Slow Render with High-End System?

Question

Im currently working on a (in my opinion) high-end Windows system for video editing. The system is about 2 years old and has cost a fortune in that time. So Im expecting significantly better speed. So heres my problem:
Im working primarly in Premiere Pro and After Effects. All the media I work with are as a video imported Jpeg Sequences. Often I have multiple Sequences (up to 7 or 8) overlayed and tweaked with Dissolves and plugins like Twixtor. I also use the Adobe Dynamic Link from After Effects to Premiere and vice versa. All the footage is currently in 1080p but in future I will want to render 4K. Im aware that a 4K workflow is probably a pain in the *** so Im surely going to edit offline with 1080p. However  I cant get any real-time playback with all my sequences. I ALWAYS have to render a preview to watch my edits. I dont know if Im just having too high expectations for my system, but Im kinda sure there has to be an issue for this lack of performance. Maybe the Dynamic Link is slowing my system down?
System Specs:
Model : HP Z400 Workstation 103C_53335X
Mainboard : HP 0B4Ch
System BIOS : HP 786G3 v03.15 10/29/2010
RAM : 12GB ECC DIMM DDR3
Processor : Intel(R) Xeon(R) CPU           W3530  @ 2.80GHz (4C 3GHz/3GHz, 2.13GHz IMC, 4x 256kB L2, 8MB L3)
Chipset:
     HP X58 I/O Hub 2x 2.4GHz (4.79GHz)
     HP Core Desktop (Bloomfield) UnCore 2x 2.4GHz (4.79GHz), 3x 4GB ECC DIMM DDR3 1GHz 192-bit
Graphic card : NVIDIA Quadro 4000 (8CU 256SP SM5.0 950MHz, 512kB L2, 2GB 2.81GHz 256-bit, PCIe 2.00 x16)
Harddisks:
      4x WDC WD2002FAEX-007BA0 (1TB, RAID10/SATA600, SCSI-4, 3.5", 7200rpm) : 932GB (C:)
     Intel Raid 1 Volume (4TB, RAID, SCSI-4) : 4TB (D:)
     HL-DT-ST BD-RE BH10LS30 (SATA150, BD-RE, DVD+-RW, CD-RW, 4MB Cache) : k.A. (E:)
Thank you very much in advance for your help and I apologize for any grammatical mistakes since english is not my main language.

Answer 1

Valentin,
I have always called a raid10 a solution for the paranoid in a hurry. It takes 4 drives to give you the capacity and performance of two disks, but gives you security by the mirroring.
Before going into your specific situation, allow me to tell something about volumes and drives, because they can be confusing and at the same time they are very important for optimal performance of a system.
Single disk, not partitioned is 1 disk = 1 volume to the OS
Single disk, partitioned is 1 disk = multiple volumes (not a good idea BTW)
Multiple disks in one raid array is Many disks = 1 volume
Multiple disks in one raid array with partitions is Many disks = multiple volumes (not a good idea either)
Each volume has a distinct drive letter for access.
Partitioning is a thing of the past and should not be used at all on modern systems.
You have to think about volumes more than about number of disks. In my current system I have 27 different physical disks but only 4 volumes. In the old one I have 17 disks and 5 volumes.
Now that we are clear what we are talking about, volumes with distinct drive letters, we can address your situation.
You have TWO volumes, C: (single disk) and D: (4 disks in raid10). Spreading the load across two volumes is more demanding and gives slower performance than using more volume, unless one or more volumes are very fast, as I tried to explain in a previous reply (remember Isenfluh/Sulwald?). If you add a SSD as you intend, you have increased the number of volumes to 3, which will definitely help performance, because SSD's are faster than conventional disks, the pagefile can be stored on the SSD, so all your performance will go up.
Compare your setup with mine with rough estimated figures:
Volume
Valentin
Harm
Transfer rate Valentin
Tranfer rate Harm
C:
1 HDD
1 SSD
125 MB/s
450 MB/s
D:
4x Raid10
1 SSD
250 MB/s
450 MB/s
E:
NA
21x Raid30
NA
2,700 MB/s
F:
NA
1 HDD
NA
150 MB/s
These figures are indicative, but do show where the major differences are. In my experience disk setup is overlooked quite often, but has a huge impact on a system's responsiveness. It is the weakest (slowest) link in the chain, afterall and with your workflow, doubly so.
But in your specific case there is something else, and that is your disappointing hardware MPE score. 100 seconds is extremely slow, even for a Quadro 4000. I would be quite normal to see a score around 8 - 9 seconds on such a system, well maybe around 12 - 13 seconds with your ECC memory, but 100 is way too slow. Some background services or processes are interfering with the hand-over from memory-GPU-VRAM-GPU-memory. This can be caused by a myriad of things, but a good starting point would be the BlackViper list of services to set to manual or disabled and taking a closer look at the processes running with Process Explorer. There should normally be less than 50 processes running.
Hope this helps.