Comments (2)
Is the compressionLevel in ZSTD_sequenceProducer_F function always the same value as ZSTD_c_compressionLevel set by ZSTD_CCtx_setParameter?
Yes.
It's mostly informational btw. The sequence producer is free to do whatever it wants with this information, including ignoring it.
My assumption is that ZSTD_c_compressionLevel somehow influences the 'compress sequences' step.
It's true. It's more a side-effect though.
There are multiple cpu / compression trade-offs in the pipeline. They are roughly influenced by the compression level.
Now, because there are so many of them, I can't tell for sure which ones have impacted your observations.
As a simple example, the "minimum compression" threshold, which is both a guess and a verification deciding to not compress literals if it doesn't produce at least a certain amount of savings compared to "just store uncompressed", this threshold changes depending on compression level.
And there are others too. For example, there is a block splitter which is triggered at some point. This advanced strategy decides to cut a block into 2 or more parts if it can produce more adjusted statistics for each part, resulting in better compression ratio. This is expensive, but compared to the rest of the pipeline, it's negligible at high compression levels. This might be the one you observed in your tests.
So there are multiple heuristics decisions across the pipeline, and their behavior is influenced by the compression level. Sometimes, there are some advanced flags that can take over from the heuristics to unconditionally enable or disable the feature, sometimes not. This is mostly because these features were added later on, one by one, and there was a lack of common framework, resulting in each feature implementing its own slightly different way to be controlled and triggered. A refactoring could be needed at some point, but until then, the general idea is that "higher compression level" means "more cpu available for compression ratio", but the exact technical translation is more complex to detail (and changes with versions).
from zstd.
Thanks for the explanation.
from zstd.
Related Issues (20)
- MSVC CMake build failed on v1.5.6
- v1.5.6 Windows binary downloads are double zipped HOT 4
- Raise version's in win32 binaries header HOT 3
- Why was the new release 1.5.6 removed? HOT 15
- long file names are cut off in output HOT 3
- Should zstd check archive consistency before overwriting files? HOT 1
- Should zstd delete incomplete archives? HOT 5
- 32-bit x86 build failure with 1.5.6 HOT 3
- v1.5.6 breaks 32-bit Windows clang-cl build HOT 3
- Decompress multiple zstaa backups on FAT32 drives HOT 4
- Replication of bug #3517 HOT 16
- Separate dictionary references to enable dictionary usage for any combination of window size and content size HOT 1
- Decompression speed regression in zstd 1.5.6 (win)
- Embed hash of raw dictionary in compressed resource (optionally) HOT 4
- Decompression crash after upgrading from zstd 1.4.5 to 1.5.6 HOT 12
- Missing check on failed allocation leads to NULL-ptr dereference
- libzstd.lib missed in package, also VC sample seems include wrong mem.h or ambigious including!
- Environment variable for --memory HOT 2
- Improve misleading wording in the streaming decompression howto HOT 1
- erro
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 zstd.