The encoding process is not I/O limited but CPU and/or GPU limited, so it's not surprised that you are not seeing improvements on SSD. Neither I am.

Most of the reads and writes during encoding are sequential, so you don't benefit from the lower access time of SSD. Also, the encoding rate is typically under 10MB/s, vastly less than what either SSD or HDs can offer.

I am using 2x128GB SSDs in RAID 0, and 2x2TB HDs in RAID 0 and I am not seeing notable encoding speed improvements on SSD either.

My RAID-0 SSD speed in the benchmarks like Crystal disk mark is about 900MB to 1 GB/s . My RAID00 HD speed is about 200 - 300MB/s.
MSI X99A Raider
Intel i7-5820k @ 4.4 GHz
32GB DDR4 RAM
Gigabyte nVidia GTX 960 4GB
480 GB Patriot Ignite SSD (boot)
2 x 480 GB Sandisk Ultra II SSD (striped)
6 x 1 TB Samsung 860 SSD (striped)

2 x LG 32UD59-B 32" 4K
Asus PB238 23" HD (portrait)

>>My RAID-0 SSD speed in the benchmarks like Crystal disk mark is about 900MB to 1 GB/s . My RAID00 HD speed is about 200 -300MB/s.

Thanks for that -- I downloaded crystaldiskmark and my single 128 mb ssd c: drive scored 442.1 and 131.3 Mb/sec sequential read and write on the 1000 mb default size. The striped hd scored 233.9 and 234.2 respectively. The graphics card is Radeon HD 5770 in crossfire configuration and the operating system is windows 7 home premium 64 bit.

I also downloaded CrystalMark 2004R3 [0.9.126.451] a suite of system benchmarks from softronic site and on the hdd test the c drive scored 45261 and the striped drive scored 18973.

I am planning a great deal of graphics work using PDir 10 this winter and 85 minutes to render a bdmv from this typical ts file is going to be a burden. I have hardware decoding unchecked in preferences as it causes ghosting on fast motion captured video. I could enable it for rendering events and will give that a try.

Looking for elements to upgrade for better results.

Quote:
I am planning a great deal of graphics work using PDir 10 this winter and 85 minutes to render a bdmv from this typical ts file is going to be a burden. I have hardware decoding unchecked in preferences as it causes ghosting on fast motion captured video. I could enable it for rendering events and will give that a try.

Looking for elements to upgrade for better results.

I don't think there are many components to upgrade to improve rendering time unfortunately.
Are you using Quicksync for encoding ? I don't have it on my AMD but heard it is pretty fast on Intel CPUs.
The Ivy bridge is supposed to have a better implementation of Quicksync also, maybe upgrade the CPU.
Or add a second machine if you have multiple projects ...

How long is the project that's taking you 85 minutes to render ?
MSI X99A Raider
Intel i7-5820k @ 4.4 GHz
32GB DDR4 RAM
Gigabyte nVidia GTX 960 4GB
480 GB Patriot Ignite SSD (boot)
2 x 480 GB Sandisk Ultra II SSD (striped)
6 x 1 TB Samsung 860 SSD (striped)

2 x LG 32UD59-B 32" 4K
Asus PB238 23" HD (portrait)

The file is 2 hours and 12 minutes running time. Your post is my first hearing of Quick Sync. Cannot find an implementation for PD 10. Unchecked all hardware acceleration boxes everywhere I could find them in PD hoping the CPU would introduce QS but apparently not -- the burn process rendered to BMDV file taking 15 minutes more. Suggestions would be welcomed -- searching the forum and Cyberlink help screens has been no help.

I do not seem to be getting whatever rendering benefits sandy bridge is supposed to bring, at least where PD 10 is concerned. Unless that is, 85 minutes is considered good by nominal rendering standards.

I hesitate to just wade into this workload if there is a rendering bunny running around somewhere out there.

Edit: Found much more to research in the forum -- was spelling quick sync incorrectly as quick synch. Correct spelling turned up five pages of hits.

Well Jullien Pierre --

You seemed to have put your finger on what I should be looking for to solve my Pdr 10 rendering issue. There are some active quick sync advocates on this forum and they succeeded in whetting my appetite for that feature. However, --

it turns out my early adopter DP67BG motherboard did not enable that function for the sandy bridge cpu that it came with . That is if the review found at: http://www.computershopper.com/components/reviews/intel-dp67bg is accurate in their statement that --

"You’ll need a motherboard based on the Z68 or H67 chipsets if you want to use the built-in graphics on the latest Sandy Bridge chips, as well as the processor’s hardware-based video-transcoding acceleration (a.k.a. Intel's Quick Sync Video)."

And, this also is probably why your ssd outran mine 2 to 1 as the review goes on to say --

"The P67 chipset also lacks support for caching with solid-state drives (SSDs) using Intel’s Smart Response technology."

So I have my answer -- change out the motherboard. And, your earlier suggestion to install the third generation Ivy Bridge processor at the same time would seem in order. May do the mb first as the pricing on those will have come in a little bit while the cpu will be getting full price at this early stage in the cycle.

Probably do one then the other later. If you or anyone have another thought, it will be welcome -- Thanks JP for your insight.

Lewer, to be fair, I have not used the Quicksync myself.
I currently use AMD CPU - an FX-8350 8 core, overclocked to 4.6 GHz. It doesn't have Quicksync as that's an Intel feature. This CPU is in the price range of the Intel i5 .

My GPU is an nVidia GTX 560 Ti , and it renders about twice as fast as the CPU. I tested a top of the line nVidia GTX 690 that I bought at the store and saw zero improvement in rendering speed unfortunately.

The Intel mobo and CPUs are pricier. And Intel has a bad habit of changing their socket seemingly every year.

I don't think the SSD caching support on the motherboard matters at all. Just add a bunch of RAM and let the OS do the caching. I have 32GB which allows many of the video clips to stay completely cached. This makes for extremely fast editing. However, it does nothing for rendering speed.
The reason my SSD is very fast is that it's two SSD in RAID. Both of them are SATA III (6 Gb/s) . The max read speed on SATA III is about 600MB/s per port. Effective on the SSD I have (Crucial m4) is about 450- 500 MB/s . But when using RAID-0, you double that. So it's 900 MB to 1 GB/s. Of course the SSD is only 256 GB. That's small, so my video clips are mostly on the HD (4 TB RAID). But again the OS does the RAM caching for the files on the HD as well.

The nVidia GTX 690 was supposed to improve encoding speed with the nvEnc chip (google their press releases on nvEnc). But apparently the support for PowerDirector still has not been enabled. That's why I returned my GTX 690 to the store. Perhaps some day they will enable this support in PD finally.

I personally will wait for faster GPUs to improve the encoding speed.

I do have two machines on a KVM. The other is an AMD Phenom II x6 OC at 3.6 GHz and with 16GB of RAM. So I can use both to render. The HD/SSDs are shared through Gigabit ethernet. This way I can render 2 projects at once if I wish.
MSI X99A Raider
Intel i7-5820k @ 4.4 GHz
32GB DDR4 RAM
Gigabyte nVidia GTX 960 4GB
480 GB Patriot Ignite SSD (boot)
2 x 480 GB Sandisk Ultra II SSD (striped)
6 x 1 TB Samsung 860 SSD (striped)

2 x LG 32UD59-B 32" 4K
Asus PB238 23" HD (portrait)