Comments (3)
Investigation 1:
std:;unique_ptr<>
does have a negligible performance impact compared to raw pointers. In fact, in one test, gcc and clang report slightly faster speed than raw pointers. PGI seems to suffer from std::unique_ptr<>
overhead though.
More detailed results can be found in this PDF:
Performance Impact of STL Structures - Sheet1.pdf
from sperr.
Investigation 2:
The vast majority of time consuming std::vector<>
operations are on push_back's of the LIS and LSP lists.
More specifically, the data movement incurred when push_back causes a capacity increase.
PR #6 partially addressed this problem.
from sperr.
Investigation 3:
After implementing #11 , the performance characteristics changed a bit. It's now obvious that the cost from std::vector<>
is mostly related to memory allocation/deallocation.
However, with the current state of the code, there are simply arrays with unknown sizes: LIS and LSP. I have already tweaked LIS so that it still reserves 2X capacity even when garbage collection, and LSP only grows but won't shrink. So, I don't think there's much I can wrt memory usage here.
from sperr.
Related Issues (20)
- Bit budget calculation in fixed-size mode is still not quite right
- Compression fails for large quantization values HOT 1
- Consider removing the divide by rms preconditioning feature HOT 1
- Bit packing can benefit from std::bitset using clang HOT 1
- Not recording chunk info when there is only one chunk
- Outlier Coding can perform a few more levels of refinements to improve accuracy gain
- Outlier Coding termination can be smarter
- Profile memory again and potentially use PMR in local objects HOT 1
- CLI11 weird behavior HOT 1
- SPERR 0.5 occasionally violates error tolerance HOT 2
- PWE tolerance is violated when very small HOT 9
- SPERR does not build with OpenMP on macOS HOT 1
- Data is compressed wrong HOT 19
- Investigate performance implications of ZFP bitstream HOT 1
- Investigate performance implications of boost::dynamic_bitset HOT 2
- Potential improvement in function m_partition_S_XYZ() HOT 1
- More refined set size ordering in function m_partition_S_XYZ() HOT 1
- Potential optimization: a small enough set would just keep the subvolume, instead of indices. HOT 1
- Where did all the symbols go? HOT 2
- A few things to consider
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from sperr.