NVIDIA Delays RTX 3080 Founders Edition Reviews, and shares GeForce RTX 3070 launch date

It is just strange its delayed two days before the review and after leaked lets say not accurate benchmarks.

 

I wouldnt worry, you'll get your fix soon enough

 

2 days delay does literally nothing to em , who ever wanted to preorder already did i bet , so the most logical reason i can think of is either a) shipping problem to retailers or reviewers or both or b) they needed to delay a bit to add some final touches to the drivers .

 

Maybe totally off target but am saying big navi jitters and last minute driver optimisation

 

Perhaps identified a driver issue with "leaked" benchmarks.

 

I thought you couldn't even preorder them yet, at least officially?

 

You can't, they cancelled all the Amazon preorders people had (AFAIK anyway, both my Amazon/B&H preorder got cancelled and there was a thread about it on reddit)

This thread is just filled with people that are a step away from thinking the earth is flat.

The mental gymnastics one has to have to think Nvidia can just magically optimize drivers in 2 days to deliver gains enough to render AMD's non-existent lineup obsolete is just mind boggling.

What benchmarks were inaccurate? How do you even have accurate benchmarks? Accurate to what?

 

I'll bet a lot of ppl are looking at 1080p benches and snickering what a fail Ampere is.

 

This is weird, what two more days will give them anyway? Ugh... RTX 3080 please be at least 40% faster than RTX 2080 Ti, I want to switch to RTX 3080 16GB or 20GB version in October..

 

I know, it's ridiculous. In the other thread I'm convinced @wavetrex wasn't even doing the math between cards properly, he was just taking the difference from videocardz graph and assuming that was the uplift but % difference is not the same as % change.

As it stands:

Nvidia stated "Up to 2x 2080" - which literally means 1 game can benchmark 2x a 2080 and the rest can do 1% more than a 2080 and that statement would be technically accurate.

Digital Foundry is the most reliable source of information we currently have. They showed two games pushing past 2x (Doom/Quake) and they showed others at roughly 70-80% faster than a 2080. One of which was Rise of the Tomb Raider.

The leaked Video Cardz benchmarks don't show the two games that pushed passed 2x at all, so it doesn't invalidate those claims - but they do show Rise of the Tomb Raider and they confirm the 70-80% faster and show a few other synthetic benches at 70-80% faster. Which falls in line with what we know.

So I don't know how you sit there and say the leaked benchmarks show that it's "inaccurate" when there is no base level of accuracy - "up to 2x a 2080" can literally be anything faster than a 2080 and the benchmarks we did gain, show it in line with what Nvidia is basically reporting.

It gives time for the reviewers to review the cards and that's about it. There is no magic driver optimization that's occurring in 2 days like others in this thread suggest. There may be a patch but it will be to fix a specific issue or thing. In the past when reviewers did get last minute patches they always disclosed it anyway.

The thread is just filled with people that would rather write some weird fantasy narrative - for most of these users it's a reoccurring one. It's gotten to the point where I click on these threads and know exactly whose posting in them and what they are going to say.

 

If it ends up being like the 7000 series from AMD back in the day, we might be in for a treat, my 7970 saw driver patches of 30-50+ % increases sometimes, that was the last crazy generation I remember when it came to driver wars

 

Well, that's the thing - people rarely read between the lines like you did now and that's exactly why nVidia marketing works so well. People see "2x" and decide that's exactly what it will be regardless of the "fine print".

I agree that DF is reliable but this time they were offered to run a few nVidia approved games within an nVidia approved controlled environment. That's why I take these results with a huge bag of salt as they may only show the absolute best case scenario.

 

found in a comment on videocardz supposedly leaked prices from caseking.de
https://www.tomshardware.com/news/p...market-cards-revealed-on-german-retailer-site
if true and not with some fishy over-pricing in each of our respective countries Nvidia made probably one of the worst "finger flipping" to previous customers I've seen
2080ti last week were still (converted) 1300USD in switzerland, if true you will be able to buy a 20+% better card for 60% of the price it's crazy !

I didn't buy a RTX 2000 but I feel for the buyers as I lost "slightly" more than that when I bought a Subaru WRX STI right before some crisis made Subaru drop the price here by 4000USD less than 6 months after I bought the car :/ (it's not even sold anymore here they gave up)

hey while I'm here please find a way to use borderlands 3 in badass settings as a benchmark it's HEAVILY gpu limited..so much that
1) it makes my 1080Ti have the same or higher usage than gpu stress tests like kombustor or furmark (except you don't stress test your gpu for 4 hours usually)
2) I tested myself on a 9900k how Cpu bottlenecked it was and found that between oc 5.1Ghz and a voluntarily underclocked 4.2Ghz I only had a 0,5fps fps difference, this game is 99% gpu 1% cpu

EDIT : MSI seems the 1st to fire a price reduction in my country it's 2080Ti Gaming Z Trio dropped last week from 1200$ to 899$ other brands still show insane prices 1300-1600$
you can see the price drop in a graph here https://www.toppreise.ch/price-comp...-Z-TRIO-GeForce-RTX-2080-Ti-V371-234R-p608394

 

Nvidia, in the slide with the chart, stated UP TO 2x at 4k, but in the presentation slide of the 3080 stated twice the performance of the 2080

https://www.guru3d.com/articles_pages/geforce_rtx_3080_and_3090_what_we_know,2.html

So I wouldn't get NVIDIA marketing completely off the hook...

And 2070s≠2080≠2080s...

https://www.guru3d.com/articles_pages/geforce_rtx_2080_super_review,20.html

 

Maybe reviewers don't want to have to work entire weekend doing benchmarking. So they can do it on Monday/Tuesday instead.

Also it is obvious now to me who are the "A grade" youtube channels according to NVidia and who are not.... from those that got a 3080 to review and those that did not (one that did not get one even said it on video and complained he would have to buy one himself now)

 

Tbh, reviews on YT are the last I look at and only if they offer something unique that may not be in printed reviews. Printed text reviews are what work for me, because you can instantly refer back to the charts and info you are looking for at a moments notice. Cant do that on YT. Plus printed reviews are more thorough and detailed.

 

Same here, if anything the "FineWine" was AMD and developers managing to extract the idle bubbles in GCN for more performance. I think that Ampere will fly in Async Compute scenarios, and I trust that it will take Nvidia much less time than it took AMD to extract the performance.

 

I think review will be on 3080 AND 3090

 

These target as are a scam.