We should keep ZSTD_CCtx_s.params as structure
instead of pointer.
Thanks & Regards,
Vaneet Narang
ich uses negligible stack (< 200 bytes).
>I think I'll schedule the patchset for 5.2-rcX so that zstd is actually
>usable on arm in 5.2. Does that sound OK?
OK
Regards,
Vaneet Narang
ntinue_internal)-> 136 -> 88
> (ZSTD_compressCCtx) -> 192 -> 64
> (zstd_compress) -> 144 -> 96
Regards,
Vaneet Narang
changed:
>e24dd080sub sp, sp, #128 ; 0x80
Regards,
Vaneet Narang
#define M2_MAX_LEN 8
#define M3_MIN_LEN 3
#define M3_MAX_LEN 33
#define M4_MIN_LEN 3
#define M4_MAX_LEN 9
#define M1_MARKER 0
#define M2_MARKER 64
#define M3_MARKER 32
#define M4_MARKER 16
Similarly for LZ4 Dyn, we have used 1 bit as a marker to determine offset
length.
Thanks & Regards,
Vaneet Narang
rcptInfo.txt
Description: Binary data
a32k
lzbench 1.7.3 (32-bit Linux) Assembled by P.Skibinski
Compressor name Compress. Decompress. Compr. size Ratio Filename
memcpy 5273 MB/s 5320 MB/s 32768 100.00 data/data32k
lzo1x 2.09 -1 283 MB/s 465 MB/s 14292 43.62 data/data32k
Regards,
Va
n to both
>LZ4 and LZ4_DYN decompression speed.
Since there is no backward compatibility of our solution with original LZ4 so we
are bit confused if we should make it as separate API to avoid overhead of
dynOffset
checks every where in the code and also to avoid changing prototype of exported
functions
LZ4_decompress/LZ4_compress OR we should keep these checks to avoid redundant
code.
Kindly suggest
Thanks & Regards,
Vaneet Narang