pro2 & multiple encoder processes

[l8nights]

I was just trying to get this set-up on my new p.c. and having rebuilder 1.08.1,procoder2 I can't get neither eclpro0.52 or the eclpro test build for dualdvdrb to give me multiple encoder processes functionality I just keep getting error, something about invalid preset's
this happened on two different source files

I thought this was already addressed but a search left me puzzled (lost all my book marks).

any one got it up and running?? is it possible?

thnx 4 any feedback

[wmansir]

Last time I (accidentally) tried ECLPro wouldn't run two instances simultaneously. That was probably an old version though. Even if it works now, I'm not sure ProCoder itself is able to run two instances.

[l8nights]

When using the tool dualdvdrb and the test eclpro robot made for it. I remember it ran fine I just didn't see any increase w/ the p4 now that I got a 920 I was ready to test again but with rb's built in settings I just can't get it to go right!

preset's are invalid?


that's all I get, I thought I read about ti somewhere here but I continually search and can't find the answer??

*edit=fat fingers =)

[kmac61]

i've got it working fine though with a ht the gain is very little
maybe 7-10% at best.
I didn't install anything new other than updating to 1.08.1 and then checking the box in settings for m.e.p.
I would suggest getting procoder to run normally 1st. and not using anything not in the 1.08.1 full install

[jdobbs]

I saw that error too when I first implemented it. There is some kind of timing problem involved with the dual processing mode of ProCoder -- because you could run the same job twice and get it one time but not the other. It appeared that the two processes were trying to access something at the same time. So I added a minimal delay time separation to ensure no two ProCoder processes would start at exactly the same time. On my system it stopped the error from occurring.

Frankly, I don't really know enough about ProCoder to even come close to calling myself expert in it -- I'm not even sure when, how, or why the "presets" are loaded...

[l8nights]

it may be the source,exorcism of emily rose R1
here is what happened while using cce instead
- Creating M2V for VTS_01 segment 13
- Creating M2V for VTS_01 segment 14
- Creating M2V for VTS_01 segment 15
- Creating M2V for VTS_01 segment 16
- Creating M2V for VTS_01 segment 17
- Creating M2V for VTS_01 segment 18
- Creating M2V for VTS_01 segment 19
- Creating M2V for VTS_01 segment 20
- Creating M2V for VTS_01 segment 21
-- V01002100001022 no .M2V, retrying.
- Creating M2V for VTS_01 segment 21
- Creating M2V for VTS_01 segment 22
- Creating M2V for VTS_01 segment 23
- Creating M2V for VTS_01 segment 24

it don't look like a problem but the thing w/ procoder2 happened w/ walk the line as well
so now I'm all kinda confused!

[robot1]

I'm investigating the problem.

@l8nights
I've sent you an email.

[l8nights]

just got the message and hope we can work this out it's getting to me not having procoder on the new dual core.

check your inbox =)

[A64]

I was running into the "Presets Are Invalid" message on two X2 3800 and one Opteron 165 system until I did the following:

1 - Turn Off The Preview In Procoder 2.04 (both windows)
2 - Replaced The Buffalo 512Mb DDR 400 Memory (3-3-3-8) with the Patriot Low Latency DDR 400 Memory (2-3-2-5)
3 - Run Dual Channel rather than Single Channel

I am running 1 ASRock 939Dual-Sata2 (Opteron), 1 ASRock 939SLI32-eSATA2 (X2), and 1 ASRock 939NF4G-SATA2 (X2) for the motherboards.

Now if I change one of the three items listed above, I start running into the message. I have been getting about a 37% gain on the Opteron 165 and about a 32% gain on the X2 3800 over a single instance of Procoder. Now if you try to overclock, the USB key tends to quit talking. I did not try any memory timings in between CL3 and CL2.

[l8nights]

these things didn't seem to help me maybe the ram is the issue on the new p.c. as it's ddr2 I got it set to 4-4-4-12 but that's about as low as I can push it or so I've read!


I was just able to get quite a large file (8+gb's) do-si do'ed ( three stepped) in 135 mins using procoder 2.04 building iso's w/ the new 920@3.6ghz, that's a big step from my ole p4 3.0's times

[shoarthing]

I was running into the "Presets Are Invalid" message on two X2 3800 and one Opteron 165 system until I did the following:

1 - Turn Off The Preview In Procoder 2.04 (both windows)
2 - Replaced The Buffalo 512Mb DDR 400 Memory (3-3-3-8) with the Patriot Low Latency DDR 400 Memory (2-3-2-5)
3 - Run Dual Channel rather than Single Channel

I am running 1 ASRock 939Dual-Sata2 (Opteron), 1 ASRock 939SLI32-eSATA2 (X2), and 1 ASRock 939NF4G-SATA2 (X2) for the motherboards.

Now if I change one of the three items listed above, I start running into the message. I have been getting about a 37% gain on the Opteron 165 and about a 32% gain on the X2 3800 over a single instance of Procoder. Now if you try to overclock, the USB key tends to quit talking. I did not try any memory timings in between CL3 and CL2.
- of those variables; 1) looks a sensible-enough Procoder issue; 2) looks like a hardware issue, 3) is genuinely interesting.

In a few weeks I'll run some 2x Procoder tests on a proper SMP box [2x Opteron 252s & 8 sticks ECC/Reg 2-3-2-6 stuff] in NUMA & non-NUMA memory configs . . would also suggest that 512MB/1GB [assume this is what you use] is very little memory to run this sort of testing [Procoder eats the stuff like it was candyfloss].

30-plus% is a worthwhile gain . . is this all at mastering quality?

[jdobbs]

I'm pretty sure it's timing related. I think the two processes are attempting to do "something" that conflicts... the difference in hardware just changes the timing slightly. I'll increase the minimum separation between processes a little for the next version and see if that helps.

[robot1]

If there is a user which can repeat the error, I could send him by email a couple of EclPRO versions to check if we can solve.
Send me email by PM.

[A64]

- of those variables; 1) looks a sensible-enough Procoder issue; 2) looks like a hardware issue, 3) is genuinely interesting.

In a few weeks I'll run some 2x Procoder tests on a proper SMP box [2x Opteron 252s & 8 sticks ECC/Reg 2-3-2-6 stuff] in NUMA & non-NUMA memory configs . . would also suggest that 512MB/1GB [assume this is what you use] is very little memory to run this sort of testing [Procoder eats the stuff like it was candyfloss].

30-plus% is a worthwhile gain . . is this all at mastering quality?


By shutting off the Preview, it actually speeds up the encode slightly (a couple of minutes across a full disk encode). Now the switching of the memory to the 2-3-2-5 in dual channel also gave a slight boost in the encoding performance. I am only running Procoder at the Mastering level. I have only tried the 1Gb memory configuration. I will have to try doing the 2Gb in the system here later.

On a sample encode of just one segment of an episodic disk, it took 31 minutes on an Opteron 150 (S939) single encoder vs 25 minutes on the Opteron 165 (S939) dual encoder. I just swapped out the Opteron 150 for the Opteron 165 on the board so everything else was the same.

[l8nights]

shoarthing just to test a theory are you using any memory above 400mhz?? like say, ddr2 @533 or 667?

[shoarthing]

shoarthing just to test a theory are you using any memory above 400mhz?? like say, ddr2 @533 or 667? - the Opteron s940 platform uses Registered DDR1 @ a max of 400MHz [ie uses PC2100/2700/3200]. I am using a pair of 2.6GHz 2-way s940 Opterons with 8 sticks of 2-3-2-6 [2T] Reg/ECC 400MHz DDR1: mobo is a workstation SMP board called the Iwill DK8EW. It is non-trivial to overclock these workstation boards & most folk don't.

The upcoming s1207 platform will use DDR2 @667 & above [some folks believe the new memory controller will allow for DDR3 as well]; but at present no s1207 SMP boards are available & there is no worthwhile performance advantage [over low-latency DDR1] till you get to 800MHz or greater DDR2.

Point of what I was saying before is that high-end SMP boards can be set to configure their memory to what is called 'NUMA' configuration - this basically = a local memory pool to each socket, so where you're running plural independent processes [affinity to a socket] each socket has fastest possible access to its own memory.

[l8nights]

I'm thinking the ram setting's may be responsible for the preset's are invalid error message, but haven't had a lot of time to test theories! thsnx for the input =)

[robot1]

This is a new test build for EclPRO.
I need some feedback, so if any user can test... expecially the one who experienced problems with the previous build.

I try to always disable the preview in the Procoder window, as some users have reported problems with multiple encoder processes and preview.
To disable the preview, I have to change a key in the registry.

Changelog:
fixed "settings of this preset are invalid" for filenames with "&" char.
fixed a bug with dots (".") in the filename.
Disabled preview while encoding for ProCoder2 (should fix bugs in multiple encoder processes)




Warning:
This is a test build.
The program could change a value in your windows registry, so if you are not familiary with the registry, please don't test it.
You should not use it on a production machine.

[l8nights]

this version has successfully fixed the annoying "preset's are invalid" error that would not let me use procoder2 what ever the issue was.

I also noticed that if I wanted to see the previews I could by simply ticking them on, just have to tinker =)

no bugs to report =(
good work, robot1