PMT - Parallel Multithreaded Encoder/Decoder

[KaktoR]

Not yet, but it's ok neverthless because rz is lightning fast in decompression

What other compressors do you suggest? What about zstd?

[78372]

ZSTD has mt and way too fast to use with PMT.
Try with slow tools.
BCM, precomp, rzm etc. can work better.
We have.alternatives for them though.

[Cesar82]

Has anyone done any tests using PMT + pZLib3?

If yes please share the values for PMT.ini and Arc.ini.
Thanks!

[78372]

Originally Posted by Cesar82

Has anyone done any tests using PMT + pZLib3?

If yes please share the values for PMT.ini and Arc.ini.
Thanks!

What's the point? pZlib is already MT and fast. Moreover, you have xtool now.

[Cesar82]

@78372
Thanks for the information. I know little about compression methods.
So of the compressors used in CIU only RAZOR has benefits if used with PMT.

CIU supports these compressors in addition to some RAZOR tools for specific games.
SREP, ZTool, XTool(not v2019), pZLib3, UELR, AFR, OodleRec [4,5,6,7,8], MSC, MPZ, OGGRE, BPK, 7z, LOLZ, ZSTD, RAZOR

Does any of this list other than razor benefit if used with PMT?

[78372]

Originally Posted by Cesar82

@78372
Thanks for the information. I know little about compression methods.
So of the compressors used in CIU only RAZOR has benefits if used with PMT.

CIU supports these compressors in addition to some RAZOR tools for specific games.
SREP, ZTool, XTool(not v2019), pZLib3, UELR, AFR, OodleRec [4,5,6,7,8], MSC, MPZ, OGGRE, BPK, 7z, LOLZ, ZSTD, RAZOR

Does any of this list other than razor benefit if used with PMT?

mpz only I guess, at least for now.

[KaktoR]

mpz with PMT
Compressed 64 files, 209,434,083 => 180,728,551 bytes. Ratio 86.29%
Compression time: cpu 0.20 sec/real 38.46 sec = 1%. Speed 5.45 mB/s

Tested 64 files, 180,728,551 => 209,434,083 bytes. Ratio 86.29%
Testing time: cpu 0.19 sec/real 36.54 sec = 1%. Speed 5.73 mB/s

mpz only
Compressed 64 files, 209,434,083 => 180,670,089 bytes. Ratio 86.27%
Compression time: cpu 0.30 sec/real 103.93 sec = 0%. Speed 2.02 mB/s

Tested 64 files, 180,670,089 => 209,434,083 bytes. Ratio 86.27%
Testing time: cpu 0.36 sec/real 95.90 sec = 0%. Speed 2.18 mB/s

Tested with settings -b64m -t100p for encode/decode with a Ryzen 5 2600 (12 threads).

I just wonder why ratio is a little bit better without pmt

[78372]

The same reason for everything, chunking.

[L0v3craft]

Is possible to add this?

https://github.com/encode84/bcm

it is BCM v1.5. If is not possible someone has a cls dll for BCM to make it multithreaded? Thanks.

[78372]

Originally Posted by L0v3craft

Is possible to add this?

https://github.com/encode84/bcm

it is BCM v1.5. If is not possible someone has a cls dll for BCM to make it multithreaded? Thanks.

BCM 1.4 works with this. I will make a stdio patch for 1.5 when I have time.

[CAT8K]

Thanks very much for this tool "78372"
If possible, I want the latest version from you "PMT_rel3"

[78372]

Originally Posted by CAT8K

Thanks very much for this tool "78372"
If possible, I want the latest version from you "PMT_rel3"

I don't think that the project requires an update, I haven't received such requests and I do not use it anyways.

[FitGirl]

Originally Posted by Simorq

@78372
Does not work with CLS-BPK.

C'mon, I have already answered to this. PMT is a BLOCK-BASED TRIMMER. IT SPLITS FILES INTO CHUNKS. BPK is a FILE-BASED ENCODER. It deals only with full files. PMT split bink files into chunks and they CANNOT BE COMPRESSED with BPK anymore. "L" for logic.

It's intended for codecs which rely on pure stream data - generic compressors, precompressors, MPZ (cause it deals with frames in MP3, not full files).

[DiCaPrIo]

Originally Posted by L0v3craft

Hi. Is possible to add bpk compatibility? Thanks.

see at first page