increase parallel block for LZ4 data compression #196
rourabpaul1986
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Hi I was using https://github.com/Xilinx/Vitis_Libraries/blob/master/data_compression/L2/demos/lz4_streaming/
I found this design has PARALLEL_BLOCK value of 8 (default). It also mentioned "Overall throughput can still be increased with multiple compute units."
Could you please inform me which file I should modify to increase the number of PARALLEL_BLOCK in each computation unit for better throughput?
I have already tried two approaches
I have found line 140 and and line 146 in [Makefile] mentioned DPARALLEL_BLOCK=8. I changed those DPARALLEL_BLOCK, but did not get the result.
I found #define PARALLEL_BLOCK 8 in the line 43 of lz4_compress_stream.hpp. I have changed this value but did not work for me
Beta Was this translation helpful? Give feedback.
All reactions