Not real world, but the only VR benchmarks i could find.
https://www.overclockersclub.com/reviews/nvidia_geforce_rtx3080_founders_edition/4.htm
Third of the way down the page. VRmark.
Not real world, but the only VR benchmarks i could find.
https://www.overclockersclub.com/reviews/nvidia_geforce_rtx3080_founders_edition/4.htm
Third of the way down the page. VRmark.
MSI RTX 3080 GAMING X TRIO scores 113 fps with superposition benchmark
compared to my 2080ti overclocked ( no shunt yet ) of 97.68 at 4k
vr mark is on sale at steam for only £2.24p
just downloading it now to compare
Interesting, when buying via the NVidia site, the XC3 black would cost 84⬠more than on Alternate (where it was 699ā¬). FE cards are also sold out here in Germany. (Not relevant for me anyways because of the case requirements).
Letās see which effects these bins have, if EVGA pre-selects, then the XC3 black is likely more in one of the lower bins. Iām more in the moderate undervolting than the overclocking camp, so it might not be as important for me.
Ugh, this went fast, the price for the 3080 XC3 Black on Alternate was raised from 699⬠to 779⬠already. Lucky meā¦
Consistent with a 40% to 55% improvement. I am betting closer to 40%
(113/97)*(10496/8704)
(85/66)*(10496/8704)
It would make my day if 3090 VR SLi support was announced.
3080 Vulcan anf VR performance compared to my 2080ti o/c card
Vulcan performance with basemark gpu test
MSI 3080 trio =200 fps
2080ti o/c ( no shunt ) = 163 fps
23% increase
vr mark orange room
3080 FE = 12649
2080ti o/c ( no shunt ) = 11925
6% increase in vr mark orange room
If the 3080 can only overclock 5% or so , itās only about 10% faster than my ( not very well) overclocked 2080ti in this vr benchmark
Since the 3080 and 3090 are similarly rated in terms of boost speeds, most of the difference between these cards will probably be core count.
10496/8704=1.206
Any 3080 benchmark can probably be multiplied that much to get a 3090 benchmark.
Are you suggesting that a 3090 will only be about 30% faster than an overclocked 2080ti for VR stuff ??
About 40%, yes.
Looks like another 10% could be squeezed out by chilling to -30degC or even -50degC. I am seriously considering that now.
minus 274C⦠202020202
My guess is the running cost of LN2 would be dollars per hour when cooling a kilowatt load, and would obviously require purpose designed equipment to be cost effective. But that would extend GPU performance almost 30%.
https://xdevs.com/guide/2080ti_kpe/#thermals
EDIT: Since -30degC is below the boiling point of R134a, and presumably most other common refrigerants, just getting 10% already calls for specialized equipmentā¦
@mirage335 do you think 3080ti will have better clocks?
Also would you rather to wait for 3080ti or get 3090?
*$700 more for 3090 to get ~10% better performance than 3080⦠what you think about this? Worth it?
There has been no evidence to suggest that a 3080Ti will even exist, much less perform better than a 3090.
At this point, I am probably just going to wait and buy whatever is best the day I have finished putting together some kind of chiller that reaches the point of diminishing returns. Given the hardware development resources I now have, and the clear desirability of -40degC coolant, that is probably not going to be using typical refrigerantsā¦
I am thinking multistage.
I am not happy with what I see from NVIDIA so far, that is for sure.
lol, nvidiaās site here just forwarded to the normal electronics stores where they have no FE versions. So no FE and before you get to the store page with the extra clicks, not to mention suddenly have to make up your mind what to then go for, theyāre all sold out.
One of the last open questions might be answered here:
Shunt modding (unlocking the power target) of a custom 3080:
Edit: He only mods the card and verifies that power measurements can be tricked as expected, but doesnāt overclock the card yet. Cliff hanger towards the next videoā¦
it looks like a 400w power modded Titan RTX will remain king.
Thank you Nvidia that, thanks to your extensive advertising for the 30xx, I get a brand new T-rex for half the price.
Besides, itās still a good time to get a cheap 11Gb 2080ti?
LG CX48 is 120Hz 4k OLED TV (HDMI 2.1) that supports Gsync.Thatās my target. Just need to make room for a 48inch screenā¦
I managed to get my 3080 ordered off Scan today but donāt yet have an 8k-X to test it out on
Where did you find that?
Also, it would not be too surprising if power modding significantly improved the 3090 as well. After all, it does seem NVIDIA was afraid of making these cards dissipate 350W, stock.
the shunt mod is an old story.
Bur you donāt find any direct benchmarks in the net.
Al comparsions are on Stock.
The weakness of the Titan is that it is capped at 280W and immediately runs in power limit. Therefore 2080ti with 330, 350 or 380w Bios are faster in this benchs.
The first Titan with Mod appeared in the OC Forum in January and are up to 22% higher than the max OCed Ti.
But the Trex are very expensive, so i have no look on it before. But thanks the 30xx, there are some people in panik mode.
I can use my old Ti watercooler, sell the Ti and have a 3090 like card for 300 + XXX Bucks for the cooler less.