Yeah, memory bandwidth helps, but more importanly the system's memory acccess latency goes down with the number of channels used to access it.

I have 64GB in my system now, not because I needed that much memory, but because I wanted 4+4 channels (each CPU has a Quad controller) to access my DDR3 memory (T7610 here, with two E5-2667 V2). So I have installed 8 sticks of 8 MB in my machine (of 16 slots total), 4 sticks for each CPU (to use their Quad channels)

Sure it's not a linear dependency, due to other bottlenecks (like the L3 level CPU, or the two QPI links, or PCI-Express 3.0 latency moving data between CPU and GPU).

PS: You really need to check the recommended Dell fill for your memory slots, especially in your case with un-equal sticks. Best way to go is to fill all slots equaly.
One of your CPU's has memory in Dual channel, the other in Quad. There will be times when the Quad ch CPU will wait for a result that is in the Dual ch CPU memory.

Quote
I have 64GB in my system now, not because I needed that much memory, but because I wanted 4+4 channels (each CPU has a Quad controller) to access my DDR3 memory (T7610 here, with two E5-2667 V2). So I have installed 8 sticks of 8 MB in my machine (of 16 slots total), 4 sticks for each CPU (to use their Quad channels)

Thanks I would have never thought of that. I had four 16GB RDIMMs sitting on the shelf which is why I ended up with 128GB total - I don't need that much memory. I shall experiment - I will borrow some memory from another workstation and play around...

Quote

Thanks I would have never thought of that. I had four 16GB RDIMMs sitting on the shelf which is why I ended up with 128GB total - I don't need that much memory. I shall experiment - I will borrow some memory from another workstation and play around...

Yeah, I don't "need" 64GB either, but I wanted to have Quad channel used on each CPU for latency and bandwidth.
My CPU has a max mem BW of 59.7 GB/s, but that's only in Quad channel mode.
Your CPU, since it has DDR4 controller, has a max mem BW of 68 GB/s... but again only in Quad mode. In Dual channel mode it halfs that.
CPU-Z is a good little software that can show various CPU-related things, including the mode of connected memory.

LE: Comparation between E5-2667 V2 and V3:

https://ark.intel.com/content/www/us/en/ark/compare.html?productIds=75273,83361

Quote Yeah, memory bandwidth helps, but more importanly the system's memory acccess latency goes down with the number of channels used to access it.
.

I have not had a chance to try using all 8 memory channels yet but I did get and interesting surprise when I added one of the new StarTech dual NVMe x8 cards. The rederening time of my test project dropped 25% and I noticed the GPU usgae went from 41% to 53%.

This is using two Samsung Evo 970 Plus NVMe SSD's. They were previously each mounted on the usual x4 adapter card. The only change in the system was removing them from the two x4 cards and moving them to the new single StarTech x8 card...

Quote This is using two Samsung Evo 970 Plus NVMe SSD's. They were previously each mounted on the usual x4 adapter card. The only change in the system was removing them from the two x4 cards and moving them to the new single StarTech x8 card...

Congrats, you just found a bottle neck. I don't think it's bandwidth related, but more of latencies.

I just installed PrimoCache, to use some of the 64GB of RAM that I have as... RAM cache. I have allocated 8GB of RAM as cache to my 1TB SSD (OS and programs) and 16GB of RAM as cache for my 7.5TB RAID5 HDD array.
It's awesome!

Quote

Congrats, you just found a bottle neck. I don't think it's bandwidth related, but more of latencies.

Yeah I am becoming a believer that how stuff is moving around is the limiting factor. - not so much the speed of any part

Hardware Encoding Utilizes My GPU where if I do not have that button pressed it will use 100% of my CPU.

SVRT does little else but use the CPU on my unit.

Asus ROG GL702VM 32 Gb RAM - 4 core 8 Thread i7 - Nvidia GeForce 1060 GPU.

Quote Hardware Encoding Utilizes My GPU where if I do not have that button pressed it will use 100% of my CPU.

SVRT does little else but use the CPU on my unit.

Asus ROG GL702VM 32 Gb RAM - 4 core 8 Thread i7 - Nvidia GeForce 1060 GPU.

That's exactly what's expected. HW = max GPU usage; no HW box checked = 100% CPU encoding; SVRT uses little to no encoding at all and the CPU does all the work.