DVD-RB, Q6600 and the new HC v0.22 Speedtest...

[DD51]

Just wanted to share a test I did to see the difference running Multiple Encoder Processes Checked vs Unchecked on a Q6600 CPU.

First off some specs...
DVD Source: The Australian Pink Floyd Show
Overclocked Quad-Core Q6600 running at 450x8 (3.6GHZ)
Memory DDR2-800 running 1:1

Here's the result running with Multiple Encoder Processes Checked
[10:01:28] One Click encoding activated...
-----------------
[10:01:28] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.7.0
- HC v0.22.0.0 encoder selected
- Source: APFS
- VTS_01: 3,034,777 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 267,000 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 71.2%
- Overall Bitrate : 3,623/2,898Kbs
- Space for Video : 3,939,598KB
- Redistributing using Base_Q: 35
- HIGH/LOW/TYPICAL Bitrates: 3,582/1,211/2,898 Kbs
[10:05:26] Phase I, PREPARATION completed in 4 minutes.
[10:05:26] Phase II ENCODING started
- Creating M2V for VTS_01 segment 0
- ...........
- Creating M2V for VTS_01 segment 22
[10:31:40] Phase II ENCODING completed in 26 minutes.

Here's the result running with Multiple Encoder Processes Unchecked
[10:54:19] Phase II ENCODING started
- Creating M2V for VTS_01 segment 0
- Creating M2V for VTS_01 segment 1
- ...........
- Creating M2V for VTS_01 segment 22
[11:26:09] Phase II ENCODING completed in 32 minutes.

It took 6 minutes more to complete.
Thanks Hank for having multicore support.

Kind regards,

[blutach]

Very jealous of that perfomance :)

Regards

[Morden]

Very nice...Im running a Quadcore Q6600 stock with 4gb of 800mhz ocz ram.

I have to agree that 0.22 is running nice and smooth with the multicore. Here's a log from Bands of Brothers NTSC disc 1

[00:01:51] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.7.0
- HC v0.22.0.0 encoder selected
- Source: DVD DECRYPTED
- VTS_01: 3,401,398 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 182,214 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 68.8%
- Overall Bitrate : 4,788/3,831Kbs
- Space for Video : 3,553,730KB
- HIGH/LOW/TYPICAL Bitrates: 4,870/786/3,831 Kbs
[00:05:48] Phase I, PREPARATION completed in 4 minutes.
[00:35:26] One Click encoding activated...
[00:35:26] Phase II ENCODING started
- Creating M2V for VTS_01 segment 0
- Creating M2V for VTS_01 segment 1
- Creating M2V for VTS_01 segment 2
- Creating M2V for VTS_01 segment 3
- Creating M2V for VTS_01 segment 4
- Creating M2V for VTS_01 segment 5
- Creating M2V for VTS_01 segment 6
- Creating M2V for VTS_01 segment 7
- Creating M2V for VTS_01 segment 8
- Creating M2V for VTS_01 segment 9
- Creating M2V for VTS_01 segment 10
- Creating M2V for VTS_01 segment 11
- Extracting STILLS for VTS_01 segment 12
[01:04:15] Phase II ENCODING completed in 29 minutes.
[01:04:16] Phase III, REBUILD started.
- Copying IFO, BUP, and unaltered files...
- Processing VTS_01
- Reading/processing TMAP table...
- Rebuilding seg 0 VOBID 1 CELLID 1
- Rebuilding seg 1 VOBID 1 CELLID 2
- Rebuilding seg 2 VOBID 1 CELLID 3
- Rebuilding seg 3 VOBID 1 CELLID 4
- Rebuilding seg 4 VOBID 1 CELLID 5
- Rebuilding seg 5 VOBID 1 CELLID 6
- Updating NAVPACKS for VOBID_01
- Rebuilding seg 6 VOBID 2 CELLID 1
- Rebuilding seg 7 VOBID 2 CELLID 2
- Rebuilding seg 8 VOBID 2 CELLID 3
- Rebuilding seg 9 VOBID 2 CELLID 4
- Rebuilding seg 10 VOBID 2 CELLID 5
- Rebuilding seg 11 VOBID 2 CELLID 6
- Updating NAVPACKS for VOBID_02
- Rebuilding seg 12 VOBID 3 CELLID 1
- Updating NAVPACKS for VOBID_03
- Updated VTS_C_ADT.
- Updated VTS_VOBU_ADMAP.
- Updated IFO: VTS_01_0.IFO
- Updating TMAP table...
- Correcting VTS Sectors...
[01:13:48] Phase III, REBUILD completed in 9 minutes.

Done.
[01:13:48] PREPARE/ENCODE/REBUILD completed in 38 min.

Keep up the great job Hank!

[linx05]

Wow, I am insanely jealous about those speeds!

[Narvi]

I agree. Jealousy.

[GZZ]

How is the load on the 4 cores, are they maxed out @ 90-100 %. Both with Multi encoder checked and unchecked. ???

[Fishman0919]

Multiple Encoder Processes with DVD-RB and the multiprocessor support (SMP) with HC is not the same.

Multiple Encoder Processes with DVD-RB make use of all the cores or CPU's by running multiple instances of the encoder for each core or cpu.

HC now has multiprocessor support (SMP) which has to be turned on by either use command line or with the GUI. This makes a single instance of HC utilize all the cores available.

[DD51]

Regarding CPU usage:
With Multiple Encoder Processes Checked
4 instances of HC open all using 166MB of memory with 100% CPU usage

With Multiple Encoder Processes Unchecked
ONLY 1 instance of HC open using 163MB of memory with CPU usage between 65% - 83%.

I tried this with 2 different discs with the same results.

With MEP checked it is way faster!

Regards,

PS. BTW the encoding that I did (1st post) had a running time of 2:28:28.
Completed in 26 minutes. Yes it is ridiculously fast!
I thought my 3.2E OC'd to 4GHZ was fast! The Quad's are so much faster it's not even funny.:p

Regards,

[laserfan]

Would y'all please tell us the mobos you're using too...!

[Fishman0919]

Some other tests with DVD-RB an Quad Core

http://forum.doom9.org/showthread.php?p=1031275#post1031275

[Morden]

My mobo is an ASUS P5E-VM DO

[ChickenMan]

I'm getting same/similar encoding times on my Q6600 with same RAM. Using an Asus P5K Premium /w WiFi mobo. A mate has the same cpu and memory but the Asus P5K Deluxe /w WiFi mobo with same speed results.

I have found HCenc 0.22 with MEP checked (cpu usage 100%) takes the same time to encode as CCE 2.70 (cpu usage 50-60%).

[A64]

If you add Encode_Processes=2 to the rebuilder.ini instead of allowing the four copies run for the MEP, it helps with HC 0.22.

Doing Ratatouille, redistributed movie only, these are the numbers I got with a stock Q6600 with 4Gb of DDR2-667 PNY memory on the ASRock 4Core1333-Viiv motherboard.

Encode with no MEP: 30 minutes
Encode with MEP - 4 instances: 25 minutes
Encode with MEP - 2 instances: 22 minutes

[Clawz]

With Procoder 2.04, I am finding that Encode_Processes=3 to the rebuilder.ini is the way to fully populate the cores. Allowing it to be =4 just slows it down by a few minutes.

[hallway]

I've only got a dual-core in my work machine but I'm *impressed* with the improvements over my home machine with an AMD XP2500+. A re-encode on the home PC will take 200-300 minutes while I haven't had a job take over 60 on the work PC. Need to update my home PC !!! :D

[jollyrogr]

with my antique SMP athlon XP machine with 1gb ram running win2k3 server:

[18:02:40] PREPARE/ENCODE/REBUILD completed in 174 min.


The prices of those new Dell quad core machines have me really tempted...

[EvanOz85]

Can someone please help. I have MEP checked, and yet I'm only getting an average of about 40% CPU usage, and the encoding times are significantly longer than the ones reported here. I haven't overclocked the processor at all, and I have 4GB of ram with Vista. Any ideas why this is happening? Thanks

[Fishman0919]

@EvanOz85

Some more info would be helpful...
What encoder are you using?
Ver of DVD-RB?
What CPU are you using?
Any and all other info you can tell would be helpful...

[EvanOz85]

Well this thread was for HC version 0.22, with the Q6600 processor, so I didn't think that was necessary. But that's what I'm using. DVD-RB 1.26.5. Here is a log from a DVD I just encoded that only needed 96.6% compression. Look how long it took. Even the rebuild took forever. This was using the "Best" quality of HC.

[21:12:46] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH UNKNOWN
- HC v0.22.0.0 encoder selected
- AVS Filters are enabled.
- Source: DVD
- VTS_01: 2,304,586 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 148,758 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 96.5%
- Overall Bitrate : 6,796/5,437Kbs
- Space for Video : 4,117,538KB
- HIGH/LOW/TYPICAL Bitrates: 6,169/1,178/5,437 Kbs
[21:15:40] Phase I, PREPARATION completed in 3 minutes.
[21:16:25] Phase II ENCODING started
- Creating M2V for VTS_01 segment 0
- Creating M2V for VTS_01 segment 1
- Creating M2V for VTS_01 segment 2
- Extracting STILLS for VTS_01 segment 3
- Extracting STILLS for VTS_01 segment 4
- Extracting STILLS for VTS_01 segment 5
- Creating M2V for VTS_01 segment 6
- Creating M2V for VTS_01 segment 7
- Extracting STILLS for VTS_01 segment 8
- Extracting STILLS for VTS_01 segment 9
- Extracting STILLS for VTS_01 segment 10
- Creating M2V for VTS_01 segment 11
- Creating M2V for VTS_01 segment 12
- Extracting STILLS for VTS_01 segment 13
- Extracting STILLS for VTS_01 segment 14
- Extracting STILLS for VTS_01 segment 15
- Creating M2V for VTS_01 segment 16
- Creating M2V for VTS_01 segment 17
- Creating M2V for VTS_01 segment 18
- Extracting STILLS for VTS_01 segment 19
- Extracting STILLS for VTS_01 segment 20
- Extracting STILLS for VTS_01 segment 21
- Extracting STILLS for VTS_01 segment 22
- Extracting STILLS for VTS_01 segment 23
- Extracting STILLS for VTS_01 segment 24
- Extracting STILLS for VTS_01 segment 25
[21:59:15] Phase II ENCODING completed in 43 minutes.
[21:59:15] Phase III, REBUILD started.
- Copying IFO, BUP, and unaltered files...
- Processing VTS_01
- Reading/processing TMAP table...
- Rebuilding seg 0 VOBID 1 CELLID 1
- Rebuilding seg 1 VOBID 1 CELLID 2
- Rebuilding seg 2 VOBID 1 CELLID 3
- Updating NAVPACKS for VOBID_01
- Rebuilding seg 3 VOBID 2 CELLID 1
- Updating NAVPACKS for VOBID_02
- Rebuilding seg 4 VOBID 3 CELLID 1
- Updating NAVPACKS for VOBID_03
- Rebuilding seg 5 VOBID 4 CELLID 1
- Updating NAVPACKS for VOBID_04
- Rebuilding seg 6 VOBID 5 CELLID 1
- Rebuilding seg 7 VOBID 5 CELLID 2
- Updating NAVPACKS for VOBID_05
- Rebuilding seg 8 VOBID 6 CELLID 1
- Updating NAVPACKS for VOBID_06
- Rebuilding seg 9 VOBID 7 CELLID 1
- Updating NAVPACKS for VOBID_07
- Rebuilding seg 10 VOBID 8 CELLID 1
- Updating NAVPACKS for VOBID_08
- Rebuilding seg 11 VOBID 9 CELLID 1
- Rebuilding seg 12 VOBID 9 CELLID 2
- Updating NAVPACKS for VOBID_09
- Rebuilding seg 13 VOBID 10 CELLID 1
- Updating NAVPACKS for VOBID_10
- Rebuilding seg 14 VOBID 11 CELLID 1
- Updating NAVPACKS for VOBID_11
- Rebuilding seg 15 VOBID 12 CELLID 1
- Updating NAVPACKS for VOBID_12
- Rebuilding seg 16 VOBID 13 CELLID 1
- Rebuilding seg 17 VOBID 13 CELLID 2
- Rebuilding seg 18 VOBID 13 CELLID 3
- Updating NAVPACKS for VOBID_13
- Rebuilding seg 19 VOBID 14 CELLID 1
- Updating NAVPACKS for VOBID_14
- Rebuilding seg 20 VOBID 15 CELLID 1
- Updating NAVPACKS for VOBID_15
- Rebuilding seg 21 VOBID 16 CELLID 1
- Updating NAVPACKS for VOBID_16
- Rebuilding seg 22 VOBID 17 CELLID 1
- Updating NAVPACKS for VOBID_17
- Rebuilding seg 23 VOBID 18 CELLID 1
- Updating NAVPACKS for VOBID_18
- Rebuilding seg 24 VOBID 19 CELLID 1
- Updating NAVPACKS for VOBID_19
- Rebuilding seg 25 VOBID 20 CELLID 1
- Updating NAVPACKS for VOBID_20
- Updated VTS_C_ADT.
- Updated VTS_VOBU_ADMAP.
- Updated IFO: VTS_01_0.IFO
- Updating TMAP table...
- Correcting VTS Sectors...
[22:14:39] Phase III, REBUILD completed in 15 minutes.

[SpazzHH]

My guess is that it's related to AviSynth, since the version is reported as "Unknown". What version are you using? I would try re-installing it.

[blutach]

Plus he's using filters, some of which can take up lots of time.

Check your DMA and reinstall AVISynth 2.5.7 or above.

Regards

[jdobbs]

Well this thread was for HC version 0.22, with the Q6600 processor, so I didn't think that was necessary. But that's what I'm using. DVD-RB 1.26.5. Here is a log from a DVD I just encoded that only needed 96.6% compression. Look how long it took. Even the rebuild took forever. This was using the "Best" quality of HC.I'm not sure I'd consider 15 minutes "forever"... What exactly is your frame of reference?

Please post your REBUILDER.INI file so we can see what filters and settings are in use.

[jarthel]

how does dvd-rb tell hcencoder to use SMP? Or does HCencoder auto-detect the number of cores?

thank you very much :)

[Boulder]

HC022 utilizes multiple threads by default, and it detects the number of processors automatically. So you shouldn't use the "Multiple encoders" function in DVD-RB if you use HC022.

[blutach]

@Boulder

Is it actually quicker on Single Encoder? I've been running 22 since it came out but with MEP and it seems fine. But are you saying that a single encoder is quicker.

Regards

[Boulder]

If I could do it in DVD-RB, I would disable SMP and run two instances as I've noticed that is faster. Running two instances with SMP disabled is also faster than using the multithreaded Avisynth with SetMTMode and running one SMP-enabled HC022.

[jdobbs]

I'll add an option for NOSMP. Maybe I shoiuld disable it automatically when DVD-RB is running multiple copies of HC?

[hank315]

Running two instances without SMP is a bit faster, about 10% on my system, there's always a penalty running multi-threaded applications.
So yes, Boulder is right.


Maybe I shoiuld disable it automatically when DVD-RB is running multiple copies of HC?
Yes, please :)
Disadvantage: running multiple (2 or 4) instances takes a lot of memory, much more than one instance with SMP enabled.

[jdobbs]

Will do.

[valnar]

I think it's safe to assume anyone with a dual or quad core these days has 1GB+ RAM. Optimize DVDRBPro for the majority.

Robert

[blutach]

Yes, make it an option, disabled by default. Those with memory woes can re-enable SMP.

Regards

[jdobbs]

I just ran it a couple of times on a DVD. On my AMD 165 Dual Core (overclocked) it runs at the same speed whether enabled or disabled -- and here I was thinking I found another way to speed it up, no joy in Dobbsville.

Anyway, I'll make it an option.

[tom942]

But please, don´t disable SMP for the Bitrate distribution phase.

In my sytem (4200+ X2) it runs nearly twice faster than 0.21.

Regards.

[jdobbs]

You're not doing a 100% encode during REDISTRIBUTION are you? That really seems to me like a waste of time... but to each his/her own.

[tom942]

Yes I do :(. I prefer to be the most accurate possible, but I don´t have tested a lot with just a 10%.

BTW, is it possible not to disable SMP in Redistribution phase?.

Regards :).

[jdobbs]

I'll change it so it doesn't. :)