note that the size of individual compressed blocks will vary more wildly with this modification.
But it seems good enough for a first test, and fix the speed regression issue.
Further refinements can be attempted later.
While it's not always strictly a win,
it's a win for files that see a noticeably compression ratio increase,
while it's a very small noise for other files.
Downside is, this patch is less efficient for 32-bit arrays of integer
than the previous patch which was introducing losses for other files,
but it's still a net improvement on this scenario.
helps when litlen==1 is cheaper than litlen==0
works great on pathological arr[u32] examples
but doesn't generalize well on other files.
silesia/x-ray is amoung the most negatively affected ones.
this works great for 32-bit arrays,
notably the synthetic ones, with extreme regularity,
unfortunately, it's not universal,
and in some cases, it's a loss.
Crucially, on average, it's a loss on silesia.
The most negatively impacted file is x-ray.
It deserves an investigation before suggesting it as an evolution.
* Add ZSTD_CCtxParams_registerSequenceProducer() to public API
* add unit test
* add docs to zstd.h
* nits
* Add ZSTDLIB_STATIC_API prefix
* Add asserts
If the relevant allocation returns NULL, ZSTD_createCDict_advanced_internal()
will return NULL. But ZSTD_createCDict_advanced2() doesn't check for
this and attempts to use the returned pointer anyway, which leads to
a segfault.
This PR introduces no functional changes. It attempts to change all
macros currently using `{ }` or some variant of that to to
`do { } while (0)`, and introduces trailing `;` where necessary.
There were no bugs found during this migration.
The bug in Visual Studios warning on this has been fixed since VS2015.
Additionally, we have several instances of `do { } while (0)` which have
been present for several releases, so we don't have to worry about
breaking peoples builds.
Fixes Issue #3830.
`HUF_DecompressFastArgs_init()` was adding 0 to NULL. Fix it by exiting
early for empty outputs. This is no change in behavior, because the
function was already exiting 0 in this case, just slightly later.
* Rename `ilimit` to `ilowest` and set it equal to `src` instead of
`src + 6 + 8`. This is safe because the fast decoding loops guarantee
to never read below `ilowest` already. This allows the fast decoder to
run for at least two more iterations, because it consumes at most 7
bytes per iteration.
* Continue the fast loop all the way until the number of safe iterations
is 0. Initially, I thought that when it got towards the end, the
computation of how many iterations of safe might become expensive. But
it ends up being slower to have to decode each of the 4 streams
individually, which makes sense.
This drastically speeds up the Huffman decoder on the `github` dataset
for the issue raised in #3762, measured with `zstd -b1e1r github/`.
| Decoder | Speed before | Speed after |
|----------|--------------|-------------|
| Fallback | 477 MB/s | 477 MB/s |
| Fast C | 384 MB/s | 492 MB/s |
| Assembly | 385 MB/s | 501 MB/s |
We can also look at the speed delta for different block sizes of silesia
using `zstd -b1e1r silesia.tar -B#`.
| Decoder | -B1K ∆ | -B2K ∆ | -B4K ∆ | -B8K ∆ | -B16K ∆ | -B32K ∆ | -B64K ∆ | -B128K ∆ |
|----------|--------|--------|--------|--------|---------|---------|---------|----------|
| Fast C | +11.2% | +8.2% | +6.1% | +4.4% | +2.7% | +1.5% | +0.6% | +0.2% |
| Assembly | +12.5% | +9.0% | +6.2% | +3.6% | +1.5% | +0.7% | +0.2% | +0.03% |
gcc in the linux kernel was not unrolling the inner loops of the Huffman
decoder, which was destroying decoding performance. The compiler was
generating crazy code with all sorts of branches. I suspect because of
Spectre mitigations, but I'm not certain. Once the loops were manually
unrolled, performance was restored.
Additionally, when gcc couldn't prove that the variable left shift in
the 4X2 decode loop wasn't greater than 63, it inserted checks to verify
it. To fix this, mask `entry.nbBits & 0x3F`, which allows gcc to eliete
this check. This is a no op, because `entry.nbBits` is guaranteed to be
less than 64.
Lastly, introduce the `HUF_DISABLE_FAST_DECODE` macro to disable the
fast C loops for Issue #3762. So if even after this change, there is a
performance regression, users can opt-out at compile time.
the flexArray in structure FSE_DecompressWksp
is just a way to derive a pointer easily,
without risk/complexity of calculating it manually.
Not sure if this change is good enough to avoid ubsan warnings though.
