-
PCIEx8 Push to the limit with Netstor NA250A
I want to demystifying the must of x16 slot.
My system
Hackintosh with
– X16 Slot ATI for gui
– X8 Slot for Netstor NA250A pcie x8 host adapterNetstor NA250A PCIEx8 Expansion Box with:
– Slot 1 – GTX480
– Slot 2 – Decklink Multibridge
– Slot 3 – Sonnet E4P
– Slot 4 – Redrocket
– Slot 5 – Free
– Slot 6 – GTX480Candle test, with different source R3D 4K 16/9 Full debayer 10bit red default 1920×1080 timeline
0 Nodes – 24fps
2 Nodes – 24fps
4 Nodes – 24fps
8 Nodes – 24fps
12 Nodes -21fps
NR4 – 24fps
NR3 – 24fps
NR2 – 21.5fps
NR1 – 7.5fpsSaludos,
Jose
-
5 Replies
-
sweet 🙂
A slice of color…
DaVinci 8.0.1 OSX 10.7
MacPro 5.1 2×2,4 24GB
RAID0 8TB eSata 6TB
GTX 470 / GT 120
Extreme 3D+ WAVE -
I wonder where the 8x will bottleneck. I wonder if we could get some benchmark from a 16x box, or an explanation where the 16x vs. 8x bottlenecks could be. 2 gtx470s and a red rocket sound like a lot to be going through one single pci8x slot.
Joseph Mastantuono
http://www.goodpostny.com
Color Grading & Post Production Consulting -
I will do some real world benchmarks tomorrow.
I just switched from a 1×570 to a 2×570 system and I cant feel ANY difference in my tree.
Not in playback or render speed.
Maybe in nodes numbers, not in fps.Funny.
A slice of color…
DaVinci 8.0.1 OSX 10.7
MacPro 5.1 2×2,4 24GB
RAID0 8TB eSata 6TB
GTX 470 / GT 120
Extreme 3D+ WAVE -
I move one of gtx480 from netstor to the computer and the speed is the same. Maybe with 2 redrockets…
-
maybe the difference would be if you have more resolution playing.
I have noticed that in some point your gpu and other io does not matter as they have so plenty headroom.
I consider it to be matter of headroom, if there is any headroom left for this resolution and node number.Correct me if i’m wrong.
—
Margus
Reply to this Discussion! Login or Sign Up