Removed constant DDR_MAX_SIZE = 512. #4372
Merged
+17
−33
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
openvino-pushbot
added
category: inference
| @@ -480,10 +469,6 @@ std::size_t Allocator::freeCMXMemoryAmount() const { | |||
| return _maxCmxSize - offset; | |||
| } | |||
|
|
|||
| std::size_t Allocator::freeMemoryAmount(const MemoryType& type) const { | |||
| return type == MemoryType::CMX ? freeCMXMemoryAmount() : freeDDRMemoryAmount(); | |||
There was a problem hiding this comment.
Not sure about removing CMX check. CMX has the same capacity both for ma2480/ma2085
| @@ -481,7 +470,7 @@ std::size_t Allocator::freeCMXMemoryAmount() const { | |||
| } | |||
|
|
|||
| std::size_t Allocator::freeMemoryAmount(const MemoryType& type) const { | |||
There was a problem hiding this comment.
We can remove this method and leave only freeCMXMemoryAmount
| @@ -77,8 +77,6 @@ void updateChildDataAllocation(const Data& data, int offsetLimitation) { | |||
| } | |||
|
|
|||
| memoryOffset += byteOffset; | |||
|
|
|||
| IE_ASSERT(memoryOffset + child->lastElemOffset() <= offsetLimitation); | |||
There was a problem hiding this comment.
You are removing an assert in CMX case
You can pass _maxCmxSize and do something like
IE_ASSERT(parent->dataLocation() != Location::CMX || memoryOffset + child->lastElemOffset() <= offsetLimitation);
IvanPoliksenov
force-pushed
the
icv/ipolikse/rem_DDR_MAX_SIZE
branch
from
1881cae to
1b26a02
Compare
IvanPoliksenov
force-pushed
the
icv/ipolikse/rem_DDR_MAX_SIZE
branch
2 times, most recently
from
c5645a4 to
1b26a02
Compare
| @@ -108,7 +110,7 @@ class Allocator final { | |||
| void extractDatas(MemoryType memType, const DataSet& from, DataVector& out) const; | |||
|
|
|||
| std::size_t freeDDRMemoryAmount() const; | |||
There was a problem hiding this comment.
You forgot to remove the declaration
| namespace { | ||
|
|
||
| void updateChildDataAllocation(const Data& data, int offsetLimitation) { | ||
| void Allocator::updateChildDataAllocation(const Data& data) { |
There was a problem hiding this comment.
style: no new line after }
Comment on lines
525
to
531
| const auto freeSpace = (memType == MemoryType::CMX ? freeCMXMemoryAmount() : -1); | ||
| if (static_cast<std::size_t>(size) > freeSpace) { | ||
| return nullptr; | ||
| } |
There was a problem hiding this comment.
if (memType == MemoryType::CMX && static_cast<std::size_t>(size) > freeCMXMemoryAmount()) {
return nullptr;
}
Comment on lines
234
to
237
| VPU_THROW_UNLESS(!failedData || failedData->memReqs() == MemoryType::CMX, | ||
| R"(Stage "{}" of type "{}" requested {} bytes in {} for output "{}", while there is only {} bytes is free)", | ||
| failedStage->name(), failedStage->type(), calcAllocationSize(failedData), failedData->memReqs(), failedData->name(), | ||
| allocator.freeMemoryAmount(failedData->memReqs())); | ||
| failedData->memReqs() == MemoryType::CMX ? allocator.freeCMXMemoryAmount() : -1); |
There was a problem hiding this comment.
failedData->memReqs() == MemoryType::CMX is always false in error message arguments. Take a look at the condition of assertion.
I'd suggest something like this:
VPU_THROW_UNLESS(!failedData || failedData->memReqs() == MemoryType::CMX,
R"(Request {} bytes in {} for output "{}" failed for stage "{}" of type "{}"",
calcAllocationSize(failedData), failedData->memReqs(), failedData->name(), failedStage->name(), failedStage->type());
AntonDudchenko
approved these changes
Apr 13, 2021
Maxim-Doronin
approved these changes
Apr 15, 2021
|
Please, rebase your branch on the latest master to re-trigger pre-commit validation |
Removed the DDR_MAX_SIZE constant as it could potentially lead to incorrect behavior of devices with a different DDR size (Prism Creek can be up to 2 GB in size). Removed the use of this constant in methods.
Signed-off-by: Ivan Poliksenov <[email protected]>
Signed-off-by: Ivan Poliksenov <[email protected]>
Signed-off-by: Ivan Poliksenov <[email protected]>
IvanPoliksenov
force-pushed
the
icv/ipolikse/rem_DDR_MAX_SIZE
branch
from
e79f95f to
48f3b09
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Removed the DDR_MAX_SIZE constant as it could potentially lead to incorrect behavior of devices with a different DDR size (Prism Creek can be up to 2 GB in size). Removed the use of this constant in methods.