amd vs intel encoding

[thanasi]

hey say I got an amd 3700 and I was wondering how high do I have to clock it to beat an intel at encoding and other apps.example-3.4 or 3.2 has better encoding right?so how high do I have to overclock it to beat my friends p4 @3.2 at those apps?

 

[stevty2889]

Well you would at least have to be a bit more specific on which applications...if they aren't multithreaded chances are you won't have to do anything to be ahead..

 

[thanasi]

no in the benchmarks like this http://www1.tomshardware.com/cpu/20050221/prescott-09.html
like encoding and other things intel favors

 

[Vegitto]

Between 10 and 30 percent.

 

[Duvie]

You will likely need to hit 2.7-2.8ghz and from a 3700+ that would be 22-27%.....BUt ofcourse that wold be equaling or bettering 800 dollar cpus.....Also remember to be fair a person can buy a 640and OC it 3.8-4.0ghz as well for relatively cheap and then you would have to be closer to 30% to catch them...


I would avoid looking at Toms reviews alone..

That being said if you can wait for a bit amd's dual cores will make this issue as mute as the lowest they will put out is better then anything Intel has....if you OC any it is just pure gravy and a higher plateau the iNtel boys have to catch up to...basically the tables have turned...

 

[thanasi]

true duvie

 

[Zebo]

It depends on the application....... Many results for video and audio where A64 is king... The P4 is unbeatable if the application is extremely optimized for SSE2/3/HT. Anything else is a wash or a victory for the A64.

AMD usually wins
Codecs: xVID..DivX 5.1.1
Consumer programs: Ulead VideoStudio 4,5,6 and 7. Roxio videowave 4,5,6 and 7. Pinnacle studio 7 and older.
MPEG2 Encoders: Canopus procoder, Ligos, bbMPEG.
DVD transcoding: DVD2AVI, VirtualDubMod
Freeware: VirtualDub.
Streaming: Quicktime.
Audio Encoders: Oggenc..LAME

I think it's a wash until you get to X2's of course which dominate everything

 

[shiranai]

These are the only two entries you have to care about:

>> Codecs: xVID
>> Freeware: VirtualDub

/fanboy

I'm fairly sure AT did some testing too ... IIRC, Divx and WMV are faster in the Intel camp. Xvid is supposed to be faster too, but only slightly. I don't think I've ever seen tests done with lots of filters on, and most real-world encoding is going to involve some filtering. I'm not sure if it would change the relative results at all or if it would simply introduce a straight framerate hit evenly across the board.

 

[djshelto]

Originally posted by: Zebo
It depends on the application....... Many results for video and audio where A64 is king... The P4 is unbeatable if the application is extremely optimized for SSE2/3/HT. Anything else is a wash or a victory for the A64.

AMD usually wins
Codecs: xVID..DivX 5.1.1
Consumer programs: Ulead VideoStudio 4,5,6 and 7. Roxio videowave 4,5,6 and 7. Pinnacle studio 7 and older.
MPEG2 Encoders: Canopus procoder, Ligos, bbMPEG.
DVD transcoding: DVD2AVI, VirtualDubMod
Freeware: VirtualDub.
Streaming: Quicktime.
Audio Encoders: Oggenc..LAME

I think it's a wash until you get to X2's of course which dominate everything

interesting...i did some very unscientific testing recently, and found the p4's outperformed the A64 when using virtualdub.

I was running a series of filters against a MJPEG compressed AVI file I capture from vhs.

On an AXP 2400+ Thorton with 512Mb and a 7200RPM ATA 100 hard drive, I averaged about 3.2 - 3.4 frames per second during processing.

On an A64 3200+ Clawhammer with 1Gb and a 7200RPM ATA100 (or 133, not positive), I got about 5.6fps.

On an HP 2.66 P4 with 512Mb and a 40Gb whatever-they-used (work machine, just ran VDUB after hours as a test), I got the same 5.6fps.

The results really surprised me, I figured the A64 would outperform the P4 by a good margin, especially given the extra RAM.

does anyone else have similar experiences? Given that I have at least 100 hours of video to clean up, i've been seriously considering switching from AMD to Intel just to speed up the processing.

 

[Zebo]

Originally posted by: djshelto

Originally posted by: Zebo
It depends on the application....... Many results for video and audio where A64 is king... The P4 is unbeatable if the application is extremely optimized for SSE2/3/HT. Anything else is a wash or a victory for the A64.

AMD usually wins
Codecs: xVID..DivX 5.1.1
Consumer programs: Ulead VideoStudio 4,5,6 and 7. Roxio videowave 4,5,6 and 7. Pinnacle studio 7 and older.
MPEG2 Encoders: Canopus procoder, Ligos, bbMPEG.
DVD transcoding: DVD2AVI, VirtualDubMod
Freeware: VirtualDub.
Streaming: Quicktime.
Audio Encoders: Oggenc..LAME

I think it's a wash until you get to X2's of course which dominate everything

interesting...i did some very unscientific testing recently, and found the p4's outperformed the A64 when using virtualdub.

I was running a series of filters against a MJPEG compressed AVI file I capture from vhs.

On an AXP 2400+ Thorton with 512Mb and a 7200RPM ATA 100 hard drive, I averaged about 3.2 - 3.4 frames per second during processing.

On an A64 3200+ Clawhammer with 1Gb and a 7200RPM ATA100 (or 133, not positive), I got about 5.6fps.

On an HP 2.66 P4 with 512Mb and a 40Gb whatever-they-used (work machine, just ran VDUB after hours as a test), I got the same 5.6fps.

The results really surprised me, I figured the A64 would outperform the P4 by a good margin, especially given the extra RAM.

does anyone else have similar experiences? Given that I have at least 100 hours of video to clean up, i've been seriously considering switching from AMD to Intel just to speed up the processing.

Somethings wrong look here: Dead even 51 (2.2 Ghz) to P4 3.2 Northwood
http://shop.ixbt.com/articles2/athlon64fx-athlon64/

Your clawhammer is a bit faster than a 51 due to non ECC and LL.

 

[AkumaX]

TMPGEnc

 

[djshelto]

actually, if i read that right, they're benchmarking with a 3200+ clawhammer.

In the summary section, they show the clawhammer being 10-20% slower than the p4 in the encoding benchmarks.

I didn't see what p4 they were using, i know it's a 478, but is it one of the HT enabled procs? i get the impression that HT helps more in encoding than in other areas, though that could be wrong.

I also recall reading that VDub favors the p4, though I never saw any reason why or supporting evidence.

 

[Lithan]

Originally posted by: Zebo
It depends on the application....... Many results for video and audio where A64 is king... The P4 is unbeatable if the application is extremely optimized for SSE2/3/HT. Anything else is a wash or a victory for the A64.

AMD usually wins
Codecs: xVID..DivX 5.1.1
Consumer programs: Ulead VideoStudio 4,5,6 and 7. Roxio videowave 4,5,6 and 7. Pinnacle studio 7 and older.
MPEG2 Encoders: Canopus procoder, Ligos, bbMPEG.
DVD transcoding: DVD2AVI, VirtualDubMod
Freeware: VirtualDub.
Streaming: Quicktime.
Audio Encoders: Oggenc..LAME

I think it's a wash until you get to X2's of course which dominate everything

Man you have a good memory.

 

[Zebo]

Originally posted by: djshelto
actually, if i read that right, they're benchmarking with a 3200+ clawhammer.

In the summary section, they show the clawhammer being 10-20% slower than the p4 in the encoding benchmarks.

I didn't see what p4 they were using, i know it's a 478, but is it one of the HT enabled procs? i get the impression that HT helps more in encoding than in other areas, though that could be wrong.

I also recall reading that VDub favors the p4, though I never saw any reason why or supporting evidence.

We never expected to see this. It seems that for everyday Divx use the Athlon64 FX 51 is the best choice

 

[djshelto]

veddy veddy interestink...

 

[carlosd]

Originally posted by: Zebo

Originally posted by: djshelto
actually, if i read that right, they're benchmarking with a 3200+ clawhammer.

In the summary section, they show the clawhammer being 10-20% slower than the p4 in the encoding benchmarks.

I didn't see what p4 they were using, i know it's a 478, but is it one of the HT enabled procs? i get the impression that HT helps more in encoding than in other areas, though that could be wrong.

I also recall reading that VDub favors the p4, though I never saw any reason why or supporting evidence.

We never expected to see this. It seems that for everyday Divx use the Athlon64 FX 51 is the best choice

Intel wins some bechmarks(synthetchic, no real world performance), AMD wins real world tests. We already knew that.

 

[HDTVMan]

Just configure a batch before you go to sleep and kick it off. When you get up in the morning all will be done. Then who cares if it takes 1 hour or 5 hours.

Unless you like watching the slider bar run for hours.

Thats what I do.

 

[djshelto]

that's a good approach for encodes upto 8-10 hours.

in my case, the encodes for a 3 hour clip were taking 26-30 hours, because of the filters involved.

 

[shiranai]

>> in my case, the encodes for a 3 hour clip were taking 26-30 hours, because of the filters involved.

^ truth. t3h filters r0xx0rz j00. I stopped doing encoding on my last rig because I was getting less than 1fps due to filtering. When you're encoding on the order of 10^4 or 10^5 frames, that's no good.