- Dec 29, 2022
- 2
- 0
- 10
I'll start by saying I'm sorry in advance if I didn't post this in the correct category- this is my first post and I couldn't find "one" specific category that would suit it so I figured this was the best spot... I should also mention that I have literally over 20 of these Samsung NVMe SSD's, so believe me, I am very aware that they aren't practical for my purposes but there's no funds going into this project; I already have all the hardware mentioned so it wasn't purchased for the intent of this project lol. I like tinkering and learning so if this is just plain stupid feel free to tell me!
Important Notes:
I will be using the Razer Core (w/ 4 X Samsung 980 PRO NVME 2TB) as an external drive for backups and file storage. I will not be using the Razer Core at the same time that I am using other intensive software. For example, I won't be using the Razer Core at the same time I'm doing work in Blender. I mention this because I do not care if the GPU's performance is TEMPORARILY hindered only while the Razer Core is in use... The goal of this is to make a removable, extremely high-performance SSD with access to the CPU's PCIe lanes. The idea was to create a perfect balance for people like me who do quite a bit of everything. I do vehicle scanning (Einscan HX via CUDA), 3d modeling (Blender / 3DS Max), CAD work, web design, software development, billing & invoicing, and quite a bit of remote desktop services. With that being said, I need my GPU every day, but I don't need the Razer Core every day.
I'm gonna make this super simple. I am currently using 20 of the 20 available PCIe lanes available to my i9 13900K- 16 with my RTX 4090 and the last 4 with the NVMe SSD. I understand that regardless of what additional hardware I add (relevant to PCIe lanes), something will lose performance or will not be able to perform to it's fullest.
This brings me to Thunderbolt 3/4- specifically e-GPU devices. Let's say I have a Razer Laptop with a Razer Core (e-GPU): you plug in the e-GPU when you want it, disconnect it when you don't. It allows the graphics card in the e-GPU enclosure to communicate directly through the PCIe lanes without being directly installed to the motherboard. I also understand this, but I would like to know how this would work with a system that is already using all of the PCIe lanes. Is this only possible because the Razer laptop is completely disconnecting the internal GPU thus freeing 16 PCIe lanes?
My main question is: if my Desktop is using all 20 of the 20 available PCIe lanes available on my system and I plug in my Razer Core (w/ 4 X Samsung 980 PRO NVME 2TB RAID_0) via Thunderbolt 3 which also requires 16 PCIe lanes, what will happen? Will my graphics card go from 16 lanes to 8 lanes and also give the Razer Core the other 8 lanes? Perhaps the bigger misunderstanding I am having is the PCIe lane capability. If a graphics card uses, say 16 PCIe lanes- is it ALWAYS using 16 PCIe lanes or is able to use UP TO 16 PCIe lanes?
My Current Hardware:
Cougar Conquest 2
Gigabyte Aorus Z690 Xtreme
Intel Core i9 13900K
128GB DDR5 5600mhz Corsair Dominator
Samsung 980 PRO NVMe 2TB
Gigabyte RTX 4090
Razer Katana 1200 Watt
Important Notes:
I will be using the Razer Core (w/ 4 X Samsung 980 PRO NVME 2TB) as an external drive for backups and file storage. I will not be using the Razer Core at the same time that I am using other intensive software. For example, I won't be using the Razer Core at the same time I'm doing work in Blender. I mention this because I do not care if the GPU's performance is TEMPORARILY hindered only while the Razer Core is in use... The goal of this is to make a removable, extremely high-performance SSD with access to the CPU's PCIe lanes. The idea was to create a perfect balance for people like me who do quite a bit of everything. I do vehicle scanning (Einscan HX via CUDA), 3d modeling (Blender / 3DS Max), CAD work, web design, software development, billing & invoicing, and quite a bit of remote desktop services. With that being said, I need my GPU every day, but I don't need the Razer Core every day.
I'm gonna make this super simple. I am currently using 20 of the 20 available PCIe lanes available to my i9 13900K- 16 with my RTX 4090 and the last 4 with the NVMe SSD. I understand that regardless of what additional hardware I add (relevant to PCIe lanes), something will lose performance or will not be able to perform to it's fullest.
This brings me to Thunderbolt 3/4- specifically e-GPU devices. Let's say I have a Razer Laptop with a Razer Core (e-GPU): you plug in the e-GPU when you want it, disconnect it when you don't. It allows the graphics card in the e-GPU enclosure to communicate directly through the PCIe lanes without being directly installed to the motherboard. I also understand this, but I would like to know how this would work with a system that is already using all of the PCIe lanes. Is this only possible because the Razer laptop is completely disconnecting the internal GPU thus freeing 16 PCIe lanes?
My main question is: if my Desktop is using all 20 of the 20 available PCIe lanes available on my system and I plug in my Razer Core (w/ 4 X Samsung 980 PRO NVME 2TB RAID_0) via Thunderbolt 3 which also requires 16 PCIe lanes, what will happen? Will my graphics card go from 16 lanes to 8 lanes and also give the Razer Core the other 8 lanes? Perhaps the bigger misunderstanding I am having is the PCIe lane capability. If a graphics card uses, say 16 PCIe lanes- is it ALWAYS using 16 PCIe lanes or is able to use UP TO 16 PCIe lanes?
My Current Hardware:
Cougar Conquest 2
Gigabyte Aorus Z690 Xtreme
Intel Core i9 13900K
128GB DDR5 5600mhz Corsair Dominator
Samsung 980 PRO NVMe 2TB
Gigabyte RTX 4090
Razer Katana 1200 Watt
Facebook
Twitter
Instagram