Anyone wish to do a group speed test ?

[JonRead]

Okay, I just wondered, does anyone want to do a group speed test for encoding times ? So people can then see if their machines seem too slow or if another machine seem ideal in comparison ? I am curious all the time about speeds of encoding with other peoples computers in comparison to mine. I know it can depend on the software used, but perhaps we could all use a freeware one for the test ?

Just an idea. Say, encode a 10 minute file in DVD format, then post your computer specs in full, along with the time taken. Of course, all tests should be done without multi-tasking.

If anyone thinks this is a good idea, reply, if it has been done recently before, then I apologise. But if you do think it is a good idea, then perhaps reply here and we can set something up and work out "rules" for it.

[auenf]

well the speed of encoding will with the footage as well ;)

Enf...

[JonRead]

Even footage encoded at constant bit rate ?

[ArdenDag]

actually, it depends on the filters and resizing because of the footage.

the filters take into account the footage that's being passed

without resizing and without filtering of any kind the tests are kind of useless

EDIT:

I'm running an AMD Duron 750 mhz, 392 mb of ram on Win 2k, I can get approximately 20 fps using a script with:

NO filters (i.e. just resizing, cropping)

But 95% of the DVDs I have require more than that, so I let them run overnight.

[auenf]

hmm, i didnt type my post too well last night :/

i meant to say that the speed of encoding varys with the complexity of the footage used.

eg, Red Hot Chili Peppers - The Zephyr Song encodes slower than Motor Ace - Pieces.

Enf...

[bastioned]

Well here is my grain:

Laptop Clevo P4 2 Ghz
512 MB DDR RAM
60 GB HD
W2000 Pro
Input PAL DVD 720x576
Avisynth -> CCE 2.50
No filters
Resize and AddBorders

Average 1.5x

Edit:

Desktop AMD XP 1500@1800
512 MB SDRAM 133
RAID0 2xIBM 80 GB :D:D:D
WXP Pro
Input PAL DVD 720x576
Avisynth -> CCE 2.50
No filters
Resize and AddBorders

Average 0.9x

I believe the last one is that slow due to the memory architecture used, I expect to jump at least to 1.5x when I switch to DDR 333, any one already with this setup could share some light , guys ?

[john@sympatico.ca]

JonRead:

If you were to make available for download:

-a small VOB (under 100 MB)
-any required files (d2v or TMPGTemplate.txt etc...)
-provide a procedure to follow to load and process the project (would have to be with freeware etc...)

this would ensure consistency between systems and participants.

I would be willing to run this through my system / provide results etc...

[JonRead]

Thanks guys for the posts. I will see if I can get some freeware/copyright free items uploaded.

One other quick thing. I always thought encoding was down to the processor rather then memory, but "bastioned" mentions that he expects encoding time to speed up with faster memory. Is this correct ? I am looking for a new PC for encoding only, and I was going to go for CPU speed over memory. Anyone know ?

[bastioned]

From my experience I can say that definitely the memory matters, and a lot. Just take into consideration that my AMD XP is faster than the P4 by a sort margin, and even so encoding is faster in the P4. The only reason I can think of is that the AMD is paired with 133 SDRAM and the P4 is with DDR 333. Even the amount of memory is the same, if you look at the specs. And the AMD is running on a RAID system: everything should make it faster than the P4, and it isn´t. The only difference: the memory architecture.

[insite]

Glad we're talking about this, because I've been scratching my head as to why I don't get better throughput. CCE will only use about 55-60% of my processing resources. Here's the setup:

Dual AMD Athlon XP 2100+
1GB PC333 DDR Ram
Adaptec AHD-2940UW SCSI Controller
(2) Striped Fujitsu 10,000 RPM 36GB HD w/8MB cache
(2) Striped IBM Ultrastar 10,000 RPM 36GB HD w/8MB cache
(1) Seagate Cheetah 10,000 RPM 18GB HD w/4MB cache

Now, it's important to note that while one instance of CCE may not run that fast, multiple instances run well. Average encoding speed with once instance of CCE running is about 1.4x, which sucks for a machine of this magnitude. If I run two instances of CCE striped accross the multiple processors, CCE crashes. If I run one instance of CCE from one account on one processor and another instance on another account on the other processor (terminal services), they both average about 1.3. So, the best I can do is to encode two simultaneous movies at 1.3 each. This is a pain in the ass.

On the upside, while CCE is running, I can be using IFO Edit to strip streams / VOB id's, using scenarist to remux, using dvd decryptor to rip, RecordNow Max to burn a DVD, AND surf the net all at the same time with no noticable performance decrease.

Basically, this setup allows me to do a ton of stuff at once even if it doesn't give me a very fast single instance encoding rate.

If anyone has any ideas on how to tweak this to spead up the single instance encoding rate, I'd love to tinker with it. I just don't think that CCE is very good at multithreading.

[bastioned]

Not just at multitasking: I have noticed that , even in single setup, CCE never uses my CPU at full speed, not in my AMD nor my P4. I think we´ll see future improvements in the CCE´s CPU usage area...

[insite]

With dual CPU's though, if I set CCE to exclusively use one processor all by itself (no other processes on it), it uses 100% of that processor. Kicking in help from the second only yields a marginal performance increase. Very strange.