BD-Rebuilder - other suggested non-cpu optimizations for better encode speed [Archive]

View Full Version : BD-Rebuilder - other suggested non-cpu optimizations for better encode speed

Pages : [1] 2

jfcarbel

16th July 2011, 09:56

I know that CPU is most important way to boost BD-Rebuilder speeds.

Can read more about that at these threads here (http://forum.doom9.org/showthread.php?t=146589&page=2)and here (http://forum.doom9.org/showthread.php?t=152770)
And talk of processor x264 benchmarks here (http://forum.doom9.org/showthread.php?t=155791&page=2)

But once someone has maxed out to top level CPU, then what other hardware/software optimizations would be next to give the biggest performance increases?

We know from others tests that memory plays very little importance in speed improvements.

But what about SSD, OS software, 32-bit vs 64-bit?

Wondering how much an SSD would factor into speed improvements or XP 64-bit or Win7 64-bit vs 32-bit XP?

Perhaps folks can post their experiences.

jdobbs

16th July 2011, 14:02

RobertM

16th July 2011, 15:22

I noticed a significant speed improvement when I installed a second HDD. I have an 80GB SSD just for the OS (Win7x64), and there was not nearly enough space left for any BD work. So, initially, I did all work on a single 1TB drive. When I installed a second HDD, 2TB, I put the rips on one and I use the other for the rebuild target. There was a noticeable increase in speed. I have a good processor (i7-950 w/ 6GB RAM) so it would seem that the reading/writing to the same drive was a bottleneck.

jfcarbel

16th July 2011, 18:58

I have a good processor (i7-950 w/ 6GB RAM) so it would seem that the reading/writing to the same drive was a bottleneck.

Ah, yes, that was a big performance increaser for me as well. The next was when I upgraded to WD Black edition drives.

Just wondering if folks get a significant boost from an SSD for the source and target and what magnitude of improvement one would see. Then again with SSD prices its unlikely we have someone here with 2 SSD drives.

Could you maybe post your results in this thread (http://forum.doom9.org/showthread.php?p=1513682#post1513682) as I do not believe anyone with i7-950 has commented in this thread.

The other is how much of a difference will an upgrade to 64-but OS help. I am not sure but I think BD-Rebuilder even on a 64-bit computer is using the 32-bit x264.exe but I could be wrong. Any 64-bit users care to comment here.

jfcarbel

16th July 2011, 18:59

From my experience once you're at/over 4GB of system memory (on Win XP, Vista, or Win7), it doesn't seem to affect speed much. I think next in-line would be disc performance, especially in large (BD-25) encodes.

:) It was actually your post in another thread that I saw when I was saying - "We know from others tests that memory plays very little importance in speed improvements."

jdobbs

16th July 2011, 19:10

Ah, yes, that was a big performance increaser for me as well. The next was when I upgraded to WD Black edition drives.

Just wondering if folks get a significant boost from an SSD for the source and target and what magnitude of improvement one would see. Then again with SSD prices its unlikely we have someone here with 2 SSD drives.

Could you maybe post your results in this thread (http://forum.doom9.org/showthread.php?p=1513682#post1513682) as I do not believe anyone with i7-950 has commented in this thread.

The other is how much of a difference will an upgrade to 64-but OS help. I am not sure but I think BD-Rebuilder even on a 64-bit computer is using the 32-bit x264.exe but I could be wrong. Any 64-bit users care to comment here. If you're on a 64 bit computer and you have selected to use LAVF, the 64 bit version of X264 is used (any other way and it would fail). My experience is that it is marginally faster on some systems and marginally slower on others. In other words -- it really isn't much of a player.

jfcarbel

16th July 2011, 20:03

If you're on a 64 bit computer and you have selected to use LAVF, the 64 bit version of X264 is used (any other way and it would fail). My experience is that it is marginally faster on some systems and marginally slower on others. In other words -- it really isn't much of a player.

Thanks, good to know. So looks like I will focus on saving up for that SSD drive, that is, of course after my donation for those upcoming BD-Rebuilder blank titles feature ;)

In terms of SSD, folks should be aware of these issues posted here (http://www.anandtech.com/show/4202/the-intel-ssd-510-review/3) and here (http://www.codinghorror.com/blog/2011/05/the-hot-crazy-solid-state-drive-scale.html) and here (http://thessdreview.com/ssd-guides/buyers-guide/the-ssd-controller/) and here (http://blog.superuser.com/2011/05/10/maximizing-the-lifetime-of-your-ssd/). Looks like I would be leaning towards an Intel based SSD. I not sure whatever happened to SanDisk since at CES 2009 they were reporting some advanced SSD tech, but they have been much overshadowed by Crucial, OCZ, and Patriot brand SSDs. It would be interesting to see their failure rates since they are backed by a 10 year warranty.

taso1000

17th July 2011, 00:19

Hi, I have 3 Kingston SSDNOW 128GB, SNV425-S2/128GB, drives in raid0. I got them when they were being blown out for $140 after rebate. They have terrible I/O performance compared to other ssd's, but 10 times better than platter hard drives. I do movie only, 1080p to 720p, auto quality, hd audio, and custom size of 60gb from 120 minutes to 180 minutes. I have the movie iso on the ssd array too. System is i7 950 stock, 6gb @ 1066, os drive is Crucial 64gb M4. As long as I have the movie, I'm willing to do a conversion for you tonight. Just give me the settings you want it done as.

jfcarbel

17th July 2011, 00:39

As long as I have the movie, I'm willing to do a conversion for you tonight. Just give me the settings you want it done as.

I would say reducing a BD50 > BD25, movie only, using the auto settings

taso1000

17th July 2011, 01:13

Ok, I'll do the last star trek movie to bd-25, auto quality

taso1000

17th July 2011, 07:22

-----------------------
[19:32:41] BD Rebuilder v0.38.04 (beta)
- Source: STARTREK11D1AC
- Input BD size: 36.85 GB
- Approximate total content: [02:06:50.269]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[19:32:41] PHASE ONE, Encoding
- [19:32:41] Extracting A/V streams [VID_00000]
- [19:49:36] Reencoding: VID_00000 (1 of 1)
- [19:49:36] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 182,464 frames
- Bitrate: 18,055 Kbs
- [19:49:36] Reencoding: VID_00000, Pass 1 of 1
- [21:24:26] Video Encode complete
- [21:24:27] Reencoding audio tracks (if req'd)
- [21:24:27] Multiplexing M2TS
[21:30:00]PHASE ONE complete
[21:30:00]PHASE TWO - Rebuild Started
- [21:30:00] Rebuilding BD file Structure
[21:35:32] - Encode and Rebuild complete
[21:35:32]JOB: STARTREK11D1AC finished.

jfcarbel

17th July 2011, 07:39

-----------------------
[19:32:41] BD Rebuilder v0.38.04 (beta)
- Source: STARTREK11D1AC
- Input BD size: 36.85 GB
- Approximate total content: [02:06:50.269]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
[21:35:32]JOB: STARTREK11D1AC finished.

Wow, just over 2 hours only!

What SSD drives do you have the source on and what is the target SDD drive?

The extraction process only took about 17 minutes, thats pretty fast also.

Since you have SSDs I assume you are running Win7. Is that 64-bit Win7?

jdobbs

17th July 2011, 14:09

Wow, just over 2 hours only!

What SSD drives do you have the source on and what is the target SDD drive?

The extraction process only took about 17 minutes, thats pretty fast also.

Since you have SSDs I assume you are running Win7. Is that 64-bit Win7? That just doesn't seem that fast to me compared to SATA platter drives. I'm running a similar job now and will post the results when it is complete. I'm using standard SATA 1TB drives (one is used to hold the source and another for the working path) and an AMD X4 945 processor @3.53Ghz with 4GB of DDR2 memory.

I'm pretty sure the i7 950 is faster than my processor...

On my job the extraction took less than 9 minutes.

taso1000

17th July 2011, 15:07

The Star Trek iso and everything else was on the Kingston array. Yes, I'm running Windows 7 Home Premium 64 bit. The Star Trek iso was mounted with Virtual Clone Drive. I just noticed I didn't have write-back cache enabled in Intel Rapid Storage. I can run it again just out of curiousity. I'll report back with the new log file.

jdobbs

17th July 2011, 15:30

Capsbackup

17th July 2011, 15:57

Well one thing I just noticed:

- [21:24:27] Multiplexing M2TS
[21:30:00]PHASE ONE complete
[21:30:00]PHASE TWO - Rebuild Started
- [21:30:00] Rebuilding BD file Structure
[21:35:32] - Encode and Rebuild complete
[21:35:32]JOB: STARTREK11D1AC finished.

The remuxing/rebuild is probably the most disc intensive operation of the whole process. The remuxing and rebuild was really fast on the SSD -- taking only about 11 minutes end-to-end for a full 25GB rebuild. That's at least twice as fast (maybe more) as I would expect from a SATA drive.

All this speed is great! The cost for all this speed increase?? :rolleyes:
The economy must be turning around! :p
Sure hope the donations are improving as well, since without BD-RB, what else would benefit from all these speed increases ? ;) :)

jdobbs

17th July 2011, 16:30

Here's the job I mentioned. Interesting that my extract was faster than the SSD drive (<9 min compared to 17) -- but the mux and rebuild was much slower -- taking over 35 minutes on my system compared to 11 on the SSD.

Since the SSD report extracted about 19% more data, adjusting the extraction would still make it better by about 50% on the SATA drive. The muxing of the two is roughly equivalent since the target size is the same, making the SSD quite a bit faster.

[06:31:04] BD Rebuilder v0.38.04 (beta)
- Source: TRUE_GRIT_2010
- Input BD size: 30.76 GB
- Approximate total content: [01:50:20.405]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=1 DTS=1 HD=0 Kbs=640
[06:31:06] PHASE ONE, Encoding
- [06:31:06] Extracting A/V streams [VID_00800]
- [06:39:43] Reencoding: VID_00800 (1 of 1)
- [06:39:43] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 158,731 frames
- Bitrate: 26,748 Kbs
- [06:39:43] Reencoding: VID_00800, Pass 1 of 1
- [07:51:38] Video Encode complete
- [07:51:38] Reencoding audio tracks (if req'd)
- [07:55:18] Multiplexing M2TS
[08:12:55]PHASE ONE complete
[08:12:55]PHASE TWO - Rebuild Started
- [08:12:55] Rebuilding BD file Structure
[08:30:42] - Encode and Rebuild complete
[08:30:42]JOB: TRUE_GRIT_2010 finished.

All this speed is great! The cost for all this speed increase??
The economy must be turning around!
Sure hope the donations are improving as well, since without BD-RB, what else would benefit from all these speed increases ? I definitely can't afford SSD drives. I'm happy to have my $99 WD 1TB SATA drives.

If I relied on BD-RB donations for my system purchases I'd probably have to do a comparison with floppy discs. ;)

JMHO, but reducing a typical job's overall time by 20 minutes certainly wouldn't justify paying more for less space.

jfcarbel

17th July 2011, 18:35

I definitely can't afford SSD drives. I'm happy to have my $99 WD 1TB SATA drives.

If I relied on BD-RB donations for my system purchases I'd probably have to do a comparison with floppy discs. ;)

That is really disturbing considering how many people are using this program and how well it performs. I mean SlySoft CloneBD is like vaporware and this is the only real choice around.

I was thinking after seeing the thread about you working on the blanking feature that perhaps you should create a thread in which you come up with about 5-6 features to add and then have people donate per each feature they really desire (i.e. maybe a different subject line for each paypal). I will be going on my 3rd donation after I see the blanking feature added, but would have certainly donated for this. This solves 2 issues - you not working on this for free or for floppies :) and users getting the features they would want most.

I hope your product is nearing completion for which you can launch a paid Pro version since you certainly deserve compensation for your efforts.

jdobbs

17th July 2011, 19:42

I'm not planning to create a "Pro" version of BD-RB. I'm hoping to just keep it based on voluntary donations. DVD Rebuilder Pro's reputation suffered quite a bit from all the terribly hacked versions that attempted to bypass the downloaded keyfile. People were constantly reporting problems that were caused by clueless hackers, thinking I wouldn't know the difference. There were even versions that were "patched" with viruses. Most of the hacked versions didn't work well, some not if at all. I downloaded a couple of them and dumped them. Patches were done to totally innocuous sections of code -- only causing issues -- by people totally devoid of any knowledge.

I laugh every time I watch one of those movies that portray hackers as "computer geniuses".

I just don't want to go through that again.

Stereodude

17th July 2011, 20:24

taso1000

17th July 2011, 23:33

Here's the results with write back caching enabled:

-----------------------
[09:12:48] BD Rebuilder v0.38.04 (beta)
- Source: STARTREK11D1AC
- Input BD size: 36.85 GB
- Approximate total content: [02:06:50.269]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[09:12:48] PHASE ONE, Encoding
- [09:12:48] Extracting A/V streams [VID_00000]
- [09:23:08] Reencoding: VID_00000 (1 of 1)
- [09:23:08] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 182,464 frames
- Bitrate: 18,055 Kbs
- [09:23:08] Reencoding: VID_00000, Pass 1 of 1
- [10:57:28] Video Encode complete
- [10:57:29] Reencoding audio tracks (if req'd)
- [10:57:29] Multiplexing M2TS
[11:02:37]PHASE ONE complete
[11:02:37]PHASE TWO - Rebuild Started
- [11:02:37] Rebuilding BD file Structure
[11:12:26] - Encode and Rebuild complete
[11:12:26]JOB: STARTREK11D1AC finished.

Dirk Diggler

18th July 2011, 02:24

[08:52:25] BD Rebuilder v0.38.04 (beta)
- Source: DRIVE_ANGRY
- Input BD size: 25.22 GB
- Approximate total content: [01:44:31.264]
- Target BD size: 23.44 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[08:52:32] PHASE ONE, Encoding
- [08:52:32] Extracting A/V streams [VID_00102]
- Extracting A/V to M2TS [VID_00102]
- Extracting A/V from M2TS [VID_00102]
- [08:54:13] Reencoding: VID_00102 (1 of 47)
- [08:54:13] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,646 frames
- Bitrate: 25,166 Kbs
- [08:54:13] Reencoding: VID_00102, Pass 1 of 1
- [08:54:57] Video Encode complete
- [08:54:57] Reencoding audio tracks (if req'd)
- [08:54:57] Multiplexing M2TS
- [08:54:58] Extracting A/V streams [VID_00103]
- [08:55:05] Reencoding: VID_00103 (2 of 47)
- [08:55:05] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 7,117 frames
- Bitrate: 25,467 Kbs
- [08:55:05] Reencoding: VID_00103, Pass 1 of 1
- [08:58:04] Video Encode complete
- [08:58:04] Multiplexing M2TS
- [08:58:08] Extracting A/V streams [VID_00106]
- [08:58:10] Reencoding: VID_00106 (3 of 47)
- [08:58:10] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,722 frames
- Bitrate: 25,106 Kbs
- [08:58:10] Reencoding: VID_00106, Pass 1 of 1
- [08:58:53] Video Encode complete
- [08:58:53] Multiplexing M2TS
- [08:58:55] Extracting A/V streams [VID_00107]
- [08:58:58] Reencoding: VID_00107 (4 of 47)
- [08:58:58] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 3,343 frames
- Bitrate: 25,414 Kbs
- [08:58:58] Reencoding: VID_00107, Pass 1 of 1
- [09:00:23] Video Encode complete
- [09:00:23] Multiplexing M2TS
- [09:00:25] Extracting A/V streams [VID_00108]
- [09:00:27] Reencoding: VID_00108 (5 of 47)
- [09:00:27] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,533 frames
- Bitrate: 25,271 Kbs
- [09:00:27] Reencoding: VID_00108, Pass 1 of 1
- [09:01:06] Video Encode complete
- [09:01:06] Multiplexing M2TS
- [09:01:07] Extracting A/V streams [VID_00109]
- [09:01:10] Reencoding: VID_00109 (6 of 47)
- [09:01:10] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,627 frames
- Bitrate: 24,935 Kbs
- [09:01:10] Reencoding: VID_00109, Pass 1 of 1
- [09:02:16] Video Encode complete
- [09:02:16] Multiplexing M2TS
- [09:02:18] Extracting A/V streams [VID_00110]
- [09:02:19] Reencoding: VID_00110 (7 of 47)
- [09:02:19] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,078 frames
- Bitrate: 25,155 Kbs
- [09:02:19] Reencoding: VID_00110, Pass 1 of 1
- [09:02:48] Video Encode complete
- [09:02:48] Multiplexing M2TS
- [09:02:48] Extracting A/V streams [VID_00111]
- [09:02:50] Reencoding: VID_00111 (8 of 47)
- [09:02:50] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,804 frames
- Bitrate: 25,467 Kbs
- [09:02:50] Reencoding: VID_00111, Pass 1 of 1
- [09:03:36] Video Encode complete
- [09:03:36] Multiplexing M2TS
- [09:03:38] Extracting A/V streams [VID_00112]
- [09:03:39] Reencoding: VID_00112 (9 of 47)
- [09:03:39] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 744 frames
- Bitrate: 25,501 Kbs
- [09:03:39] Reencoding: VID_00112, Pass 1 of 1
- [09:03:58] Video Encode complete
- [09:03:58] Multiplexing M2TS
- [09:03:58] Extracting A/V streams [VID_00113]
- [09:04:00] Reencoding: VID_00113 (10 of 47)
- [09:04:00] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,710 frames
- Bitrate: 25,255 Kbs
- [09:04:00] Reencoding: VID_00113, Pass 1 of 1
- [09:04:43] Video Encode complete
- [09:04:43] Multiplexing M2TS
- [09:04:45] Extracting A/V streams [VID_00114]
- [09:04:46] Reencoding: VID_00114 (11 of 47)
- [09:04:46] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,164 frames
- Bitrate: 25,063 Kbs
- [09:04:46] Reencoding: VID_00114, Pass 1 of 1
- [09:05:16] Video Encode complete
- [09:05:16] Multiplexing M2TS
- [09:05:17] Extracting A/V streams [VID_00115]
- [09:05:19] Reencoding: VID_00115 (12 of 47)
- [09:05:19] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,952 frames
- Bitrate: 25,570 Kbs
- [09:05:19] Reencoding: VID_00115, Pass 1 of 1
- [09:06:08] Video Encode complete
- [09:06:09] Multiplexing M2TS
- [09:06:10] Extracting A/V streams [VID_00116]
- [09:06:12] Reencoding: VID_00116 (13 of 47)
- [09:06:12] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,831 frames
- Bitrate: 25,271 Kbs
- [09:06:12] Reencoding: VID_00116, Pass 1 of 1
- [09:07:01] Video Encode complete
- [09:07:01] Multiplexing M2TS
- [09:07:02] Extracting A/V streams [VID_00117]
- [09:07:11] Reencoding: VID_00117 (14 of 47)
- [09:07:11] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 9,528 frames
- Bitrate: 25,449 Kbs
- [09:07:11] Reencoding: VID_00117, Pass 1 of 1
- [09:11:11] Video Encode complete
- [09:11:12] Multiplexing M2TS
- [09:11:16] Extracting A/V streams [VID_00120]
- [09:11:19] Reencoding: VID_00120 (15 of 47)
- [09:11:19] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,777 frames
- Bitrate: 25,203 Kbs
- [09:11:19] Reencoding: VID_00120, Pass 1 of 1
- [09:12:28] Video Encode complete
- [09:12:28] Multiplexing M2TS
- [09:12:30] Extracting A/V streams [VID_00121]
- [09:12:32] Reencoding: VID_00121 (16 of 47)
- [09:12:32] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,194 frames
- Bitrate: 25,283 Kbs
- [09:12:32] Reencoding: VID_00121, Pass 1 of 1
- [09:13:02] Video Encode complete
- [09:13:02] Multiplexing M2TS
- [09:13:03] Extracting A/V streams [VID_00122]
- [09:13:05] Reencoding: VID_00122 (17 of 47)
- [09:13:05] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,618 frames
- Bitrate: 24,805 Kbs
- [09:13:05] Reencoding: VID_00122, Pass 1 of 1
- [09:13:45] Video Encode complete
- [09:13:45] Multiplexing M2TS
- [09:13:46] Extracting A/V streams [VID_00123]
- [09:13:49] Reencoding: VID_00123 (18 of 47)
- [09:13:49] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,945 frames
- Bitrate: 25,811 Kbs
- [09:13:49] Reencoding: VID_00123, Pass 1 of 1
- [09:14:37] Video Encode complete
- [09:14:37] Multiplexing M2TS
- [09:14:39] Extracting A/V streams [VID_00124]
- [09:14:42] Reencoding: VID_00124 (19 of 47)
- [09:14:42] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 3,123 frames
- Bitrate: 25,319 Kbs
- [09:14:42] Reencoding: VID_00124, Pass 1 of 1
- [09:16:00] Video Encode complete
- [09:16:00] Multiplexing M2TS
- [09:16:02] Extracting A/V streams [VID_00125]
- [09:16:05] Reencoding: VID_00125 (20 of 47)
- [09:16:05] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,816 frames
- Bitrate: 25,683 Kbs
- [09:16:05] Reencoding: VID_00125, Pass 1 of 1
- [09:17:16] Video Encode complete
- [09:17:16] Multiplexing M2TS
- [09:17:18] Extracting A/V streams [VID_00126]
- [09:17:20] Reencoding: VID_00126 (21 of 47)
- [09:17:20] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,717 frames
- Bitrate: 25,305 Kbs
- [09:17:20] Reencoding: VID_00126, Pass 1 of 1
- [09:18:04] Video Encode complete
- [09:18:04] Multiplexing M2TS
- [09:18:05] Extracting A/V streams [VID_00127]
- [09:18:07] Reencoding: VID_00127 (22 of 47)
- [09:18:08] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,039 frames
- Bitrate: 25,683 Kbs
- [09:18:08] Reencoding: VID_00127, Pass 1 of 1
- [09:18:58] Video Encode complete
- [09:18:58] Multiplexing M2TS
- [09:19:00] Extracting A/V streams [VID_00128]
- [09:19:01] Reencoding: VID_00128 (23 of 47)
- [09:19:01] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 852 frames
- Bitrate: 25,339 Kbs
- [09:19:01] Reencoding: VID_00128, Pass 1 of 1
- [09:19:23] Video Encode complete
- [09:19:23] Multiplexing M2TS
- [09:19:23] Extracting A/V streams [VID_00129]
- [09:19:25] Reencoding: VID_00129 (24 of 47)
- [09:19:25] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,177 frames
- Bitrate: 25,369 Kbs
- [09:19:25] Reencoding: VID_00129, Pass 1 of 1
- [09:19:54] Video Encode complete
- [09:19:54] Multiplexing M2TS
- [09:19:55] Extracting A/V streams [VID_00130]
- [09:20:02] Reencoding: VID_00130 (25 of 47)
- [09:20:02] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 7,522 frames
- Bitrate: 25,116 Kbs
- [09:20:02] Reencoding: VID_00130, Pass 1 of 1
- [09:23:11] Video Encode complete
- [09:23:11] Multiplexing M2TS
- [09:23:15] Extracting A/V streams [VID_00131]
- [09:23:17] Reencoding: VID_00131 (26 of 47)
- [09:23:17] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,011 frames
- Bitrate: 25,079 Kbs
- [09:23:17] Reencoding: VID_00131, Pass 1 of 1
- [09:24:08] Video Encode complete
- [09:24:08] Multiplexing M2TS
- [09:24:10] Extracting A/V streams [VID_00132]
- [09:24:11] Reencoding: VID_00132 (27 of 47)
- [09:24:11] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 773 frames
- Bitrate: 24,911 Kbs
- [09:24:11] Reencoding: VID_00132, Pass 1 of 1
- [09:24:30] Video Encode complete
- [09:24:30] Multiplexing M2TS
- [09:24:31] Extracting A/V streams [VID_00133]
- [09:24:33] Reencoding: VID_00133 (28 of 47)
- [09:24:33] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,442 frames
- Bitrate: 25,247 Kbs
- [09:24:33] Reencoding: VID_00133, Pass 1 of 1
- [09:25:09] Video Encode complete
- [09:25:09] Multiplexing M2TS
- [09:25:11] Extracting A/V streams [VID_00134]
- [09:25:14] Reencoding: VID_00134 (29 of 47)
- [09:25:14] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 3,148 frames
- Bitrate: 25,662 Kbs
- [09:25:14] Reencoding: VID_00134, Pass 1 of 1
- [09:26:33] Video Encode complete
- [09:26:33] Multiplexing M2TS
- [09:26:35] Extracting A/V streams [VID_00135]
- [09:26:36] Reencoding: VID_00135 (30 of 47)
- [09:26:36] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 987 frames
- Bitrate: 25,494 Kbs
- [09:26:36] Reencoding: VID_00135, Pass 1 of 1
- [09:27:01] Video Encode complete
- [09:27:01] Multiplexing M2TS
- [09:27:02] Extracting A/V streams [VID_00136]
- [09:27:06] Reencoding: VID_00136 (31 of 47)
- [09:27:06] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 4,972 frames
- Bitrate: 25,810 Kbs
- [09:27:06] Reencoding: VID_00136, Pass 1 of 1
- [09:29:10] Video Encode complete
- [09:29:10] Multiplexing M2TS
- [09:29:12] Extracting A/V streams [VID_00137]
- [09:29:16] Reencoding: VID_00137 (32 of 47)
- [09:29:16] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 3,579 frames
- Bitrate: 25,977 Kbs
- [09:29:16] Reencoding: VID_00137, Pass 1 of 1
- [09:30:43] Video Encode complete
- [09:30:44] Multiplexing M2TS
- [09:30:45] Extracting A/V streams [VID_00138]
- [09:30:51] Reencoding: VID_00138 (33 of 47)
- [09:30:51] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 6,359 frames
- Bitrate: 25,540 Kbs
- [09:30:51] Reencoding: VID_00138, Pass 1 of 1
- [09:33:32] Video Encode complete
- [09:33:32] Multiplexing M2TS
- [09:33:35] Extracting A/V streams [VID_00139]
- [09:33:37] Reencoding: VID_00139 (34 of 47)
- [09:33:37] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,687 frames
- Bitrate: 25,043 Kbs

Dirk Diggler

18th July 2011, 02:25

- [09:33:37] Reencoding: VID_00139, Pass 1 of 1
- [09:34:20] Video Encode complete
- [09:34:20] Multiplexing M2TS
- [09:34:22] Extracting A/V streams [VID_00140]
- [09:34:26] Reencoding: VID_00140 (35 of 47)
- [09:34:26] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 4,125 frames
- Bitrate: 24,991 Kbs
- [09:34:26] Reencoding: VID_00140, Pass 1 of 1
- [09:36:09] Video Encode complete
- [09:36:09] Multiplexing M2TS
- [09:36:12] Extracting A/V streams [VID_00141]
- [09:36:13] Reencoding: VID_00141 (36 of 47)
- [09:36:13] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,116 frames
- Bitrate: 24,510 Kbs
- [09:36:13] Reencoding: VID_00141, Pass 1 of 1
- [09:36:41] Video Encode complete
- [09:36:41] Multiplexing M2TS
- [09:36:42] Extracting A/V streams [VID_00142]
- [09:36:51] Reencoding: VID_00142 (37 of 47)
- [09:36:51] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 10,300 frames
- Bitrate: 25,346 Kbs
- [09:36:52] Reencoding: VID_00142, Pass 1 of 1
- [09:41:12] Video Encode complete
- [09:41:12] Multiplexing M2TS
- [09:41:17] Extracting A/V streams [VID_00143]
- [09:41:39] Reencoding: VID_00143 (38 of 47)
- [09:41:39] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 21,609 frames
- Bitrate: 25,408 Kbs
- [09:41:39] Reencoding: VID_00143, Pass 1 of 1
- [09:50:39] Video Encode complete
- [09:50:39] Multiplexing M2TS
- [09:50:47] Extracting A/V streams [VID_00152]
- [09:50:50] Reencoding: VID_00152 (39 of 47)
- [09:50:50] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 3,332 frames
- Bitrate: 24,959 Kbs
- [09:50:50] Reencoding: VID_00152, Pass 1 of 1
- [09:52:14] Video Encode complete
- [09:52:14] Multiplexing M2TS
- [09:52:16] Extracting A/V streams [VID_00153]
- [09:52:16] Reencoding: VID_00153 (40 of 47)
- [09:52:16] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 630 frames
- Bitrate: 25,152 Kbs
- [09:52:16] Reencoding: VID_00153, Pass 1 of 1
- [09:52:32] Video Encode complete
- [09:52:32] Multiplexing M2TS
- [09:52:33] Extracting A/V streams [VID_00154]
- [09:52:34] Reencoding: VID_00154 (41 of 47)
- [09:52:34] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,267 frames
- Bitrate: 25,281 Kbs
- [09:52:34] Reencoding: VID_00154, Pass 1 of 1
- [09:53:07] Video Encode complete
- [09:53:07] Multiplexing M2TS
- [09:53:08] Extracting A/V streams [VID_00155]
- [09:53:10] Reencoding: VID_00155 (42 of 47)
- [09:53:10] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,069 frames
- Bitrate: 25,188 Kbs
- [09:53:10] Reencoding: VID_00155, Pass 1 of 1
- [09:54:03] Video Encode complete
- [09:54:03] Multiplexing M2TS
- [09:54:04] Extracting A/V streams [VID_00156]
- [09:54:05] Reencoding: VID_00156 (43 of 47)
- [09:54:05] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,162 frames
- Bitrate: 25,027 Kbs
- [09:54:05] Reencoding: VID_00156, Pass 1 of 1
- [09:54:35] Video Encode complete
- [09:54:35] Multiplexing M2TS
- [09:54:35] Extracting A/V streams [VID_00157]
- [09:54:38] Reencoding: VID_00157 (44 of 47)
- [09:54:38] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 2,075 frames
- Bitrate: 25,278 Kbs
- [09:54:38] Reencoding: VID_00157, Pass 1 of 1
- [09:55:30] Video Encode complete
- [09:55:30] Multiplexing M2TS
- [09:55:31] Extracting A/V streams [VID_00158]
- [09:55:33] Reencoding: VID_00158 (45 of 47)
- [09:55:33] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,624 frames
- Bitrate: 26,305 Kbs
- [09:55:33] Reencoding: VID_00158, Pass 1 of 1
- [09:56:15] Video Encode complete
- [09:56:15] Multiplexing M2TS
- [09:56:16] Extracting A/V streams [VID_00159]
- [09:56:18] Reencoding: VID_00159 (46 of 47)
- [09:56:18] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 1,575 frames
- Bitrate: 26,385 Kbs
- [09:56:18] Reencoding: VID_00159, Pass 1 of 1
- [09:56:56] Video Encode complete
- [09:56:56] Multiplexing M2TS
- [09:56:57] Extracting A/V streams [VID_00160]
- [09:57:06] Reencoding: VID_00160 (47 of 47)
- [09:57:06] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 9,939 frames
- Bitrate: 25,445 Kbs
- [09:57:06] Reencoding: VID_00160, Pass 1 of 1
- [10:00:49] Video Encode complete
- [10:00:49] Multiplexing M2TS
[10:00:53]PHASE ONE complete
[10:00:53]PHASE TWO - Rebuild Started
- [10:00:53] Rebuilding BD file Structure
[10:03:32] - Encode and Rebuild complete
- WORKFILES folder removed.
[10:03:33]JOB: DRIVE_ANGRY finished.

hoju3508

21st July 2011, 21:42

but the mux and rebuild was much slower -- taking over 35 minutes on my system compared to 11 on the SSD.

Is it possible to mux and rebuild on a different drive than the WORK directory? Shouldn't that cut off some time as well?

jdobbs

21st July 2011, 22:31

Is it possible to mux and rebuild on a different drive than the WORK directory? Shouldn't that cut off some time as well? I've answered that question many times. Yes, it's possible and I've tested it. But it makes almost no measurable difference in time.

Ch3vr0n

21st July 2011, 23:30

it be nice though if you were to add that 3rd path option for the "final rebuild" ;)

RobertM

22nd July 2011, 00:59

I have done a considerable amount of poking around trying to figure out ways to reduce my BD-Rebuilder times.

My starting point was a decent processor (i7-950 with 6GB RAM) and a 1TB drive used for both source and target. I quickly noticed the HDD bottleneck and purchased a 2TB drive so that I could put rips and rebuilds on separate drives. Rips are now on the 2Tb drive and the 1TB is the BD-Rebuilder working drive. I noticed, in particular, that "extraction" times were dramatically reduced.

This is the typical performance that I was seeing, using a Star Trek(2009) BD for testing purposes

i7-950 @ 3.2GHz
Begin Duration
Start 11:48:15 0:00:05
Extract 11:48:20 0:05:12
Re-encode 11:53:32 1:25:27
Mux 13:18:59 0:08:55
Rebuild 13:27:54 0:12:10
End 13:40:04
Total 1:51:49

I was not really unhappy with the extraction and mux/rebuild times. The main problem seemed to be the re-encode times. I noticed that the processor was only running at about 45% during the re-encoding (and it was only at about 60 C on air cooling), and almost idle during the extract/mux/rebuild processes. I also noticed that if I switch to higher quality (slower) re-encodes, or 2-pass encodes, then the processor becomes more heavily loaded.

After reading the discussions in this thread I decided to see what some moderate o/c would do:

i7-950 @ 3.6GHz
Begin Duration
Start 23:44:45 0:00:02
Extract 23:44:47 0:05:30
Re-encode 23:50:17 1:16:45
Mux 25:07:02 0:09:19
Rebuild 25:16:21 0:09:39
End 25:26:00
Total 1:41:15

The processor is still running <50% load during the re-encode, but it is running faster, saving me 10 minutes. I tried running the 0/C up to 4GHz, but I had stability issues running the benchmark tests. I'm pretty sure that it would still handle the re-encode, however, since the core loading is so low.

I then thought that, since the working drive and target drive are the same, speeding up that drive might yield benefits during mux/rebuild. So I installed another 1TB drive, putting it in RAID-0 configuration with the original, identical, 1TB drive. Sure enough:

i7-950 @ 3.6GHz, RAID-0
Begin Duration
Start 16:02:46 0:00:06
Extract 16:02:52 0:05:08
Re-encode 16:08:00 1:14:34
Mux 17:22:34 0:04:15
Rebuild 17:26:49 0:07:11
End 17:34:00
Total 1:31:14

So, it would seem that spending $90 on an additional 1TB drive has yielded considerable speed improvements at the back end of the BD-Rebuilder process. It should also speed up overall performance since most of my applications are located on that same drive, while my OS is on a different SSD drive. The only downside I can think of is the inherent doubling of the drive failure rate, which doesn't scare me TOO much since I backup regularly.

If I tweaked the O/C up to 3.8GHz then I bet I would fall comfortably below the 1.1/2hr mark. And if I could get the processor to actually pull its weight then I'd really be getting somewhere!

For reference, here's a copy of my latest rebuild log (for the last trial above):
[16:02:46] BD Rebuilder v0.38.04 (beta)
- Source: STARTREK11D1AC
- Input BD size: 36.85 GB
- Approximate total content: [02:06:50.269]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[16:02:52] PHASE ONE, Encoding
- [16:02:52] Extracting A/V streams [VID_00000]
- [16:08:00] Reencoding: VID_00000 (1 of 1)
- [16:08:00] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 182,464 frames
- Bitrate: 23,100 Kbs
- [16:08:00] Reencoding: VID_00000, Pass 1 of 1
- [17:22:34] Video Encode complete
- [17:22:34] Reencoding audio tracks (if req'd)
- [17:22:34] Multiplexing M2TS
[17:26:49]PHASE ONE complete
[17:26:49]PHASE TWO - Rebuild Started
- [17:26:49] Rebuilding BD file Structure
[17:34:00] - Encode and Rebuild complete
[17:34:00]JOB: STARTREK11D1AC finished.

Stereodude

22nd July 2011, 01:32

And if I could get the processor to actually pull its weight then I'd really be getting somewhere!The settings being used for x264 are causing it to be limited by one of the slower threads. As such it doesn't saturate the CPU. Using "slower" encoding settings may not really run much slower (timewise) since the CPU will be more saturated and offer better quality.

Ch3vr0n

22nd July 2011, 01:37

@RobertM since you're overclocking, be VERY carefull with that. x264 is an extremely powerfull but sensitive and system demanding program . The slightest flaw in your overclocking can cause x264 to become unstable and fail your encoding.

Dirk Diggler

22nd July 2011, 23:18

Phenom X6@4188

-----------------------
[08:12:40] BD Rebuilder v0.38.04 (beta)
- Source: ADJUSTMENT_BUREAU
- Input BD size: 28.55 GB
- Approximate total content: [01:45:50.344]
- Target BD size: 24.05 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Quality: Good (Very Fast), Two Pass
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[08:12:40] PHASE ONE, Encoding
- [08:12:40] Extracting A/V streams [VID_00800]
- [08:15:18] Reencoding: VID_00800 (1 of 1)
- [08:15:18] Collecting video information
- Source Video: VC-1, 1920x1080
- Rate/Length: 23.976fps, 152,256 frames
- Bitrate: 25,874 Kbs
- [08:15:18] Reencoding: VID_00800, Pass 1 of 2
- [09:13:59] Reencoding: VID_00800, Pass 2 of 2
- [10:21:58] Video Encode complete
- [10:21:58] Reencoding audio tracks (if req'd)
- [10:21:58] Multiplexing M2TS
[10:23:55]PHASE ONE complete
[10:23:55]PHASE TWO - Rebuild Started
- [10:23:55] Rebuilding BD file Structure
[10:26:34] - Encode and Rebuild complete
- WORKFILES folder removed.
- Computer SHUTDOWN initiated.

RobertM

23rd July 2011, 04:00

RobertM

23rd July 2011, 15:06

Well, I didn't expect THIS!

As stated previously, I've been playing around with some moderate O/C on my i7-950. While running IntelBurnTest I noticed that the processor was running at around 50%, which is the same as I was seeing during my x264 encoding runs. I was expecting this "torture test" to be close to 100%, and it seemed curious that, in both cases, the max loading was about exactly half of what I was expecting. The 1:2 performance hit got me thinking about hyper-threading so I ran another test with HT turned off to see what would happen. CPU load went immediately up to 100% during the test, the GFlops result jumped significantly, while temps remained the same.

I was running at 3.8GHz, and it was feeling pretty stable, so I ran my test encode of Star Trek 2009 to see what would happen:

i7-950 @ 3.8GHz, RAID-0
Begin Duration
Start 23:09:47 0:00:09
Extract 23:09:56 0:05:08
Re-encode 23:15:04 1:02:29
Mux 24:17:33 0:04:13
Rebuild 24:21:46 0:04:42
End 24:26:28
Total 1:16:41

Extraction and muxing times have remained basically unchanged, but re-encoding and rebuilding have shown significant time reductions. I was expecting HT to offer a slight advantage in x264 encoding, not a significant penalty. Seems odd to me.

Here's my log:

[23:09:47] BD Rebuilder v0.38.04 (beta)
- Source: STARTREK11D1AC
- Input BD size: 36.85 GB
- Approximate total content: [02:06:50.269]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=0 Kbs=640
[23:09:56] PHASE ONE, Encoding
- [23:09:56] Extracting A/V streams [VID_00000]
- [23:15:04] Reencoding: VID_00000 (1 of 1)
- [23:15:04] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 182,464 frames
- Bitrate: 23,100 Kbs
- [23:15:04] Reencoding: VID_00000, Pass 1 of 1
- [00:17:33] Video Encode complete
- [00:17:33] Reencoding audio tracks (if req'd)
- [00:17:33] Multiplexing M2TS
[00:21:46]PHASE ONE complete
[00:21:46]PHASE TWO - Rebuild Started
- [00:21:46] Rebuilding BD file Structure
[00:26:28] - Encode and Rebuild complete
[00:26:28]JOB: STARTREK11D1AC finished.

Dirk Diggler

23rd July 2011, 17:56

- [08:12:40] Extracting A/V streams [VID_00800]
- [08:15:18] Reencoding: VID_00800 (1 of 1)

- [10:21:58] Multiplexing M2TS
[10:23:55]PHASE ONE complete
[10:23:55]PHASE TWO - Rebuild Started
- [10:23:55] Rebuilding BD file Structure
[10:26:34] - Encode and Rebuild complete

0:02:38 to extract and 0:04:36 to mux/rebuild a BD-25??? What kind of drives are you running there bud?

5 1TB WD Black FAEX drives in a Raid0. I want to run some test using different drives for reading and writing like you have done just keep forgetting LOL. My SSD for the original to the array for the encode but until we can separate the WORK Folder from the final encode there will always be an issue of reading and writing from the same drive

RobertM

23rd July 2011, 18:00

Is your working drive the same as your source drive?

Dirk Diggler

23rd July 2011, 18:12

Is your working drive the same as your source drive?

Yes currently everything is read and written to and from the array

Dirk Diggler

23rd July 2011, 18:55

@RobertM do you have "The Adjustment Bureau" I would like to see encoding times for this release?

Dirk Diggler

23rd July 2011, 19:02

RobertM

23rd July 2011, 19:17

@RobertM do you have "The Adjustment Bureau"

No, sorry I don't. I get my BD's from a local video store after they have rented them out for a while, so it'll be a little while before they put this movie in their sales bin.

Dirk Diggler

23rd July 2011, 19:20

No, sorry I don't. I get my BD's from a local video store after they have rented them out for a while, so it'll be a little while before they put this movie in their sales bin.

I lurk so Im patient..... Whenever you get it I would be interested in encode times

jdobbs

23rd July 2011, 19:41

@RobertM

Interesting...:goodpost:

Dark Shikari

23rd July 2011, 19:55

Hyperthreading may be markerting fluff for certain workloads I think this is the reason AMD has answered intel with more physical cores instead of buying into the whole hyperthreading deal.HT typically gives a ~25% benefit to x264 encoding speed. It's hardly fluff.

Dirk Diggler

23rd July 2011, 21:34

HT typically gives a ~25% benefit to x264 encoding speed. It's hardly fluff.

Gotcha, Im not Intel savvy but can you explain his results WITHOUT HT? I guess we would need more people to chime in and setup some type of benchmark thread so we just didnt have a one off case I am not sure why anyone hasnt set it up yet including myself

RobertM

23rd July 2011, 22:30

HT typically gives a ~25% benefit to x264 encoding speed.

I expected some kind of benefit like that, but I'm seeing something completely different. This morning I re-ran the StarTrek encode job, making no changes whatsoever to the BD-Rebuilder settings, but turning HT back on. Results:

[10:56:08] PHASE ONE, Encoding
- [10:56:08] Extracting A/V streams [VID_00000]
- [11:01:17] Reencoding: VID_00000 (1 of 1)
- [11:01:17] Collecting video information
- Source Video: MPEG-4 (AVC), 1920x1080
- Rate/Length: 23.976fps, 182,464 frames
- Bitrate: 23,100 Kbs
- [11:01:17] Reencoding: VID_00000, Pass 1 of 1
- [12:11:30] Video Encode complete

So the only variable that changed was turning on HT in BIOS, and the encode time lengthened from 1:02:29 to 1:10:13. Note that the core loading with HT OFF was near 100% while it dropped to around 50% with HT ON. With the core loading cut in half I might expect the processing time to have increased by a significant amount (perhaps double?) but the actual processing time increased only 12%. So perhaps HT is actually doing a good job of helping out a lightly loaded processor. Maybe if I had a process that would actually load up the core fully, with HT enabled, I might see that HT is offering significant speed gains. I dunno; I'm just guessing here.

I have noticed in the past that my core loading DOES increase to close to 100%, with HT enabled, when running the 2nd pass of a 2-pass encoding job. In that situation, perhaps turning off HT would result in a performance drop. I'll explore that possibility overnight and report back.

RobertM

24th July 2011, 14:13

At this point I must say that Dark Shikari is generally correct (surprise?) about the benefits of HT.

I ran a series of small trials using a 5 min section from the start of Star Trek 2009. I first verified that HT was off and then I re-encoded the test clip using BD-Rebuilder, selecting "good", "better", "high" and "highest" quality settings, all 1-pass. Then I enabled HT in my BIOS and I ran the batch again.

Results:

Duration HT OFF HT ON HT Benefit
Good 0:02:26 0:02:47 -13%
Better 0:06:24 0:05:05 26%
High 0:10:47 0:08:28 27%
Highest 0:15:54 0:12:34 27%

CPU load HT OFF HT ON
Good 100% 50%
Better 100% 100%
High 100% 100%
Highest 100% 100%

CPU temp HT OFF HT ON
Good 68c 68c
Better 68c 75c
High 68c 75c
Highest 68c 75c

So I am seeing a definite benefit using HT in most cases, almost exactly to the extent that Dark Shikari said I would.

But there is an actual downgrade in speed when I encode using the "good" quality setting, and the same thing happens when I run the IntelBurnTest "torture" test. Core loading goes down by half, and core temps drop. Maybe it is just MY processor? But there is a lot of discussion about HT on the net, with some people saying that their games/apps seem to run slower with HT, and others disagreeing. That sort of Jekyll/Hyde behaviour seems to be in accord with what I found above; it helps sometimes, and hinders at other times.

Anyone else out there with an i7 who could confirm this on their machine?

And it leaves me wondering what to do? I'm perfectly happy with "good" quality encodes, and x264 encoding is, by far, the most intense thing that I, personally, do on this computer. So I'm leaning towards just leaving HT off, and enjoying faster encodes and lower temps. Who cares if the kids games run a little slower (I'm pretty sure they aren't reading this forum ;) )?

At least I know what to do if I want to start running higher quality encodes in the future.

Stereodude

30th July 2011, 22:23

Anyone else out there with an i7 who could confirm this on their machine?FWIW, I did similar testing with x264 (not using BD Rebuilder) and posted the results here (http://forum.doom9.org/showthread.php?p=1516774). I didn't find any slowdown from HT.

massiah

5th August 2011, 05:44

I have been using BD-RB for awhile. I love the program!
My last system setup was Win7 Ultimate(64Bit)8GB RAM, Intel Q9650 , Multiple 7200RPM HDDs and My OS Drive was a VRaptor 300GB.
I was typically doing BD25 Rebuilds in about 3-5 hours.

I just finished my new system build last week and have been reeeeeaaaaallll slow getting programs installed on it. LOTS OF WORK!!!!
This build has Win7 Ultimate(64Bit), 24GB RAM, Intel Core i7980X, 240GB OCZ Vertex 3 OS Drive, Samsung 256GB SSD(SATA II), Kingston 128GB SSD(SATA II) 2TB USB3.0 External HDD.
I keep the files on a 1TB WD 7200RPM HDD and process them on the 2TB External HDD.

I just did The Town with this setup and finished the BD25 conversion process before burn in less than 20 minutes!!!!!

[00:10:55] BD Rebuilder v0.38.08 (beta)
- Source: THE_TOWN
- Input BD size: 18.37 GB
- Approximate total content: [02:30:17.049]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[00:10:55] PHASE ONE, Encoding
- [00:10:55] Extracting A/V streams [VID_00008]
- [00:20:58] Reencoding: VID_00008 (1 of 1)
- [00:20:58] Collecting video information
- [00:20:58] Keeping original video (no reencode)
- [00:20:58] Reencoding audio tracks (if req'd)
- [00:20:58] Multiplexing M2TS
[00:27:18]PHASE ONE complete
[00:27:18]PHASE TWO - Rebuild Started
- [00:27:18] Rebuilding BD file Structure
[00:30:46] - Encode and Rebuild complete
[00:30:46] BURNING BD structure to disc

hoju3508

5th August 2011, 08:24

@massiah, it wasn't necessary to re-encode since your Input BD size is less then the Target BD size :).

Do you have either Star Trek 11 or The Adjustment Bureau since that's what a lot of people are using?

massiah

6th August 2011, 04:12

No,I do not have The Adjustment Bureau or Star Trek 11.

I know the size is smaller than a BD25 encode but, I was going for the rebuilding the Extended Version of the movie. As I stated before, this is a new build. So, I didn't have a couple of helpful tools on this machine, yet.

I will try something more standard for sure in the very near future and post the results!

massiah

7th August 2011, 03:54

OK...
So I did the movie HEAT tonight(Movie Only). Reason being, I didn't want the main movie to be degraded by two hours worth of extras.

The file size was 34GB so, obviously it was going to take a little longer than my last encode. I'm not too pleased with the results because, last night's encode used up over 16GB of my 24GB RAM while encoding. Tonight? It barely reached 3GB. I didn't change any settings other than selecting Movie-Only.

-----------------------
[20:51:10] BD Rebuilder v0.38.08 (beta)
- Source: HEAT
- Input BD size: 34.86 GB
- Approximate total content: [02:50:27.383]
- Target BD size: 22.95 GB
- Windows Version: 6.1 [7601]
- MOVIE-ONLY mode enabled
- Auto Quality: Good (Very Fast), ABR
- Audio Settings: AC3=0 DTS=0 HD=1 Kbs=640
[20:51:10] PHASE ONE, Encoding
- [20:51:10] Extracting A/V streams [VID_00003]
- [21:15:52] Reencoding: VID_00003 (1 of 1)
- [21:15:52] Collecting video information
- Source Video: VC-1, 1920x1080
- Rate/Length: 23.976fps, 245,212 frames
- Bitrate: 14,978 Kbs
- [21:15:52] Reencoding: VID_00003, Pass 1 of 1
- [22:15:48] Video Encode complete
- [22:15:48] Reencoding audio tracks (if req'd)
- [22:15:48] Multiplexing M2TS
[22:18:37]PHASE ONE complete
[22:18:37]PHASE TWO - Rebuild Started
- [22:18:37] Rebuilding BD file Structure
[22:27:44] - Encode and Rebuild complete
[22:27:44] BURNING BD structure to disc

I'll keep testing to see what I get.

jdobbs

7th August 2011, 14:20

I've had a couple of other reports similar to yours. I even saw it once myself. Somehow the exact same process uses ridiculous amount of memory, but when stopped and restarted it doesn't. I haven't been able to put my finger on the when-where-why-and-how of it because it is very rare and is unpredictable. Since it is highly unlikely that the same app would act differently on the same source with the same settings, I suspect it is a Windows 7 issue (likely related to disc caching).