Riser card Lower Performance?

Mr_BobDobelina

Reputable
Feb 28, 2014
46
0
4,540
So I bought a Generic 190mm riser card for not so much, around €11.
I didn't buy a premium one since they're hard to get. I'll post a picture of the card/cable later on.

I benchmarked my GPU (GTX 1080) before I installed the card and benchmarked it afterwards, using Unigine Heaven. My score without the card was around 3800 and without i got 3650/3700. Does my risercard influence my performance or something?

I'll put anything I know about the card down below.

cab-e1010-pcie16x-fm-a-228x228.png

16X Riser Card Extender PCI-Express Extension Cable, 19cm [CAB-E1010-PCIE16X-F/M]

CAB-E1010-PCIE16X-F/M

Features:
- Connect more devices easily & quickly to your computer
- Gold plated contacts for best connectivity and long life
- Install PCI-E device freely according to dimension and position of computer case.
- Ideal for small computer & server case, like 1U & 2U.

Condition: 100% brand new and high quality.
Length: about 19cm
Interface: One end is PCI-E 16X male and another end is PCI-E 16X female.


Thanks in advance,
Bob
 
Solution
The riser probably does affect performance a bit. 3700 vs 3800 is only like 2%, so in the grand scheme of gaming if you were getting 90FPS in a game before, with the riser you'd get 2% less, so 1.8 less FPS, or 88.2 FPS. No real difference you'd see.
The riser probably does affect performance a bit. 3700 vs 3800 is only like 2%, so in the grand scheme of gaming if you were getting 90FPS in a game before, with the riser you'd get 2% less, so 1.8 less FPS, or 88.2 FPS. No real difference you'd see.
 
Solution

bloodroses

Distinguished
It shouldn't considering it's still hard wired. The only issues I've heard of is cards not receiving enough power when the cable is too long, in which the cables have molex connectors on them as well. My computer case (RVZ01) uses a riser card, and I was using a ribbon cable beforehand for testing a custom case build. I haven't noticed a performance difference in any of the situations vs in the pcie slot, but it was also with a less powerful video card (7770).
 

Mr_BobDobelina

Reputable
Feb 28, 2014
46
0
4,540


Thanks for the fast answers guys, I'll just accept it for how it looks :)
 



Good question, and to be honest.. Not a question I will try to answer since I have no idè :D
 

Light travels about 30 cm in 1 nanosecond, which is 1 billionth of a second. So in one cycle of a 4 GHz CPU, an electrical signal only travels about 7.5 cm. Yes, CPUs have gotten so fast you actually do have to worry about how far light can travel in a single CPU cycle.

So basically his test found no change in framerate because he ran benchmarks which pre-loaded everything (models, textures) onto the video card. If he'd been playing a real game which required dynamically loading models and textures, there would in fact be a slight performance degradation. 3 meters corresponds to a signal delay of 40 CPU cycles (for a 4 GHz CPU). This delay is repeated each time the CPU/RAM/SSD has to send data to the GPU, and vice versa. It's not much, but it is there.


The thinner your wires and the higher the current, the more voltage and power is lost per length of wire the signal has to travel. For a PCIe connector transmitting 75 Watts @ 12 Volts, the current is 6.25 Amps.

http://www.calculator.net/voltage-drop-calculator.html

According to the calculator, using 18 gauge wire (most common gauge for hand-wired electronics), the voltage drop over 3 meters (10 ft) is 6.67%, or 0.8 V. The power reaching the card would thus be only 70 Watts. Not great, but not terrible either.

If you go down to 24 gauge (kind used in HDMI cables), the voltage drop over 3 meters would be 27%, or 3.2 V. Your 12V power would now be 8.8 V, and the card probably wouldn't run. Even if it did, it would only be getting 55 Watts.
 

dan3141592653

Reputable
Feb 23, 2015
300
0
4,780


you're pretty good