7 Replies Latest reply on Sep 12, 2012 10:30 AM by Fabio Pis

    Hardware Update

    Fabio Pis Level 1

      Hello to all.


      Yesterday I did a major upgrade to my system I have changed ram, motherboard and CPU


      Previous hardware:

      gigabyte x58 ga ud3r with i7 980X and 24gb Kingston Hyperx genesis kit cl9


      new hardware:

      asrock x79 extreme 9 with i7-3960x (C2 stepping) and 32 gb Kingston genesis kit cl9


      ppbm5.5 results are nice:








      os disk Force GT 240 gb

      Project disks 2 600 gb velociraptor in raid 0


      Same GFX (gtx 680) in pci-e 2.0 with latest betadrivers 306.02(I have no applied the hack to enable pci-e 3.0 but I'll do very soon).


      System is stable at 4.4 Ghz with a phanteks ph-tc14pe cooler


      Previous system was stable at 4.0 Ghz with noctua nh-d14 cooler.

      I had reached good results at 4.3 Ghz with PPBM5.5 but not stable for my needs.


      New system is stable also at 4.6 Ghz with PPBM5 benchmark but for dayly 3d use, temperature are too high in heavy load (about 85°)



      with pci-e 3.0 enabled:



      with no MPE:



      I think it's a good results.


      Bill & Harm, May I submit again?

        • 1. Re: Hardware Update
          JEShort01 Level 4



          Awesome new build! A 138 net time (1st place time for CS5.5 submissions) for CS5.5 on PPBM5 shows that not only is your hardware scorching, but that you have tuned / tweaked your system very well too. Great job!


          Comments / questions:

          - regarding submitting PPBM5 result multiple times, yes that is common; just give your new submission a new name (fabio-pc2, fabio-pc x79, etc.).

          - regarding retiring a pretty darn fast 6-core PC, I have to ask why? (for fun, need to edit RED 4k, etc.)

          - please explain why you think turning on pci-e 3.0 helps when you seem to be running all drives to your motherboard controllers? Does it help the video card performance in ways that help Premiere?

          - do you have 5 total drives now? And where do you put what files when you ran PPBM5 (i.e. projects/media/scratch on 2x600 R0, media cache/exports on 2nd 2x600GB, Win7 swap on, etc.)?





          • 2. Re: Hardware Update
            Bill Gehrke Most Valuable Participant

            That really is a great system and as Jim said please do submit those results. 


            And yes, it would be very interesting to know the results with just your GTX 680 on PCIe 2.0 versus PCIe 3.0 at any given clock speed.  Also do not forget that any result is subject to at least +/-1 second because we only have a 1 second resolution clock.  Also with disk drives because they are asynchronous, on repeated testing runs with Arrays you will find +/- 2 seconds typical variance and as the drives fill up the Disk I/O performance will decrease.

            • 3. Re: Hardware Update
              Fabio Pis Level 1

              Hello Jim and Bill.


              My config:


              Corsair 950 Tx

              case Cooler master 690 II advanced

              i7 3960X @ 4.4 Ghz (automatic asrock oc tweaker)

              32 gb Kingston cl9 (9  9  9  27) 1600 DDR3

              Mb asrock x79 extreme9

              Phanteks PH-TC14PE cpu cooler (2 140 fans)

              Soundcard included with Mb in pci -e slot

              Corsair Force GT 240 gb

              2 Hd velociraptor 600 Gb each configured in in raid 0

              2 hd Wd 2Tb (crap ears version)

              Lg BH10ls30 bluray recorder

              Gainward GTX 680 2gb (for testing no overclock on this)

              Windows 7 ultimate 64 bit sp1 all updated

              Production Creative Suite CS 5.5 updated

              All mb drivers installed and Nvidia Geforce 306.02 beta installed

              A basic tweaking with msconfig (disabled acrobat and other unusefull services)

              No antivirus

              No other software installed.


              My system in this config had passed 1 hour of limpak in OCCT and standard prime 95


              Boot disk (corsair force GT 240 gb)is installed on sata 3 intel connector

              Raid 0 disk is installed on marvel 9220 connector


              With aida 64 disk benchmark I had best result with ssd drive than raid0 drive setup but worst in PPBM5.


              Os and creative suite  disk, and swap disk:

              Corsair force gt


              Project disk:



              With these settings I achieved previous posted results.

              The only differents I noticed in pci-e 3.0 vs 2.0 are related in the first values

              48-55 vs 49-51



              With ssd drivers for project disk I have the total benchmark 6 points greater than raid 0


              For Jim,

              Thank you

              In a chat session with a guy at adobe international I have asked a link for old premiere 5.04 topush my system limits( I am a production premium legit user, but he says that it is not possible to do with my serials..)


              i am used to name may pc with my name fabio, windows add -PC, But as you say I can modify name in benchmark submitting

              I am a pc entusiast and a 3d stereoscopic expert for a part of my job. As I can,  I update my pc. I did some test with a Xeon 2687w but it is not a good choice for premiere and also with 3dsmax because the lack of overclocking part, with the same motherboard I achieved only minimal (from 100 to 102 for fsb) overclock. So I decided for 3960x


              I would like to test in a future Intel sata 3 connector for my raid 0, but I have to restore my clean image. Now, I have installed avg free antivirus and some other drivers for stereoscopic and phisique, by the way, stereoscopic service in my previous test with 980X processor had a bad influence in PPBM5 benchmark results


              For Bill,

              I have submitted yesterday my tests two time because in first submit I did an error with files.


              Message was edited by: Fabio Pis

              • 4. Re: Hardware Update
                Fabio Pis Level 1

                All my tests with simbols to understand better




                gt= Corsair Force GT 240 Gb

                Raid 0= 2 Velociraptor 600 Gb each

                clean= OS clean installation with CS 5.5 and all system and CS 5.5 updated

                pcie2 .0= gfx in pcie2.0

                pcie3.0= gfx in pcie3.0

                BV= black viper optimization

                CR1= ddr3 command rate set1

                no MPE= no mercury engine

                GFX oc= gtx 680 overclocked with custom dynamic clock table


                pcie 2.0:


                gt clean pcie2.0 =                                                         "51","55","36","4"

                gt clean pcie2.0 no MPE =                                         "51","55","36","55"


                Raid 0 clean pcie2.0 =                                                 "49","55","35","5"

                Raid 0 clean pcie2.0 no MPE =                                 "49","55","35","54"


                pcie3 enabled with nvidia enabler:


                gt clean pcie3.0=                                                           "49","52","35","4"

                gt clean pcie3.0 no MPE=                                           "49","52","35","55"


                Raid 0 clean pcie3.0=                                                   "49","52","36","4"

                Raid 0 clean pcie3.0 no MPE=                                   "49","52","36","54"


                gt clean pcie3.0 BV=                                                     "50","53","36","5"

                gt clean pcie3.0 BV no MPE=                                     "50","53","36","56"


                Raid 0 cleanpcie3.0 BV=                                              "49","53","35","4"

                Raid 0 cleanpcie3.0 BV no MPE=                              "49","53","35","56"


                gt clean pcie3.0 BV CR1=                                            "51","53","35","5"

                gt clean pcie3.0 BV CR1 no MPE=                            "51","53","35","55"


                Raid 0 cleanpcie3.0 BV CR1=                                    "48","51","34","4"

                Raid 0 cleanpcie3.0 BV no MPE CR1=                    "48","51","34","55"


                gt cleanpcie3.0 BV CR1 GFX oc=               "50","48","34","4"

                gt cleanpcie3.0 BV CR1 GFX oc no MPE=              "50","48","34","54"


                Raid 0 cleanpcie3.0 BV CR1 GFX oc=        "49","48","35","4"

                Raid 0 cleanpcie3.0 BV CR1 GFX oc no MPE=      "49","48","35","55"    


                So my best score is 136

                • 5. Re: Hardware Update
                  Fabio Pis Level 1

                  This is the last test I'll submit.. I am happy butI think  there are some problems with adobe premiere newest version mpeg dvd encoding.

                  I have the chance to try on my system a legit 5.03 version (5.0 then updated at 5.03)


                  My result:


                  "57","22","39","60" with no mpe


                  "57","22","39","4" with gtx 680 added in cuda_supported_cards.txt file



                  Is it possible?

                  • 6. Re: Hardware Update
                    JEShort01 Level 4

                    Congrats again Fabio, a great score indeed (rank #1 on PPBM5 and beating out the massive SR-2 system from ADK)! I certainly believe it; the X79 CPU really seems to crank on Blu-ray encoding compared with the earlier generation CPUs.


                    Regarding CS5 appearing faster than CS5.5/6 on the PPBM5 benchmark, it is my understanding that this has something to do with changes Adobe made to caching in the different versions. I suspect that this is a PPBM5 artifact however, and would not really benefit CS5 doing real work where video clips are not repeated on a given timeline. Perhaps Harm or Bill have more to add.





                    • 7. Re: Hardware Update
                      Fabio Pis Level 1

                      Thank you Jim,

                      Maybe I could achieve better results If I had used Black Viper windows optimization services config but I didn't in 5.03.

                      In my system there is no different at all if I use raid 0 (two velociraptor 600 gb each one) or a single ssd 240gb for project disk.

                      By the way, installation was clean and I forgot to set pci-e 3.0 nvidia enabler ( I have to add to my clean os and drivers image), default gfx clocks


                      Yes blu ray encoding is very fast, but in disk's test I have 10 seconds delay after encoding is finished, this result can change about 10 seconds in different start(from 57 to 67 s in cs 5.0 suite)


                      I have tried a xeon 2687w with a similar setup of my test configuration  and it is slower.

                      I would like to hear some comments from Harm and Bill also.


                      For me the secret is to keep clean OS with less services installed .

                      In my dayly use ( with 3dvision driver,3dsmax satellite, itunes and other programs installed) my result are greater about 7 points also disabling related services....