Are calculation speed difference depending on the interface?
For example, there a difference when rendering speed, even if wrong, as the bandwidth PCI-E 3.0 X16 with PCI-E 2.0 X4 or X8?
PCI-E bandwidth..
Moderator: juanjgon
Nope, barely difference. Search the forums.
Most likely only the sceneloading speeds will be affected a bit.
Most likely only the sceneloading speeds will be affected a bit.
Octane 2022.1.1 nv535.98
x201t - gtx580 - egpu ec
Dell G5 - 16GB - dgpu GTX1060 - TB3 egpu @ 1060 / RTX 4090
Octane Render experiments - ♩ ♪ ♫ ♬
x201t - gtx580 - egpu ec
Dell G5 - 16GB - dgpu GTX1060 - TB3 egpu @ 1060 / RTX 4090
Octane Render experiments - ♩ ♪ ♫ ♬
-
BorisGoreta
- Posts: 1413
- Joined: Fri Dec 07, 2012 6:45 pm
- Contact:
PCI express speed determines how fast the scene data is loaded into GPU memory. Once it is loaded it doesn't affect render speed.
PCI Express 2.0 x1 500 MB/s
PCI Express 2.0 x4 2GB/s
PCI Express 2.0 x8 4GB/s
PCI Express 2.0 x4 15.75GB/s
So an average scene of 2GB loads in 4s, 2s, 1s or 0.25 s.
Plus maybe some data preparation overhead.
If you are rendering animations the complete scene data gets transfered only at the beginning of a sequence, for every consecutive frame very little extra data is transfered, camera motion, possible mesh deformations etc. so here PCI speed doesn't matter that much.
If you have an Octane network render node than PCI speed doesn't matter at all because your network node is connected via UTP cable which can support only 100MB/s (Gigabit betwork) where the slowest PCI Exptess speed is 500MB/s so it doesn't make sense to have higher PCI speeds in your node. 10 Gigabit networks are emerging but are very expensive. They support 1GB/s speeds.
PCI Express 2.0 x1 500 MB/s
PCI Express 2.0 x4 2GB/s
PCI Express 2.0 x8 4GB/s
PCI Express 2.0 x4 15.75GB/s
So an average scene of 2GB loads in 4s, 2s, 1s or 0.25 s.
Plus maybe some data preparation overhead.
If you are rendering animations the complete scene data gets transfered only at the beginning of a sequence, for every consecutive frame very little extra data is transfered, camera motion, possible mesh deformations etc. so here PCI speed doesn't matter that much.
If you have an Octane network render node than PCI speed doesn't matter at all because your network node is connected via UTP cable which can support only 100MB/s (Gigabit betwork) where the slowest PCI Exptess speed is 500MB/s so it doesn't make sense to have higher PCI speeds in your node. 10 Gigabit networks are emerging but are very expensive. They support 1GB/s speeds.
19 x NVIDIA GTX http://www.borisgoreta.com
that might become a bottleneck if You have something like those: ASRock WS-E/10G motherboard offers 22Gbps networking =)BorisGoreta wrote:
If you have an Octane network render node than PCI speed doesn't matter at all because your network node is connected via UTP cable which can support only 100MB/s (Gigabit betwork) where the slowest PCI Exptess speed is 500MB/s so it doesn't make sense to have higher PCI speeds in your node. 10 Gigabit networks are emerging but are very expensive. They support 1GB/s speeds.