AMD v. Intel which is better for DVD Shrink?

[madoka]

I've been doing some backups using DVD Shrink 3.2. I was under the impression that P4s were better at video encoding, but I've found that my Sempron was actually faster. My P4 3.2 Prescott, 1gig GEIL Ultra, 300gig SATA 16mb cache takes upwards of 40 minutes longer to encode a DVD than my Sempron 2400+, 512mb Kingston Value Ram, 250gig UATA does.

Is this just because AMD processors are better for DVD Shrink or is there something wrong with my setup?

 

[Duvie]

No you likley have something f'd up......My P4 2.4 @ 3.2ghz is faster at dvdshrinking same gone in 60 seconds DVD 64% (ful disc) then my system in my sig.....A 3.2e is likley slower then my 3.2 was since I was ocing and it was a northwood....However the sempron is a POS!!!! I had one...the clock speed is actually only equal to a 2100+ tbred...NO WAY and I repeat NO WAY will a 2100+ xp beat the p4 at 3.2ghz (E or not) in DVDshrink....

I dont know of a DVD that even takes me 40 minutes let alone 40 minutes longer...YOU GOT SOMETHING SCREWED UP!!!

The best of the P4's will beat the best of AMD64's in dvdshrink.....I dont think a 2400+ sempron would beat a p4 2.4ghz...my 3200+ barton couldn't beat my 2.4ghz P4 at 3.0ghz, and that was with the barton using pc3200, 512kb of l2 cache, etc....

Did you do deep analysis on one and not the other??? The DVDshrink program is actually HT aware, however most need to remember it is extremely IO dependent as well.....Are you ripping from the DVD straight??? Same dvd drive in both??? IDVD drives will not rip the same...My liteon 16x will kick my NEC 8+/- DVD-R/RW arse by 2-3 times....It hits a maximum of 8.7x ripping on the Gone in 60seconds DVD while the NEC maxes out at 3.5x...It appears not to be a good ripper. Also HDD speed is very important....

 

[MisterChief]

For pure coding, the P4 is the leader. For gaming, the AMD64 rules. So, to answer your question, I would stick with Intel; unless you plan on gaming...You can wait a couple extra minutes for your DVD's

 

[Wingznut]

Originally posted by: Duvie
Are you ripping from the DVD straight??? Same dvd drive in both??? IDVD drives will not rip the same...My liteon 16x will kick my NEC 8+/- DVD-R/RW arse by 2-3 times....It hits a maximum of 8.7x ripping on the Gone in 60seconds DVD while the NEC maxes out at 3.5x...It appears not to be a good ripper.

Agreeing with Duvie... There's a huge difference between drives, especially using a DVD-ROM and DVD-RW.

 

[TronX]

When it comes to burners they suck at ripping speeds. There is also a speed limiter built in that you must bypass with a hacked firmware. As for the best speeds if you don't use a Lite-on you will need to use a regular DVD player. Your speeds should range from about 7X to 12X during a rip.

Now for DVDShrink if you uncheck sharp it will run much faster. However if you have the time it's worth using because it makes the movie look better.

Making DVDShrink run faster is more MHz speed. 2GHz vs 4Hz is only 2X faster, so don't think things will jump right on DVD disk in minutes.

If you reauthor the movie you can sometimes get away with zero compression and this will be the fastest way overall.

 

[gwag]

Originally posted by: Duvie
No you likley have something f'd up......My P4 2.4 @ 3.2ghz is faster at dvdshrinking same gone in 60 seconds DVD 64% (ful disc) then my system in my sig.....A 3.2e is likley slower then my 3.2 was since I was ocing and it was a northwood....However the sempron is a POS!!!! I had one...the clock speed is actually only equal to a 2100+ tbred...NO WAY and I repeat NO WAY will a 2100+ xp beat the p4 at 3.2ghz (E or not) in DVDshrink....

I dont know of a DVD that even takes me 40 minutes let alone 40 minutes longer...YOU GOT SOMETHING SCREWED UP!!!

The best of the P4's will beat the best of AMD64's in dvdshrink.....I dont think a 2400+ sempron would beat a p4 2.4ghz...my 3200+ barton couldn't beat my 2.4ghz P4 at 3.0ghz, and that was with the barton using pc3200, 512kb of l2 cache, etc....

Did you do deep analysis on one and not the other??? The DVDshrink program is actually HT aware, however most need to remember it is extremely IO dependent as well.....Are you ripping from the DVD straight??? Same dvd drive in both??? IDVD drives will not rip the same...My liteon 16x will kick my NEC 8+/- DVD-R/RW arse by 2-3 times....It hits a maximum of 8.7x ripping on the Gone in 60seconds DVD while the NEC maxes out at 3.5x...It appears not to be a good ripper. Also HDD speed is very important....

"NO WAY and I repeat NO WAY will a 2100+ xp beat the p4 at 3.2ghz" it did otherwise he wouln't have posted it cant you read. The Sempron is not a POS just a slower CPU, HDD is not "very important" any modern drive will be close to the speed of the fastest drives a crappy drive can move 25MB/s dvd shrink wont use more than 8MB/s.
some DVD drives simply dont go that fast when its a video DVD by design. Lite-on do not suffer from this problem my 4X dvd-rw will do 10X near the end of discs.

 

[Duvie]

Originally posted by: gwag

Originally posted by: Duvie
No you likley have something f'd up......My P4 2.4 @ 3.2ghz is faster at dvdshrinking same gone in 60 seconds DVD 64% (ful disc) then my system in my sig.....A 3.2e is likley slower then my 3.2 was since I was ocing and it was a northwood....However the sempron is a POS!!!! I had one...the clock speed is actually only equal to a 2100+ tbred...NO WAY and I repeat NO WAY will a 2100+ xp beat the p4 at 3.2ghz (E or not) in DVDshrink....

I dont know of a DVD that even takes me 40 minutes let alone 40 minutes longer...YOU GOT SOMETHING SCREWED UP!!!

The best of the P4's will beat the best of AMD64's in dvdshrink.....I dont think a 2400+ sempron would beat a p4 2.4ghz...my 3200+ barton couldn't beat my 2.4ghz P4 at 3.0ghz, and that was with the barton using pc3200, 512kb of l2 cache, etc....

Did you do deep analysis on one and not the other??? The DVDshrink program is actually HT aware, however most need to remember it is extremely IO dependent as well.....Are you ripping from the DVD straight??? Same dvd drive in both??? IDVD drives will not rip the same...My liteon 16x will kick my NEC 8+/- DVD-R/RW arse by 2-3 times....It hits a maximum of 8.7x ripping on the Gone in 60seconds DVD while the NEC maxes out at 3.5x...It appears not to be a good ripper. Also HDD speed is very important....

"NO WAY and I repeat NO WAY will a 2100+ xp beat the p4 at 3.2ghz" it did otherwise he wouln't have posted it cant you read. The Sempron is not a POS just a slower CPU, HDD is not "very important" any modern drive will be close to the speed of the fastest drives a crappy drive can move 25MB/s dvd shrink wont use more than 8MB/s.
some DVD drives simply dont go that fast when its a video DVD by design. Lite-on do not suffer from this problem my 4X dvd-rw will do 10X near the end of discs.

NO shite sherlock...i can see what he wrote...I am telling him he has something messed up cause it should not happen unless he has extremely limited the IO ability of the p4...I find that hard to do to make the p4 40 minutes slower....

The sempron 2400+ is a POS!!! Its speed is a 2100+ tbred so ppl need to remember its rating is not the same as a 2400+ tbred.

My liteon will hit speeds of 12x but on some types of media, dual layer disc, etc the speeds dont get as high....

HDD is not "very important" any modern drive will be close to the speed of the fastest drives a crappy drive can move 25MB/s dvd shrink wont use more than 8MB/s.

Quit speaking no nothing.... I have tested this (going to be in a review I will release once I have my raid setup ready to test) that shows different combination of drives...yes combination of drives (HDDs) can result in speeds 25% faster. Usually you can gain faster dvdshrink speeds when you read the file on one drive and write to another...Both HDDs need to be on their own channel ofcourse.

I can see a difference with Raptor to 80gb segate SATA drive...size of drive, amount of data filled onto it, fragmentation are also factors...

Now when you read from the DVD drive the HDD as little effect since the DVD drive is the bottleneck ....Often times I will just rip the whole disc to the HDD with the liteon and then dvdshrink later.

 

[zakee00]

ok people...duvie knows more about this subject than ANY of you. so listen to him, don't even argue.

 

[Sc4freak]

:thumbsup:

 

[madoka]

I checked both computers, along with my A64 3000+ and noticed something interesting with the task manager. With both the AMDs (the A64 and the Sempron), the CPU usage is at 100% while using DVD Shrink, yet with the P4, the CPU usage is stays between 15-35%. All the computers have the low priority option unchecked.

Is there a reason why there is such a big difference between processors?

 

[Duvie]

Are you ripping directly from the DVD and shrinking at same time??? What I have seen is it is hard to keep cpu at 100% why ripping from DVD cause the bottleneck of data being delivered does not keep the cpu busy 100% of the time....Using my Liteon DVD-rom which is a better ripper it keeps my usage higher then the NEC but not as high as the HDD..none of them holds at 100%...

Liteon - Gone in 60 seconds (full rip) 53% compression *** This is about as high as most would go without any real degradation of content*** I would mainly dvdshrink the movie only with this high of compression and thus movie alone on this disk would only be a 92% compression....

http://home.comcast.net/~duvie23/dvdshrink_usage-dvdrom-53_compression.JPG


NEC 8x DVD-r/-rw/+r/+rw - Gone in 60 seconds (full rip) 53% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-dvdrom-53_compression_NEC.JPG

Look what a crappy ripper will do for you....Take longer no doubt but make the cpu sit and wait for the data to transcode....

NOw....

Same movie ripped to HDD then 53% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-53compression.JPG

Look at how it is higher then the Liteon but way higher then the NEC...the story is the cpu doesn't have to wait as long. BUt in contrary to numbnuts above it is IO dependent as reading and writing speeds undoubtedly hold it back somewhat....remember 100% cpu usage means the cpu has to wait for nothing and all the IO systems are keeping the cpu saturated with info. Perfect world no doubt....

NOw the fact is the amount compression also determines the cpu usage. The less the encoding or less percentage of shrinking will result in less work for the cpu but mean the data is handed faster to the drive to basically be rewritten. Again it seems that becomes more and more IO dependent as it merely copying the file to another folder.


I took same file on HDD and reduced it to 65% compression by dropping the foreign soundtracks...Less work right???

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-65_compression.JPG

notice how the usage drops....less trancoding per frame means the cpu is waiting more for the read write operations....

NOw I still frame some of the extras and get it to 85% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-85compression.JPG

notice the pattern???



Remember PPL!!!! when a P4 says 35% it is more like 70% due to the fact how HT works, and how the cpu task manager reports it....It is actually not doing just 35%...

In a perfect world without IO limitations the cpu would not be limited as thus would hold 100%....It would not have to wait and would just get done as fast as it could...GO and BLOW!!!! Obviously this is not the case...i have noticed ram speed and system FSB thus bandwidth has some effect on the time as well....play with it...

 

[gwag]

Originally posted by: Duvie
Are you ripping directly from the DVD and shrinking at same time??? What I have seen is it is hard to keep cpu at 100% why ripping from DVD cause the bottleneck of data being delivered does not keep the cpu busy 100% of the time....Using my Liteon DVD-rom which is a better ripper it keeps my usage higher then the NEC but not as high as the HDD..none of them holds at 100%...

Liteon - Gone in 60 seconds (full rip) 53% compression *** This is about as high as most would go without any real degradation of content*** I would mainly dvdshrink the movie only with this high of compression and thus movie alone on this disk would only be a 92% compression....

http://home.comcast.net/~duvie23/dvdshrink_usage-dvdrom-53_compression.JPG


NEC 8x DVD-r/-rw/+r/+rw - Gone in 60 seconds (full rip) 53% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-dvdrom-53_compression_NEC.JPG

Look what a crappy ripper will do for you....Take longer no doubt but make the cpu sit and wait for the data to transcode....

NOw....

Same movie ripped to HDD then 53% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-53compression.JPG

Look at how it is higher then the Liteon but way higher then the NEC...the story is the cpu doesn't have to wait as long. BUt in contrary to numbnuts above it is IO dependent as reading and writing speeds undoubtedly hold it back somewhat....remember 100% cpu usage means the cpu has to wait for nothing and all the IO systems are keeping the cpu saturated with info. Perfect world no doubt....

NOw the fact is the amount compression also determines the cpu usage. The less the encoding or less percentage of shrinking will result in less work for the cpu but mean the data is handed faster to the drive to basically be rewritten. Again it seems that becomes more and more IO dependent as it merely copying the file to another folder.


I took same file on HDD and reduced it to 65% compression by dropping the foreign soundtracks...Less work right???

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-65_compression.JPG

notice how the usage drops....less trancoding per frame means the cpu is waiting more for the read write operations....

NOw I still frame some of the extras and get it to 85% compression

http://home.comcast.net/~duvie23/dvdshrink_usage-HDD-85compression.JPG

notice the pattern???



Remember PPL!!!! when a P4 says 35% it is more like 70% due to the fact how HT works, and how the cpu task manager reports it....It is actually not doing just 35%...

In a perfect world without IO limitations the cpu would not be limited as thus would hold 100%....It would not have to wait and would just get done as fast as it could...GO and BLOW!!!! Obviously this is not the case...i have noticed ram speed and system FSB thus bandwidth has some effect on the time as well....play with it...

True genius!!! figures out that is DVD shrink works faster if you rip it to your hard drive first!! Boy you should defiantly write an article about this for all the 2nd graders who haven?t figured this out yet.

 

[Duvie]

That wasn't the whole point.....I showed how the dvd drive effected this and basically took your stupid statement and poked holes in it...DVDshrink is extremely IO limited. MOre so with DVD ripping and less so with HDD, but nonetheless it still is....It gets amplified more so when less compression is needed....

So when will I be seeing you in my 2nd grade class??? Dumb-ass!!!

I was also trying to explain his cpu usage issue.

 

[Duvie]

Sidenote:

A good reason to rip the movie to HDD using a dvdripper program like dvdcryptor is problems that are arising from Sony's new DVD prtotection called ARccOS...DVDshrink 3.2 wont rip and shrink these DVDs.....

I noticed issue with my Bruce Almighty and now just found out why.....Also had it happen helping my buddy doing a Journey Concert Video...I bypassed it by using decryptor first then DVDshrink and it worked...I didn't know about this issue until just about 5 minutes ago....He made the copy (he owned the original, I saw it) so fiends could borrow it and not F up his original. If ppl let friends borrow DVDs or family members, then you will understand a legitmate use for DVDshrink and copying DVD movies....

 

[madoka]

Originally posted by: Duvie
Are you ripping directly from the DVD and shrinking at same time???

If I understand you correctly, yes I am. But I was doing that with all three computers, so I don't understand why the P4 would only be about 20-35% when the AMDs are at 100%. Also, when I later tried DVD Decrypter (it's much more tolerant of scratches on disks and useful for Sony's new copy protection scheme on new titles like Resident Evil 2 and the Forgotten), I found that the P4 would be using 100% CPU usage when I switched over to DVD Shrink. BTW, the P4 is using a Lite-On DVD Drive and a NEC 3500 burner. Could there be a riplock on the Lite-On?

 

[ribbon13]

http://techreport.com/reviews/2005q1/pentium4-600/index.x?pg=10

I hate Intel, but I won't deny them where they're better.

 

[Technonut]

madoka:

When you hit the Back-Up! button, click the Options tab, and un-check "Run backup in low priority mode".

I always use DVD Decrypter to rip the DVD to my hard drive first... I just used DVD Shrink to backup "The Village" with the P4 rig in my sig. With "Run backup in low priority mode" unchecked, Windows Task Manager reported 100% processor usage throughout the entire encoding process.

From the start of DVD Shrink to burned disc took 47 Minutes 3 Seconds.... This was with Deep Analysis and High Quality Adaptive Error Compensation (Maximum Sharpness) enabled...

EDIT: I am currently running some comparisons at different DVD Shrink settings.. I will post the results shortly...

 

[Technonut]

Hmmm... I used DVD Decrypter to rip Alien vs Predator to my 120GB WD 1200JB.

I went ahead with a Deep Analysis with DVD Shrink, then just encoded to the HDD with the following results:

HQAEC: Enabled (Maximum Sharpness)
Run backup in low priority mode: Not Checked
Time: 28min 53sec

HQAEC: Enabled (Maximum Sharpness)
Run backup in low priority mode: Checked
Time: 28min 51sec

HQAEC: Disabled
Run backup in low priority mode: Not Checked
Time: 12min 8sec

HQAEC: Disabled
Run backup in low priority mode: Checked
Time: 12min 0sec

Even though I noticed the processor usage dropping at times with low priority checked, it was still faster than running without it checked, and the processor usage staying maxed out...

I think I will see what happens when I use my Raptor for the rip, and encode to the 1200JB....

EDIT: Here are the times with the rip placed on the Raptor, encoding to the 1200JB:

HQAEC: Enabled (Maximum Sharpness)
Run backup in low priority mode: Not Checked
Time: 28min 6sec

HQAEC: Enabled (Maximum Sharpness)
Run backup in low priority mode: Checked
Time: 28min 16sec

HQAEC: Disabled
Run backup in low priority mode: Not Checked
Time: 9min 47sec

HQAEC: Disabled
Run backup in low priority mode: Checked
Time: 9min 50sec

Results are somewhat confusing... But as expected, encoding from the Raptor to the 1200JB did help the times somewhat... These results came more into line, as the high priority times were faster, compared to the low priority mode box checked...

Anyway, it appears that there is not really much of a difference between the low and high priority modes as I originally thought from watching the CPU Usage.....

I am expecting a 74GB Raptor to arrive this week... It will be interesting to see how much time is shaved encoding from Raptor to Raptor.....

 

[Duvie]

NIce technonut.....

My fastest times were seen using ripped file (decrypted to HDD first) on 2nd raptor and writing to 1st raptor (OS drive)....


You guys are running a newer version obviously. For reasons of comparing to my older system I have been running 3.0 beta 5....There obviously may be some enhancements to versions that help the P4 stay loaded fully...I will retest on the latest....same files etc....

 

[MemoryInAGarden]

Even my Celeron D 2.4 @ 3.7 does well against my 2.4C@3.4 (for the price, of course) I'd stick with Intel anyday for encoding, whether with budget procs or expensive ones.

 

[Duvie]

version 3.2 will hold me at 100% at everything...whether adaptive is check ior not, whether low priority is checked or not....

Funny thing is 3.0beta 5 only hold me at 58% and yet it only takes 3 seconds longer with adaptive off and set to regular priority...if I ran a program like mp3downloading and talking on the net while dvdshrinking the program with 3.0 beat I still get 3:46 second in my test file....With 3.2 it bumps it up to 4:03 seconds while running mp3downloading...

So the question is what changed??? It holds 100% but doesn't appear to get it done any faster....the adaptive thing probably helps the visual quality though I find 75% or higher compression rarely effects quality that would even be visually distinguishable.....

I also see plenty of bugs ebing reported wit this one....I think I would stick with the 3.0 version...I have liek the 3.1 version I may try as well...

Why hold the cpu at 100% versus 58% and get it done at near same time....Unless they say visual clarity is being enhanced I think it seems wasteful for a cpu like the A64 which wont benefit from multitasking like the HT....Cant steal my cpu cycles here cause it will effect me more so....

I apologize to qwag...cause had I only ran test on this version I would not think there was an IO limitation as well....