1 person found this helpful
The performance increase for each update has been reported and duplicated by many including us. The 5.0.1 update was a memory hog and did not optimize ram allocation like 5.0.2 with the MPE. The MPE optimizations further increased with 5.0.3 which uses less than 50% of the ram that 5.0.1 use to use with the same project.
The notable difference is in the MPEG encoding. Apart from improvements in each version of PR that Eric told you about, this significant improvement was caused by getting rid of a number of background processes, that lifted PR over a memory barrier to cause this. I have no other explanation for it and it surprised me as well. OTOH, Baz, Studio North, encountered a similar situation when going from 12 to 14 GB memory. I guess that killing some memory intensive background processes are the 'culprit'.
These is Just a good indicator of how many ram demand the CPU test MPEG2-DVD I had the same question before and after closing many many procces, services ang to distribute the priorities of the procces I got a very good result to. I will send them today.
Ya i should have been more specific with the 50% memory usage that atleast certain codecs use less than 50% of the ram they use to. The mpeg2 codec still uses quite abit of ram over AVCHD or H264. I need to check R3D material and see about that.
On a side note, I'd just like to say that Jim's little table there is WORLDS easier to read than the one currently in use on the web site.
Some food for thought.
OK Harm, you've improved your PPBM5 score yet again
(Disk IO was 72 on 12Jan11, now 64 on 2Feb11) without any of the specifications changing;
what was the secret this time!?
With the Areca ARC-1680iX-12 and 12HD IN RAID30 I realy dont get how he was not able to be under that time before. I'm ussing the raid controler from the M.B. and my time is 74 sec wiht 3WD,VR in Raid0.
There is not secret's.
You look like you have a I7, dont you? I can giveyou good imputs tune your pc for maximun performance.
Jes. I guess is you who have the i7860...
Use 1 of your HD 7200RPM for the O.S. and the SSD RAID0 for the proyect. I promes your Disk I/O TEST Will go done.
And please tell me which RAM do you have, and the cl latency, These is very important, even I´m sure some do not agree whit me.
To set the perfect balance in the performance of the PC is not a easy job.
We need to read more to understang the comunication off all PC components.
I was testing again with a new H.264 time-line for a future version, as well as running tests comparing direct Export versus the Queue. So obviously you tune your run by killing processes or suspending them and run the tests. And I had just rebooted after a disk optimization session.
Interesting, because there are some people claiming that direct export is a lot faster than queuing, I did not find any proof of that, apart from AVI.
Now these figures are NOT comparable to the PPBM5 benchmark, apart from the AVI results.The AVI results are based on the same time-line in the benchmark.
Why? Because both the H.264-BR and the MPEG2-DVD time-lines have additional tracks, non-MPE effects added, in short made a lot more complex. The only reason I'm showing these outcomes is to show that for most export formats direct Export is not faster than Queuing in AME.
There is one weird thing though, why is direct export with AVI so much faster than with AME?