PCI Express Scaling Game Performance Analysis review
Click here to post a comment for PCI Express Scaling Game Performance Analysis review on our message forum
k3vst3r
stevevnicks
just wondered because it says in your sound card part SB X-FI Titanium , which i thought od that cant be right lol
ruiner13
The highest bus usage I've seen on my rig with 780s in SLI at 3.0x16 is 45%, and that was during the Monster Hunter benchmark. I'm guessing at 2.0x16 that it would be near maxed. Any chance you could do a test with that benchmark, Hilbert?
Cave Waverider
I'd like to see the bridgeless CrossfireX tests. Also, I'd like to see 8x and 4x tests on single cards and SLi/CrossfireX.
4x single card would be interesting to see how much of an impact it would have to run a state of the art video card with Alienware Graphics amplifier for their newest line of laptops which only runs at 4x.
Fox2232
If I buy new graphics card for $700 and it would under perform by 3%, I would not go and buy new intel system with PCIe 3.0 x16 for another $800.
If I needed more graphical power, I would simply go and buy 2nd GPU for another $700.
And I would laugh as this would be under performing by 10%, but would still provide 80% higher performance of guy who got 1 GPU + 1 extra modern MB+CPU+Memory.
DeskStar
My x16-x8-x8-x8 is screaming along just fine. I will have to max out that monster hunter benchy today and see what the bus speeds report on this particular setup.
"I'm very excite....!"
Agonist
I was kinda worried about running 2x R9 290 on my motherboard as it is 2x 16 2.0 and 1 2.0 4x.
I run a 9800gtx+ for hybrid phsyx in the x4 slot.
I knew there wasnt that much of a difference but was still worried 2.0 slots would hurt 2 R9 290 crossfire alot.
Thanks for the review/info HH. Awesome job.
albertfu
I am wondering if a 3-way crossfire with x16/x16/x8 configuration will make a difference with x16/x8/x8 or x8/x8/x8, also the impact on performance in crossfire caused by bandwidth limitation under different resolutions.
Earlier in a review done by TPU, I saw the impact decreases as resolution increases, but it was a single card test with 7970/680
I am currently using 290x-290-290 3-way crossfire on a P9X79-WS board which can only do x16/x8/x8, and rampage x79 board can do well x16/x16/x8. But again X99 rampage board can only do x16/x8/x8 according to ASUS's site.
So it would be interesting to see if there is a difference, since Hawaii uses PCIE bus instead of crossfire bridges, so normally PCIE bandwidth would play a more important role in crossfire.
mbk1969
I wonder whether GPU reading system RAM (as I take it CUDA can do it, and modern AMD cards too) depends on PCIE speed. Such reading goes through PCIE bus. Although system RAM is the slowest component here I guess.
iancook221188
when folding A18 work unit my 780 at X8 2.0 use 65% interesting
Reddoguk
It's of no interest really for single card users like myself.
This only matters if you are building a thermonuclear power plant in your house.
The sale of different PCI-e Gens is one of those not really needed upgrades that we are seem to fall for, gotta get that latest an greatest.
X fire with no bridge will be interesting though no doubt.
Good read as usual though from HH, Thanks.
BLEH!
Cool article, cheers boss. I see no need to upgrade yet, not for 5%. You gunna do an AMD one when the new 39X series are out?
Prince Valiant
Nice article, very curious to see how Crossfire will go if you do decide to do a review on it.
lmlim
Would like to see xfire AMD 290x in PCIE 2.0 vs 3.0
tsk2k
Awesome article by Hilbert as always! I for one would love to see crossfire results as well.
Aura89
Cyberdyne
It's pretty cool to see 1.1, a 10 year old standard, still perform so well. Pointless, but interesting.
DeskStar
rl66
rl66