• AnyStream is having some DRM issues currently, Netflix is not available in HD for the time being.
    Situations like this will always happen with AnyStream: streaming providers are continuously improving their countermeasures while we try to catch up, it's an ongoing cat-and-mouse game. Please be patient and don't flood our support or forum with requests, we are working on it 24/7 to get it resolved. Thank you.

CloneBD 1.1.5.0 not making full use of 10 core XEON

Qwikie

Active Member
Thread Starter
Joined
Oct 2, 2016
Messages
29
Likes
2
The first 3 bars are 100% and CloneBD only uses 60% (6 cores?) of my cpu.

Do I have to wait for the 10 core Kalby Lake is written for to use my full 10 core XEON.

Also, this only seems to be happening in the last few versions.
 
hover with your mouse over the bars. The top 3 are supposed to be at 100%
 
still only using about 60% of cpu (5% other) and says it can go faster. Seems like it is not making full use of all cpu cores. So why again are top 3 bars at 100% if there is cpu % free? If queues are full then cpu is not the limiting factor so why can't the queues be made bigger?

Still do not understand.

Also thanks for answering my Furious 7 audio watermark questions.
 
If you're using Intel quick sync it's normal. The CPU's GPU component does the bulk of the work then. Same thing if I use my GTX 1080 for hardware encoding. Top 3 bars are at 100% and CPU load at like 10-20, which is normal cause my GPU is doing all the work.

Verstuurd vanaf mijn Nexus 6P met Tapatalk
 
Thanks for the response. 10 core XEON, no INTEL Quick Sync and using all software decoding. Could you please elaborate more since I still do see why.

Thanks.
 
I'm using a 6950X 10 core and have similar issues with the software encoder. Initially, the software was crashing with more than 16 threads so I believe the software was then limited to use less than that during encoding. That said I find the software encoder to sometimes use as little as 30 % CPU utilization.

I've reported this issue in the past at least a couple times but the focus lately has been the new GPU based encoding so little has been done (IMO) to address the issues with the software encoder.
 
I have to look into this, I think the bottleneck is the software image rescaler.
 
Thanks to both of you for noticing this issue. Please do not forget us 10 core folks.

I have a GTX980, but found even being limited in cpu my cpu was faster. My last test was a month or two ago.
 
Thanks to both of you for noticing this issue. Please do not forget us 10 core folks.

I have a GTX980, but found even being limited in cpu my cpu was faster. My last test was a month or two ago.
Can you please post a CloneBD log file - we're totally in the dark regarding your settings and intents (what kind of conversion, what kind of codec etc....)
 
still only using about 60% of cpu (5% other) and says it can go faster. Seems like it is not making full use of all cpu cores. So why again are top 3 bars at 100% if there is cpu % free? If queues are full then cpu is not the limiting factor so why can't the queues be made bigger?
MOST important is: are you converting to AVC or HEVC?
The HEVC software encoder in ffmpeg is limited to a max. number of threads.

Then: the "Bars" in CloneBD: I understand, that the first 3 are maxed and the last one is not (screenshot would help a lot in getting picture of things)?
That would mean, the encoder is limiting.

Then there is another thing: do you have Cinavia removal enabled? (Downconvert audio and AnyDVD removes Cinavia).
For technical reasons, Cinavia removal needs to be serialized, so it can't use multiple cores. With many audio streams, this can be a limiting factor, too.

Regarding GTX980 being slower than your limited CPU: I find that very hard to believe. Can you give us numbers? (fps on an average AVC source Blu-ray with and without nVidia acceleration).
I have a 950 and frame rate usually is >300fps (if not limited by the source or destination medium - so using SSD for both).
 
MOST important is: are you converting to AVC or HEVC?
The HEVC software encoder in ffmpeg is limited to a max. number of threads.

Then: the "Bars" in CloneBD: I understand, that the first 3 are maxed and the last one is not (screenshot would help a lot in getting picture of things)?
That would mean, the encoder is limiting.

Then there is another thing: do you have Cinavia removal enabled? (Downconvert audio and AnyDVD removes Cinavia).
For technical reasons, Cinavia removal needs to be serialized, so it can't use multiple cores. With many audio streams, this can be a limiting factor, too.

Regarding GTX980 being slower than your limited CPU: I find that very hard to believe. Can you give us numbers? (fps on an average AVC source Blu-ray with and without nVidia acceleration).
I have a 950 and frame rate usually is >300fps (if not limited by the source or destination medium - so using SSD for both).

Did not check all your questions in prior rips. Will be getting new BRs this week and will do complete comparisons and logs then. Do not remember how to know AVC vs HEVC, but remember an occasional VC-1 that were like molasses.

I can tell you what I know for sure now.

I have the default, unchecked, setting which I believe doesn't enable Cinavia removal.

I always rip my movies to MKV with lossless audio and directly onto my HDD drives. I do have an EVO 850 SSD, but never use for rips as far as I know. When doing TV I think my ISOs are written directly to the HDD.

I prefer not to do more movie rips than necessary, e.g 4-way SSD+980 SSD+CPU HDD+980 HDD+CPU, not to mention combinations involving any temp drives or other variable combo's. Can you let me know what setting and disk combinations you would most be interested in where at least one involves the HDD as destination and if needed HDD temporary drive.

Again thank you for your continued interest in my equipment and process. Look forward to helping us learn more regarding 10 core cpu's
 
Just got 5 disc TV show. Made no compression, lossless audio, bit for bit with cpu decoding. It took 29 minutes for rip using HDD and saved cdblog file.
 
Just got 5 disc TV show. Made no compression, lossless audio, bit for bit with cpu decoding. It took 29 minutes for rip using HDD and saved cdblog file.
So, can you please post a CloneBD log file (any maybe an AnyDVD log file, too)? So we can go through your settings and try to find the bottle-neck.
 
Sorry I did not get back sooner.

Here are my logs for HAVEN Final season TV rip lossless 1-1 to ISO didn't check if cpu not fully utilized. sorry.

Am doing Furious 7 both ways cpu/gpu for MVK, AVC and lossess

EDIT > Just realized 1-1 ISO rips aren't using CPU, but do have new logs and same question regarding MKV rips. < EDIT
 

Attachments

  • Haven D1.cbdlog
    172.3 KB · Views: 0
  • AnyDVD_8.1.0.1_Info_K_HAVEN_THE_FINAL_SEASON_D1.ziplog
    695.3 KB · Views: 0
Last edited:
I just finished Furious 7 MPEG-4 AVC lossless MKV rip of 32.33GB 2:19:54 time.

EDIT : Temp and Final destinations were both on HDD. Documents on HDD. Windows and Programs, etc. on SSD

GPU was surprisingly about 3 minutes faster than CPU, but CPU still seems it could use more cores, and better queue capacity for encoder % usage. So if all 10 cores and queue capacity were better used could duration be improved 3+ minutes?

CPU Stats
Total Duration : 22:19
Data Read 32.33 GB
Data Write 30.63 GB
Avg CPU 56% top about 65%
FPS avg 210 max 346

GPU Stats
Total Duration : 19:36
Data Read 32.33 GB
Data Write 29.81 GB
Avg CPU 3%
FPS avg 210 max 346

Thanks again for checking into this.
 

Attachments

  • Furious 7 CloneBD 1_1_1_5_0 GPU log.cbdlog
    27.2 KB · Views: 1
  • Furious 7 CloneBD 1_1_1_5_0 CPU log.cbdlog
    383.3 KB · Views: 1
  • default-settings-AnyDVD_8.1.0.1_Info_K_FAST7_NA.ziplog
    336 KB · Views: 0
Last edited:
GPU was surprisingly about 3 minutes faster than CPU, but CPU still seems it could use more cores, and better queue capacity for encoder % usage. So if all 10 cores and queue capacity were better used could duration be improved 3+ minutes?
The log files suggest, that the source is an actual disc in a BD drive?
In that case I doubt, there is anything to gain.

32GB in 20 minutes, that's about 26MB/s - I'd be surprised, if your drive can do more than that.
 
Temp and Final destinations were both on HDD. Documents on HDD. Windows and Programs, etc. on SSD
The process is limited by the source component. The optical drive would be the bottle neck in this case and even though the GPU completed faster than the CPU, both were limited by the speed at which the optical drive could feed the data to the encoders. You would probably find the optical drive to be the slowest component with increasing rates going from 5400rpm HDD to 7200rpm HDD to SSD to NvME SSD. The destination drive also has an effect but it's much less pronounced than the source.

Of course, ripping to the HDD or SSD takes time too. So in those cases you would have to consider that time as well and add it to the total to determine which overall process works best in your case.
 
yes, everything you say makes sense to me, but in many of those cases CBD says source drive is the limiting factor. However in some of these cases it does not specifically say that on display, but shows encoder queue capacity filled so still in those cases why doesn't it state the source drive limit?

Also, always at first stage of rip CBD ALWAYS displays Source as the limiter, but almost always before end the source no longer displays as a limiter (avg vs max fps). Still seems to me in those later rip portions that are limited it isn't the source.

PS : I just replaced my P55VT25 with a LG B6. I really loved that panny and it was the longest I ever kept a 1080P tv.
 
An optical drive will ALWAYS be slower than an HDD, it'll even be slower than a 5400RPM mechanical drive. As i stated before, the top 3 bars are SUPPOSED to be at 100%, it's if the 4th bar fills up there's a problem.
 
Back
Top