källkod 
SWDEV-137270 - Add findLocalWorkSize to OpenCL runtime ROC device
	- Add support for kernel using image argument
	- fix the issue of not using the max workgroup size specified by the environment variables "maxWorkGroupSize*"

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rockernel.cpp#33 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rockernel.hpp#19 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocsettings.cpp#31 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocvirtual.cpp#50 edit
This commit is contained in:
foreman
2018-01-19 13:33:55 -05:00
förälder edb9746999
incheckning 87ad3f9692
4 ändrade filer med 110 tillägg och 44 borttagningar
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/runtime/device/rocm/rocvirtual.cpp
+67 -44
Visa fil
@@ -1401,6 +1401,7 @@ void setRuntimeCompilerLocalSize(hsa_kernel_dispatch_packet_t& dispatchPacket,
amd::NDRangeContainer sizes, device::Kernel* devKernel,
const roc::Device& dev) {
Kernel& gpuKernel = static_cast<Kernel&>(*devKernel);
const size_t* compile_size = devKernel->workGroupInfo()->compileSize_;
// Todo (sramalin) need to check if compile_size is set to 0 if dimension is not valid
@@ -1425,53 +1426,75 @@ void setRuntimeCompilerLocalSize(hsa_kernel_dispatch_packet_t& dispatchPacket,
// Find threads per group
thrPerGrp = devKernel->workGroupInfo()->size_;
size_t tmp = thrPerGrp;
// Split the local workgroup into the most efficient way
for (uint d = 0; d < sizes.dimensions(); ++d) {
size_t div = tmp;
for (; (sizes.global()[d] % div) != 0; div--)
;
sizes.local()[d] = div;
tmp /= div;
}
// Assuming DWORD access
const uint cacheLineMatch = dev.info().globalMemCacheLineSize_ >> 2;
// Check if partial dispatch is enabled and
if (dev.settings().partialDispatch_ &&
// we couldn't find optimal workload
((sizes.local().product() % devKernel->workGroupInfo()->wavefrontSize_) != 0 ||
// or size is too small for the cache line
(sizes.local()[0] < cacheLineMatch))) {
size_t maxSize = 0;
size_t maxDim = 0;
for (uint d = 0; d < sizes.dimensions(); ++d) {
if (maxSize < sizes.global()[d]) {
maxSize = sizes.global()[d];
maxDim = d;
}
}
if ((maxDim != 0) && (sizes.global()[0] >= (cacheLineMatch / 2))) {
sizes.local()[0] = cacheLineMatch;
thrPerGrp /= cacheLineMatch;
sizes.local()[maxDim] = thrPerGrp;
for (uint d = 1; d < sizes.dimensions(); ++d) {
if (d != maxDim) {
sizes.local()[d] = 1;
}
}
if (gpuKernel.imageEnable() &&
// and thread group is a multiple value of wavefronts
((thrPerGrp % devKernel->workGroupInfo()->wavefrontSize_) == 0) &&
// and it's 2 or 3-dimensional workload
(sizes.dimensions() > 1) &&
((dev.settings().partialDispatch_) ||
(((sizes.global()[0] % 16) == 0) && ((sizes.global()[1] % 16) == 0)))) {
// Use 8x8 workgroup size if kernel has image writes)
if (gpuKernel.imageWrite() || (thrPerGrp != dev.settings().preferredWorkGroupSize_)) {
sizes.local()[0] = 8;
sizes.local()[1] = 8;
}
else {
// Check if a local workgroup has the most optimal size
if (thrPerGrp > maxSize) {
thrPerGrp = maxSize;
}
sizes.local()[maxDim] = thrPerGrp;
sizes.local()[0] = 16;
sizes.local()[1] = 16;
}
if (sizes.dimensions() == 3) {
sizes.local()[2] = 1;
}
}
else {
size_t tmp = thrPerGrp;
// Split the local workgroup into the most efficient way
for (uint d = 0; d < sizes.dimensions(); ++d) {
size_t div = tmp;
for (; (sizes.global()[d] % div) != 0; div--)
;
sizes.local()[d] = div;
tmp /= div;
}
// Assuming DWORD access
const uint cacheLineMatch = dev.info().globalMemCacheLineSize_ >> 2;
// Check if partial dispatch is enabled and
if (dev.settings().partialDispatch_ &&
// we couldn't find optimal workload
((sizes.local().product() % devKernel->workGroupInfo()->wavefrontSize_) != 0 ||
// or size is too small for the cache line
(sizes.local()[0] < cacheLineMatch))) {
size_t maxSize = 0;
size_t maxDim = 0;
for (uint d = 0; d < sizes.dimensions(); ++d) {
if (d != maxDim) {
sizes.local()[d] = 1;
if (maxSize < sizes.global()[d]) {
maxSize = sizes.global()[d];
maxDim = d;
}
}
if ((maxDim != 0) && (sizes.global()[0] >= (cacheLineMatch / 2))) {
sizes.local()[0] = cacheLineMatch;
thrPerGrp /= cacheLineMatch;
sizes.local()[maxDim] = thrPerGrp;
for (uint d = 1; d < sizes.dimensions(); ++d) {
if (d != maxDim) {
sizes.local()[d] = 1;
}
}
}
else {
// Check if a local workgroup has the most optimal size
if (thrPerGrp > maxSize) {
thrPerGrp = maxSize;
}
sizes.local()[maxDim] = thrPerGrp;
for (uint d = 0; d < sizes.dimensions(); ++d) {
if (d != maxDim) {
sizes.local()[d] = 1;
}
}
}
}