-
looks fantastic and thanks for the work.
when can we expect xxh3 1.0?
hiqbn updated
5 years ago
-
Hi,
Line 63 of xxh3.h defines out a C++ keyword and leaves it defined in the header. This is a problem when interacting with other code. Could it add an #undef restrict somewhere or, better yet, us…
-
[rpmbuild@rhel5-x86-64 xxHash-0.7.0]$ gcc -v
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --enable…
-
According to @hgnan when running on Jetson Nano with CUDA, LLVM throws fatal error `Cannot select: intrinsic %llvm.nvvm.atomic.add.gen.i.sys`:
```
jetson@jetson-nano:~/taichi$ python3 mpm.py
…
-
First of all thanks for adding XXH3. It's an awesome addition to xxHash. While the recent sources build fine with MinGW, they failed to compile with MSVC (version19.16.27027, Visual Studio 2017 15.9.9…
-
Hi Reini,
please consider taking a look at my latest-n-fastest etude - 'Yorikke':
The source code along with the superheavy (1 trillion unique keys) benchmark is freely downloadable at:
www.sanma…
-
If you use the crc32 instruction properly, available since Nehalem (SSE 4.2), you can achieve throughput of 1.17 cycles per 8 bytes, which would be a theoretical performance of 20.5 GB/s on a 3ghz pro…
eloff updated
5 years ago
-
This causes this warning (or error with `-Werror`):
```
In file included from xxhash.c:1017:
xxh3.h:50: error: "NDEBUG" redefined [-Werror]
#define NDEBUG
```
-
Is the benchmark source code available anywhere (["Presenting XXH3" article](https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html))?
I would like to use the same tool for measurements (a…
-
Dear rurban:
wyhash have been updated significantly with the "threat" of XXH3. The cycles per hash have been reduced by 1.5 on average. The long key speed is much higher. Also portability has been …