PPBM5 is now available. at DVinfo.net

Harm Millaard · July 24th, 2010, 05:04 AM

The Premiere Pro CS5 Benchmark is now available for testing your system.

What are the major differences in comparison to the PPBM4 test?

1. The testing procedure has been greatly simplified, making it a lot easier to perform the tests and less error prone.

2. The tests themselves are much more realistic, using actual footage instead of synthetic tests.

3. Measurement errors have been greatly reduced, increasing the reliability of test results.

4. MPE testing and H.264-BR testing have been added.

5. An extensive explanation has been added on how to interpret the results.

Bill and I would like to take this opportunity to thank all those who voluntarily made actual footage available in different formats, used to come up with this new test and to those who helped us with beta testing through several builds and their constructive and helpful comments despite busy schedules.

Guys, thank you so much for all your much appreciated help in making this possible.

Look here for full descriptions: PPBM5 home page

David Dwyer · July 26th, 2010, 02:20 AM

Great I'll test it later on today

Hannu Korpinen · July 26th, 2010, 03:20 AM

I don't know what I am missing but I can run only to step.3. When I run statistic.vbs it gives me alarm

Output.txt is also empty.

Harm Millaard · July 26th, 2010, 04:43 AM

This is indicative of a file not being found, in this case the Benchmark.avi file. This can be caused by the fact that in step 2, you did not follow the instructions to the letter. When running through step 2, check carefully that the file names are exactly as shown in the AME screen shot. If you have a dual monitor setup, display the ReadMe file on one monitor and AME on the other. Otherwise print the ReadMe file and follow the instructions to the letter.

Hannu Korpinen · July 26th, 2010, 05:08 AM

This is all I have in Folder PPBM.

Harm Millaard · July 26th, 2010, 05:37 AM

Try deleting all Benchmark.* files with the EXCEPTION of Benchmark.prproj. Leave that one intact. Next rerun the AME encoding as described in step 2 and proceed from there.

David Dwyer · July 26th, 2010, 02:04 PM

Ekk sent my results through - Didn't get the email bounce back this time.

319;secs Total Benchmark Time
;Performance index
86;secs AVI Encoding Time
126;secs MPEG Elapsed Time
97;secs H.264 Elapsed Time
10;secs Rendering Time

I need to turn up my OC!

Harm Millaard · July 26th, 2010, 03:17 PM

David,

You have my mail address. Please send me your Output.txt and Output-MPE.txt files, but also include in one of them your hardware setup, if you are not still using your Proliant 115GS system.

David Dwyer · July 26th, 2010, 03:19 PM

Quote:

Originally Posted by Harm Millaard

David,

You have my mail address. Please send me your Output.txt and Output-MPE.txt files, but also include in one of them your hardware setup, if you are not still using your Proliant 115GS system.

Yeah I have emailed them to the ppbm@att address and included the updated PC spec in the MPE output. Let me know if you didnt get them and I'll get them sent again.

Harm Millaard · July 26th, 2010, 04:24 PM

Bill will have gotten them by now.

Randall Leong · August 1st, 2010, 10:41 PM

I recently ran PPBM5 on my system, both with 6GB and 12GB of RAM installed in my system. I could not believe how much faster the CPU tests ran with 12GB than it did with 6GB! When I had only 6GB, the MPEG-2 encoding test took all of 189 seconds even with the CPU overclocked to 3.7GHz. With 12GB installed, I have the CPU overclocked to a more modest 3.5GHz, and it still took less than 80 seconds in the same MPEG-2 test. (To double-check my confirmation, I re-ran the older PPBM4 test with both 6GB and 12GB, with a very similar improvement on the MPEG-2 test with 12GB versus 6GB in percentage.)

This definitely confirmed my previous suspicions: Although Premiere Pro CS5 can run with as little as 4GB of RAM, I truly feel that 12GB or more per CPU is strongly recommended for optimal performance. (And I would not count the i7-860 or any other Lynnfield CPU as a true Nehalem due to its relatively low maximum memory and bus bandwidth, which is reflected in its relatively low performance in all of the PPBM5 tests compared to a true triple-channel Nehalem CPU.)

Thus, I am glad I replaced the Intel-branded X58 motherboard (which had only four DIMM slots) with a Gigabyte X58 motherboard with six DIMM slots. Had I stuck with the Intel-branded board, I would have needed to purchase three 4GB DIMMs, which still cost $400 to $500 compared to the $250 to $300 I spent on six 2GB DIMMs.

Harm Millaard · August 2nd, 2010, 12:53 PM

Randall,

Those exactly mirror our own experiences during testing. However, there are indications that doubling memory from 12 to 24 GB will give equal MPEG2-DVD improvements. I'm still waiting to hear from Adobe on that.

Randall Leong · August 2nd, 2010, 08:11 PM

Quote:

Originally Posted by Harm Millaard

Randall,

Those exactly mirror our own experiences during testing. However, there are indications that doubling memory from 12 to 24 GB will give equal MPEG2-DVD improvements. I'm still waiting to hear from Adobe on that.

And I would like to add that the CPU score also has some effect on the other scores (disk, H.264 and CPU/GPU) as well.

July 24th, 2010, 05:04 AM	#1
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	PPBM5 is now available. The Premiere Pro CS5 Benchmark is now available for testing your system. What are the major differences in comparison to the PPBM4 test? 1. The testing procedure has been greatly simplified, making it a lot easier to perform the tests and less error prone. 2. The tests themselves are much more realistic, using actual footage instead of synthetic tests. 3. Measurement errors have been greatly reduced, increasing the reliability of test results. 4. MPE testing and H.264-BR testing have been added. 5. An extensive explanation has been added on how to interpret the results. Bill and I would like to take this opportunity to thank all those who voluntarily made actual footage available in different formats, used to come up with this new test and to those who helped us with beta testing through several builds and their constructive and helpful comments despite busy schedules. Guys, thank you so much for all your much appreciated help in making this possible. Look here for full descriptions: PPBM5 home page

July 26th, 2010, 03:20 AM	#3
Hannu Korpinen Regular Crew Join Date: Jul 2009 Location: Keuruu, Finland Posts: 67	I don't know what I am missing but I can run only to step.3. When I run statistic.vbs it gives me alarm Output.txt is also empty. Attached Images

July 26th, 2010, 05:08 AM	#5
Hannu Korpinen Regular Crew Join Date: Jul 2009 Location: Keuruu, Finland Posts: 67	This is all I have in Folder PPBM. Attached Thumbnails

July 26th, 2010, 02:20 AM	#2
David Dwyer Major Player Join Date: Jan 2006 Location: Chester, North West Posts: 565	Great I'll test it later on today

July 26th, 2010, 04:43 AM	#4
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	This is indicative of a file not being found, in this case the Benchmark.avi file. This can be caused by the fact that in step 2, you did not follow the instructions to the letter. When running through step 2, check carefully that the file names are exactly as shown in the AME screen shot. If you have a dual monitor setup, display the ReadMe file on one monitor and AME on the other. Otherwise print the ReadMe file and follow the instructions to the letter.

July 26th, 2010, 05:37 AM	#6
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	Try deleting all Benchmark.* files with the EXCEPTION of Benchmark.prproj. Leave that one intact. Next rerun the AME encoding as described in step 2 and proceed from there.

July 26th, 2010, 02:04 PM	#7
David Dwyer Major Player Join Date: Jan 2006 Location: Chester, North West Posts: 565	Ekk sent my results through - Didn't get the email bounce back this time. 319;secs Total Benchmark Time ;Performance index 86;secs AVI Encoding Time 126;secs MPEG Elapsed Time 97;secs H.264 Elapsed Time 10;secs Rendering Time I need to turn up my OC!

July 26th, 2010, 03:17 PM	#8
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	David, You have my mail address. Please send me your Output.txt and Output-MPE.txt files, but also include in one of them your hardware setup, if you are not still using your Proliant 115GS system.

July 26th, 2010, 04:24 PM	#10
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	Bill will have gotten them by now.

August 1st, 2010, 10:41 PM	#11
Randall Leong Major Player Join Date: Nov 2009 Location: Melrose Park, Illinois, USA Posts: 936	I recently ran PPBM5 on my system, both with 6GB and 12GB of RAM installed in my system. I could not believe how much faster the CPU tests ran with 12GB than it did with 6GB! When I had only 6GB, the MPEG-2 encoding test took all of 189 seconds even with the CPU overclocked to 3.7GHz. With 12GB installed, I have the CPU overclocked to a more modest 3.5GHz, and it still took less than 80 seconds in the same MPEG-2 test. (To double-check my confirmation, I re-ran the older PPBM4 test with both 6GB and 12GB, with a very similar improvement on the MPEG-2 test with 12GB versus 6GB in percentage.) This definitely confirmed my previous suspicions: Although Premiere Pro CS5 can run with as little as 4GB of RAM, I truly feel that 12GB or more per CPU is strongly recommended for optimal performance. (And I would not count the i7-860 or any other Lynnfield CPU as a true Nehalem due to its relatively low maximum memory and bus bandwidth, which is reflected in its relatively low performance in all of the PPBM5 tests compared to a true triple-channel Nehalem CPU.) Thus, I am glad I replaced the Intel-branded X58 motherboard (which had only four DIMM slots) with a Gigabyte X58 motherboard with six DIMM slots. Had I stuck with the Intel-branded board, I would have needed to purchase three 4GB DIMMs, which still cost $400 to $500 compared to the $250 to $300 I spent on six 2GB DIMMs.

August 2nd, 2010, 12:53 PM	#12
Harm Millaard Trustee Join Date: Aug 2006 Location: Rotterdam, Netherlands Posts: 1,832	Randall, Those exactly mirror our own experiences during testing. However, there are indications that doubling memory from 12 to 24 GB will give equal MPEG2-DVD improvements. I'm still waiting to hear from Adobe on that.