Releases: fcorbelli/zpaqfranz
Windows 32/64 binary
In sum() new hasher whirlpool
zpaqfranz sum * -whirlpool
Exposed zfs commands
- zfslist => list existing snapshots
- zfspurge => delete snapshot
- zfsadd => freeze into a zpaq archive many snapshot at once
zpaqfranz -h zfs
zpaqfranz zfslist "*" "*"
zaqpfranz zfsadd "/tmp/47/zpaqfranz" "/temporaneo/kongo7.zpaq" "tank/d@2021" "--7d"
zfspurge "tank" "--7d"
Windows 32/64 binary
Windows 32/64 binary
Just for fun in the b (benchmark) now the
-all (multithread)
-tX (when used with -all limit to X thread. Note NO SPACE)
Beware: high CPU load brings high temperature, watch your cooler!
Benchmark all: b
Benchmark all on 1.86GB: b -minsize 2000000000
Benchmark all on 1MB: b -minsize 1048576
Benchmark SHA256 and BLAKE3: b -sha256 -blake 3 -minsize 1048576
Benchmark for 10 second each: b -n 10 -sha256 -blake3 -minsize 1048576
Cook the CPU (all cores): b -all -n 20 -blake3
Cook the CPU (8 cores): b -all -t8 -n 20 -blake3
Windows 32/64 binary
Windows 32/64 binary
Windows 32/64 binary
I found a bug in gcc 10.3 (default for latest Ubuntu 21)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101558
so I try a workaroud (I hope that everything will runs fine!)
Then another strange behavior of some Linux gcc compiler (!), that do strange things with 64-byte aligned data.
Therefore I made a massive rework of XXH3 interface.
Now the default hasher is XXHASH64, as the -128 (or -xxh3) needs more test.
So yes, even the compiler can broke your valuable software!
Now
-xxhash = XXHASH64: new default
-xxh3 = XXH3 (XXHASH 128 bit)
Experimental -wyhash (just a test)
Windows 32/64 binary
New switch -filelist
zpaqfranz a z:\1.zpaq c:\nz -filelist
zpaqfranz x z:\1.zpaq -filelist
New behavior: store by default XXH3 and CRC-32 of every file (to be completed)
New switches -checksum and -summary in list
zpaqfranz l z:\1.zpaq -checksum (show CRC and hash if any)
zpaqfranz l z:\1.zpaq -checksum -summary (compact output)